The Software Freedom Conservancy, a non-profit organization dedicated to legally protecting open-source software, said that Bambu Lab has violated the licenses that its slicer used as it looked into the company after the 3D printer manufacturer threatened to sue independent developer Paweł Jarczak.

Jarczak built an OrcaSlicer fork (called OrcaSlicer-bambulab) that bypassed the company’s Bambu Connect service, and was soon served with a cease-and-desist and a demand for the removal of their work from GitHub. According to the SFC, the company’s restrictions violate AGPLv3, the license its software is based on, which says one cannot put additional limitations on the rights granted under the copyleft agreement.

The company’s troubles began when it accused Jarczak’s fork of impersonating Bambu Studio, bypassing their authorization controls, and violating their Terms of Use. The 3D printer manufacturer also alleged that the developer used “reverse engineering” of its closed-source software and said it “could allow modified forks to send arbitrary commands to printers.” The company says that the bambu_networking plugin included in Bambu Studio, which handles all communication for the slicer, is proprietary. Because of this, it says reverse engineering the protocol that the plugin used, which Jarczak did to create their own Rust implementation, violates Bambu’s Terms of Service.

The company told All3DP in a statement, "The AGPL, the DMCA, and Bambu Lab’s terms do not permit reverse engineering that violates applicable protocols, rules, or circumvents technical protection measures protecting our cloud services." It also added, "From the beginning, our preference has been dialogue, not confrontation. At this stage, rather than escalating conflict, we are focusing on strengthening our own infrastructure and protection measures moving forward. Interim measures have already been implemented. Security will continue to be strengthened in future releases, and we recommend that users update to the latest version in a timely manner."

On the other hand, the SFC contends that because bambu_networking is required to run the Bambu Studio, it falls under AGPLv3, which the Bambu Lab’s slicer is licensed under. Bambu Studio itself is a fork of PrusaSlicer, which was also based on Slic3r. The original Slic3r uses AGPLv3, meaning every derivative under it, including Bambu Studio, must use the same license. The fact that bambu_networking is flagged as proprietary breaks the copyleft license and violates several clauses in the license agreement.

This has been going on for years now, but it’s the company’s actions against Jarczak that finally caused an uproar in the 3D printing community. Joseph Prusa himself, the chief of Prusa Research and the creator of PrusaSlicer on which Bambu Studio is based on, said that it was violating the AGPLv3 license, and that Bambu Lab’s actions raise red flags when it comes to security. Right-to-repair advocates are also rallying behind the developer, with YouTuber Louis Rossman saying that he will support Jarczak up to $10,000 for their legal expenses, with Gamers Nexus matching that amount. Furthermore, Rossman said that he’ll host the fork on his own FULU (Freedom from Unethical Limitations) Foundation GitHub and dared the company to take legal action against him.

Bambu Lab is one of the biggest names in the 3D printing industry, with the company reportedly boasting a valuation of billions of dollars and having overtaken Creality as the world’s best-selling budget 3D printer brand. This is why many enthusiasts are concerned, as it apparently moves away from open-source software and towards proprietary apps. But as long as it uses software based on AGPLv3-licensed components, it has no choice but to comply with those requirements.

As for the trouble that Bambu Lab's response on Jarczak's fork stirred, the company said to All3DP: "We nonetheless regret that our reference to Terms of Service, legal context and a potential C&D understandably came across as a legal threat. That was not the outcome we wanted."

Microsoft continues to make some of the earliest chapters of its operating system history open-source and freely available. Earlier this week, it announced that Tim Paterson's DOS listings, containing source code of the 86-DOS 1.00 kernel, various PC-DOS 1.00 pre-release kernels and utilities, and the Microsoft BASIC-86 Compiler runtime library, were available on GitHub. Microsoft VP Scott Hanselman tied the release to 86-DOS 1.00’s 45th anniversary. The exec confirmed that the code, transcribed from reams of old dot matrix printouts found in a garage, was perfect, "and recompiles byte for byte to the original binaries.”

If you head on over to the GitHub page to snag the code, you will see a photo of Tim Paterson standing in his garage with a pile of yellowed dot matrix printouts in the foreground. These pages contain the code for the software mentioned in the intro, and you can even see the original scans in PDF and PNG format via a link to the Internet Archive. These include the coder’s handwritten notes.

Probably more important to tinkerers, though, is the fact that the work of transcribing the printed code has been completed (for those three mentioned wares). Tips to compile and assemble the sources can also be found on Paterson’s GitHub.

From 86-DOS to MS-DOS

In case you aren’t familiar with the place of 86-DOS (or Tim Paterson) in Microsoft’s history, here's a short refresher. Microsoft took a shortcut and gained a foothold in the OS software market by purchasing 86-DOS from Seattle Computer Products and inventor Tim Paterson for a figure in the region of $75,000.

In the GitHub repository, you can see 86-DOS’s transformation into the PC-DOS 1.00 kernel, alongside code for some well-known utilities (still in use today) such as CHKDSK. As the Microsoft blog asserts, this work “offers rare insight into how MS-DOS/PC-DOS came to be, and how operating system development was done at the time, not as it was later reconstructed.”

So, we have another old DOS release to tinker with. In April 2024, we reported on Microsoft releasing the code for MS-DOS 4.00 under the generous MIT License, allowing tinkerers free rein. It did the same with MS‑DOS 1.25 and 2.11 in 2018. Also in 2024, we coincidentally covered a video demo featuring 86-DOS version 0.1C being taken for a test drive (via the Internet Archive), and now version 1.00 of this OS has hit GitHub, straight from the files squirreled away in Tim Paterson's (the creator’s) garage.

We’re still waiting for any version of Windows to be open-sourced. You have to dig through leaks if you are curious enough to want to investigate the source code for Windows XP, for example.

"Nice radar you got there," followed by "thanks, I just had it jammed," might be the new word exchange among buddy electronics enthusiasts. In a move that might ruffle the feathers of many large companies with exceedingly pricey wares, a Moroccan electronics engineer named Nawfal Motii has designed the open-source Aeris-10 radar system that is purportedly comparable to commercial systems costing $250,000.

Aeris-10 comes in two variants: 10N Nexus with a 3-km range and an 8x16 patch antenna array, and the 10E Extended, capable of reaching up to 20 km thanks to its 32x16 slotted waveguide array. Motii published the entire project on GitHub, including all the necessary schematics, PBC layouts, components, firmware, and software with a GUI for controlling and monitoring the system.

On the technical end, the Aeris-10 uses an XCA7A50T FPGA as a central brain for doing its FFT math, along with Moving Target Indicator (MTI), Doppler-effect estimation of moving object speed, and CFAR false alarm detection control. The figurative spinal cord is an STM32F746xx microcontroller that orchestrates the frequency synthesizers, ADCs, DACs, the GPS, barometer, stepper motors, and cooling setup.

The fact that Aeris-10 offers a true phased array system and ±45° elevation/azimuth adjustments are seemingly its differentiating factors. Prices for electronics are exceedingly floaty in these ship-shinking days, but a brief estimate pins the bill of materials at $5,000 for the 10N and $7,200 for the 10E. Despite the number of zeros on those figures, they're pocket change compared to amounts commanded by off-the-truck offerings. A cursory look puts commercial phase-array systems at somewhere $120k and $200k, and well past those prices for longer-range units.

Motii claims that military surplus radars can be had for $10k to $50k, but those are invariably decades-old tech with next to no spare parts availability. He says that building a DIY system is also a hard ask for a small team, as the testing gear can cost $50k on its own. Describing himself as "a guy in a workshop in Morocco with a soldering iron and an obsession," he took it upon himself to fix that particular problem.

Anyone can hit the project's GitHub page and get their own radar system going, but not everyone might have the necessary electronic and mechanical skills necessary for building one. To that effect, Motti says he's reached an agreement with the Crowed Supply platform, aiming for a Q3 2026 release. The site isn't your standard dice-rolling crowd-sourcing platform, though, as apparently it only accepts fully-designed projects with functional prototypes, rejects 90% of submissions, and claims it never had a scam.

Interestingly enough, this project was originally licensed under the MIT license, but Motti was advised that said license does not protect physical hardware, so it changed to the CERN-OHL-PT license. Should you elect to build your own unit, be aware that the frequencies it operates in are almost assuredly highly regulated in your legal jurisdiction.

The creator of a 3D printer that mines Bitcoin took part in an interview on the Home Mining Podcast over the weekend. Host Two Sats quizzed creator PizzAndy about his adapted 3D printing prototype. In summary, the hybrid device is based on an open-source 3D printer design with a custom bed that is heat-controlled by throttling the Bitcoin mining ASICs attached to it. Yes, the printer bed is basically a heatsink. The compact prototype is capable of a humble 500 GH/s, says PizzAndy, but there are already plans for scaling and a tile-based model using Intel ASICs targeting at least 10 TH/s.

How did PizzAndy come up with this interesting blend of 3D printing and Bitcoin mining technology? In the interview, the creator says the thought initially popped into his head five years ago, when he was heating his basement workspace (in the cold months) to help his 3D printers work better.

Then, in 2024, PizzAndy attended a presentation about heat reuse — and something clicked. The 3D printer/BTC enthusiast thought it would be “a cool thing” to mix these two technologies, but wasn’t sure about its commercial value. Since that time, he has become convinced that a commercial product is worthwhile.

Podcast host Two Sats advises his guest that this kind of hybrid product has to be top quality, a respectable performer in its field, with Bitcoin as a bonus. Then the pair discusses the recent emergence of easy-to-market home heating products, which are dual-purpose Bitcoin mining hardware.

Round the clock business

PizzAndy notes that “this product is mostly intended to exist in a print farm, where printers are printing around the clock, all the time… that’s when the economics of a print farm are not unlike a Bitcoin mine. And if your machines aren’t printing, you’re not making money.”

Learning about the prototype

Andy says that the prototype achieves 500 GH/s when the bed is at 75 °C. The bed temperature is important to the process, so the BTC mining chips are run to target this particular thermal environment. This contrasts with typical BTC Miner optimization.

At this time, PizzAndy notes that the miner chip waste heat isn’t used to heat the extruder – only the 3D printer bed at this time.

Inside the prototype are four ‘BM 1362 AK chips’ according to the podcast transcript. A custom heatsink attached to these chips directly warms the bed. “That’s the heart of this printer, and its mining,” explains PizzAndy.

An open source printer design (Voron) forms the bones of this project, but Andy “designed everything on the Z-axis” with some supplementary electronics on the back. The ‘Bitcoin orange’ filament flourishes are the chef’s kiss of the design.

Roadmap and scaling

PizzAndy talks about a modular bed tile, which means you can scale up your printer bed and mining operation “as big as you want, theoretically.” Work on this is about to start, and the creator reckons there will be 16 mining chips per tile.

The mining hash rate of the new tiles at 75 °C is still to be determined, but estimates fall roughly between 10 and 30 TH/s with Intel BZM2 ASIC chips. “At least 10 TH/s or we riot,” quips Andy. And he teases there will be a prototype “soon.”

Interestingly, there is a configuration/efficiency perk of using cryptomining ASIC arrays to warm your 3D printer bed. Andy points out that you can selectively heat up certain zones of the printer bed, as each chip/tile is individually addressable. Also, different parts of the bed can be set to different temps to suit a range/mix of materials.

It will be interesting to see if this project takes off, or at least whether the bed tech gets bought/adopted by some of the bigger names in 3D printing. In essence, the underlying concept is like some of the Bitcoin water and space heater consumer products we’ve seen before.

If you're tired of the modern internet, then why not dial back to the 1990s? One nostalgic developer has recreated the pinnacle of early Linux operating systems with the so-called CDE Time Capsule. Posted as an open-source project on GitHub under the GPL license, but accessible via its own website, the project has faithfully recreated the appearance of a Debian Linux installation, circa 1994.

For those who didn't get to experience those early Unix-based operating systems, they borrowed a lot from the -Unix part of their name. CDE, or the Common Desktop Environment, was the desktop environment used on Unix-based systems until successors like GNOME and KDE were released by the end of the decade. It was jointly developed by a number of big firms, including HP, IBM, and Sun, and first launched back in June 1993.

This new project, fully available for you to try on its own domain but available for you to run locally, seeks to capture that mid-1990s experience in full. The developer, Victor Larios, describes it as "a modern Progressive Web App that brings 1990s Unix to any device. Desktop, tablet, phone—the experience adapts. Touch gestures on mobile. Keyboard shortcuts on desktop. Always authentic."

Ambitious, indeed. It certainly looks the part, with 76 authentic color palettes and 198 original XPM backdrops used in the original release. It has a pseudo boot sequence that lets you watch as the "system initializes," after which you're pushed straight into the CDE desktop.

This isn't just a visual spectacle, as it actually works, at least partially. It has a desktop, an icon bar at the bottom, a top bar with "system info" and the time, and a workspace switcher that lets you switch between four virtual desktops. It has its own web browser (Netscape, naturally), a terminal, the XEmacs text editor, a file manager to manage your files, along with system apps to control the pseudo system processes. There are heavy customization options available, too, to let you switch up how the environment looks and feels.

Five minutes of playing around with the environment show how faithfully it appears to be replicated, but enough to show up its limitations. For obvious reasons, the Netscape browser doesn't quite work as you'd expect, with only a limited selection of hard-coded pages, with their faithful '90s designs, able to be viewed. Still, the project is quick, responsive, and (from what I can tell) bug-free, and works well on different devices, including on mobile.

While the limitations do exist, the project is well documented over at its GitHub site, with guides for both new users and power users to get the most out of it, as well as developers who might want to contribute. If you want to "experience the legend," then you can, without installing it natively, over at the specially-hosted project website, although installing it is as simple as double-clicking the 'Install PWA" button in any modern browser.

A developer has ported Doom to a pair of earbuds. That kind of statement might not have the impact it once had, after tales of high jinks and ports of Doom stretching from quantum computers, to USB chargers, to lawn mowers. However, as earbuds don’t have displays (yet), Arin Sarkisian also devised a canny method to stream the Doom action to another device, or even via the internet.

You can’t just use any earbuds for this latest Doom-on-x episode of development gymnastics. Currently, this Doom port only works on the PineBuds Pro, which Arin-S says are the only earbuds with open source firmware. And, yes, these buds are made by the same folks behind the RISC-V-powered soldering irons, like the Pinecil V2 we reviewed in Aug 2022.

In terms of improbable CPU horsepower, the PineBuds Pro are right up there. Inside their snug ear-fitting shells, alongside the audio drivers and battery cells, there is an Arm Cortex-M4F processor. Arin-S tweaked the open-source firmware to boost the CPU clock from 100 to 300 MHz (an astounding OC nowadays), and disabled the low-power mode for the most Doom-tastic experience.

Some more hurdles of using the PineBuds were the less-than-ample RAM and storage. The developer managed to get Doom to run in under 1MB by “pre-generating lookup tables, making variables const, reading const variables from flash, disabling Doom's caching system, removing unneeded variables.” Moreover, the shareware Doom 1 WAD assets file was 4.2MB, just over the 4MB storage on the PineBuds. This payload was reduced to 1.7MB after borrowing some pre-modded-for-size Doom resources.

Lastly, the no-screen conundrum presented its own set of challenges. Arin-S decided the PineBuds Pro’s UART connection was the best choice for achieving game visuals (the only other choice was Bluetooth). A mix of bandwidth, image compression shenanigans, and demands on the Cortex-M4F ultimately meant the best achievable performance was about 18fps, in practice. In theory, it should have been nearer 25fps, but the dev reckons the CPU hadn’t the grunt to keep up with converting the MJPEG stream at that rate.

No PineBuds Pro, no problem

Arin-S shares all the resources needed to get your own set of PineBuds Pro up and running, with Doom loaded. However, the developer has also generously decided to set up a website that streams Doom action from his own pair of PineBuds Pro.

Head on over to this nicely crafted webpage to join the queue (yes, it is popular) and partake in some online Doom powered by the remote earbuds.

Lastly, the developer says that he is currently looking for work if you think you might need someone with his unique set of skills.

A programming, electronics, and retro enthusiast has showcased an open-source Intel 486 motherboard that they claim was “made from scratch” in under six months. The M8SBC-486 isn’t based on existing designs, but on previous experimental work by the maker, Maniek86. This real Intel 486 CPU packing project originally began with the goal of creating a system that could run Linux and Doom. However, Maniek86 excelled themselves and noted that the system also runs various flavors of DOS, Windows 3.1 (kinda), various programs, and games like Prince of Persia and Wolfenstein 3D.

It is quite astonishing how quickly Maniek86 put together this working 486 system. Research on the project started in April last year, even though actual work is said to have begun in August, which adds a little more time to the achievement clock.

Another thing that helped Maniek86 was a relaxed attitude to compatibility. The target was ‘simply’ to be able to run Linux and Doom on the assembled machine. Providing a speedier route to this goal, the dev’s chipset (Codename Hamster 1) was implemented in an FPGA, as were some other essentials like input device controllers, CMOS RTC, and storage. PCB prototyping and manufacturing outfit PCBWay was also praised for its help and support.

Maniek86 also didn’t care too much about functions that weren’t essential to the original goal. Thus, “the secondary PIC and DMA” are missing. Check out the list below for the full specs of M8SBC-486, as of January 14, 2026.

• 150 x 150mm 4 layer PCB. Custom hole placement! (a bit smaller than the 170mm square Mini-ITX standard).
• PGA-168 socket for 5V 486 CPUs. FSB currently runs at 24 MHz, meaning that DX2 CPUs work at 48 MHz
• Currently operating at 24 MHz FSB
• Xilinx Spartan II XC2S100 FPGA as the chipset. Codename "Hamster 1"
• 4MB of SRAM
• 256KB (224KB accessible) ROM for BIOS
• 8254 Programmable Interval Timer (PIT)
• 8259 Programmable Interrupt Controller (PIC)
• Two 16-bit ISA slots
• PS/2 keyboard port. Controller is implemented in the FPGA
• Simple CMOS RTC and CMOS storage. Implemented in the FPGA too
• ATMega128 as reset circuit handler, nonvolatile CMOS storage and bitstream loader.

Since this project is open source, it might be interesting for like-minded readers to tinker with the source files and even build their own M8SBC-486 derivative design. Maniek86 admits that “There are still many issues,” but is gratified that the capabilities of this motherboard exceeded the initial goals already.

Meanwhile, the open source nature of the project beckons contributions. “I am pretty sure that this work could be used to build something more robust and stable or even to develop fully custom-made boards for other x86 CPUs,” says the maker. We’d also like to see the ISA slots become more useful, as graphics cards have poor or glitchy performance, sound cards are almost 100% incompatible, and a swathe of other cards are untested and not likely to work without ISA PnP feature support and DMA.

Hot on the heels of one of AMD's David McAfee suggesting that the company might dust off the blueprints for its Zen 3 CPUs for another production run, another exec let out a hint about AMD's future plans with FSR 4. In an interview at CES, AMD's president of GPU Technologies and Chief Software Officer Andrej Zdravković suggested that the FSR 4 Redstone technology might be open-sourced. The statements come in the wake of an accidental release of the FSR 4 code earlier this year.

The hint came during a Q&A that Tom's Hardware attended when Chips and Cheese's George Cozma asked Zdravković if an open-source release was in the cards. The software head responded that the accidental public release in August was unexpected, but that it intends to release the source for the FSR4 library while keeping the core technology closed, so as not to give Nvidia engineers an advantage.

He continued by stating that AMD intends to "work as openly as possible," and was then asked to be specific about FRS 4's open release. Zdravković then said "that's the long-term plan," seemingly corroborating an earlier remark that "open sourcing is in [AMD's heart and mind]".

While the Radeon chief's words aren't a direct statement, it's likely that FSR 4 will see an official open-source release sooner rather than later. The August leak was by way of a GitHub repository that was part of FidelityFX SDK, which itself bears a broad MIT license, except for a handful of specific files.

Many have taken this to mean that although AMD pulled the source code down, theoretically it can't put the genie back in the bottle, as all of the published data, including the core AI model and its weights (a critical piece to replicate the technology), was arguably MIT-licensed for a brief moment in time.

Keeping FSR 4 as an exclusive selling point for 9000-series Radeon cards might not really be a realistic option anymore, either. As soon as they got their hands on the source code, industrious modders quickly tweaked the algorithm to work on RDNA 3 and older cards, and even on 3000-series GeForces.

While this comes with a hefty performance penalty versus running it on contemporary cards, many found the quality-versus-performance tradeoff very much worth it, given the significantly better output of the FSR 4 model. Don't be surprised if that FSR 4 GitHub repository pops up again soon.

Leading Chinese fabless semiconductor company Rockchip has had one of its major software repositories taken down in response to a DMCA takedown notice. Specifically, its rockchip-linux / mpp repository on GitHub is currently inaccessible, after the FFmpeg team complained about Rockchip’s cavalier attitude concerning intellectual property ownership, attribution, and matters of copyright infringement. Rockchip uses this code to accelerate video de/coding on its popular SoCs. If the situation isn’t remedied, it could have broad implications for users of Linux multimedia stacks, SBC communities, Android builds, and more.

FFmpeg’s complaint

FFmpeg shares its multimedia processing toolkit under the LGPL-license, and it alleges that Rockchip has infringed its copyrighted work on the libavcodec library, which contains numerous audio and video parsers and decoders for various formats.

You might assume that this is free and open source software, so little harm has been done. However, Rockchip’s Linux MPP (Media Processing Platform) code on GitHub has wholesale copied swathes of FFmpeg code, removed any attribution, and re-licensed it under Apache, which is incompatible with FFmpeg’s LGPL license. LGPL requires the original license and attribution to be preserved. Moreover, Apache adds patent clauses that LGPLv2.1 does not permit.

Rockchip dragged its feet for years

FFmpeg appears to have been quite patient in trying to converse with Rockchip devs to iron out issues ahead of this takedown. There is evidence of chats between FFmpeg and Rockchip devs taking place on Twitter/X and GitHub since early 2024, as a video from Brodie Robertson shows (h/t Hackaday).

In this chat history, we see a Rockchip dev admitting to copying the FFmpeg code in the manner they did “due to lack of understanding” about LGPL and Apache licensing conflicts.

Also in 2024, the Chinese company grumbled about being busy, but pledged fixes would come. However, the last archived response from a Rockchip dev, from last November, stated that “there are too many chips to verify and suspend…” which reduced any hope of a friendly settlement, provoking action.

What now?

FFmpeg has faced some criticism for this DMCA takedown, which GitHub has honored. However, remember that its code has been copied and re-licensed with new authorship claimed, despite large sections of code being identical and preserving original FFmpeg dev comments.

A proposed remedy to the situation, from FFmpeg, is as follows: “remove false authorship claims; restore original attribution and copyright notices; distribute the code under an LGPL-compatible license (e.g., LGPL itself, GPL, AGPL, etc.).” The nuclear option of “remove all infringing files” is also available, as is a choice to rewrite all the code without leaning on FFmpeg sources.

If the DMCA isn’t resolved, it isn’t just an annoyance for Rockchip. Linux, Android, and SBC devs, reliant on Rockchip’s MPP for hardware-accelerated video playback, might have to fall back to doing this processing in software (with multiple negative impacts). If the downstream developers continue to use MPP, they risk losing trust, losing OS support, and facing legal risks.

The developers behind the FFmpeg project are again claiming major performance uplifts delivered by wielding the art of handwritten assembly code. With the latest patch applied, users should see a “100x speedup” in the cross-platform open-source media transcoding application. However, the developers were soon to clarify that the 100x claim applies to just a single function, “not the whole of FFmpeg.”

“The biggest speedup I've seen so far”

Last November, we reported on an FFmpeg performance boost that could speed certain operations by up to 94x. The latest handwritten assembly patch boosts the app’s ‘rangedetect8_avx512’ performance by 100x. If your modern processor doesn’t support AVX512, you should still see a 64% uplift with the rangedetect8_avx2 code path.

Where will you feel these speed increases? In some follow-up tweets, the FFmpeg developers admit that “It's a single function that's now 100x faster, not the whole of FFmpeg.” They would later go on to elaborate that the functionality, which might enjoy a 100% speed boost, depending upon your system, was “an obscure filter.”

The obscurity of the function means it hadn’t been prioritized by the devs until now. But we also gather that the filter code was recoded using the SIMD (Single Instruction, Multiple Data) processing concept for vastly improved parallel processing on today’s powerful chips.

Evidently, compilers – programs that take higher-level language code and spit out assembly (machine) code – are still not competitive with handwritten assembly. Or you could say, “register allocator sucks on compilers,” as FFmpeg tweeted today.

Assembly language evangelicals

If you are still computing in a workspace without hardware dials, Ploopy might have a solution for you. This USB desktop accessory adds an adaptable tactile rotary encoder to your input peripheral arsenal. Claimed to be both high precision and high resolution, the Ploopy Knob is now available for pre-order at CAD$49.99 ($36.75).

Ploopy differentiates its Knob from rival devices in a few compelling ways. Makers and computer freedom advocates will probably warm to the Knob’s 3D-printed, totally open-source design, but at Ploopy’s ticket price it is delivered ready to roll (scroll?). Out of the box, it offers “pixel-by-pixel scrolling precision” via its high-resolution sensor.

Thanks to its open source nature, Ploopy can comfortably claim that the Knob “can be programmed to do anything you want.” If vertical scrolling isn’t what you purchased the Knob for, it comes with the QMK firmware preloaded, which will probably a familiar feature to keyboard aficionados. Ploopy’s implementation is released under GPLv3. On the topic of open-source, all the hardware design files, including electronics and mechanical files, are released under OHL CERN v2-S.

As to what else you can do with the Ploopy Knob, there are a few helpful possibilities highlighted. Regular spreadsheet jockeys might prefer to use the Knob for horizontal scrolling. Similarly, music, media, and other timeline-based creative workflows may benefit from a horizontal scroller configuration. Another obvious use of dials like the Knob is for media control, particularly sound volume adjustment, as well as rewinding and fast forwarding clips.

High precision scrolling

How precise is the Ploopy Knob? On the firm’s blog, it is revealed that the sensor behind this device is the “formidable AS5600” position sensor. This offers 12 bits of position resolution, detecting rotational changes as small as one-tenth of a degree. In other words, it can detect your knob spinning even if you nudge it just 0.043 millimeters, or 0.00169291-inches.

We also noticed the Knob’s PCB features the RP2040 microcontroller, the brains behind the Raspberry Pi Pico W and many other devices.

Currently, high-resolution scrolling only works on Windows and Linux. On Macs, the dial functionality currently only works in discrete steps due to input device ‘smoothing.’

Acquiring a Ploopy Knob

DeepSeek AI has made its Fire-Flyer Fire System (3FS) parallel file system fully open-source this week, as part of its Open Source Week event. The disruptive AI company from China brags that 3FS can hit 7.3 TB/s aggregate read throughput in its own server data clusters, where DeepSeek has been using 3FS to organize its servers since at least 2019.

3FS is a Linux-based parallel file system designed for use in AI-HPC operations, where many data storage servers are being constantly accessed by GPU nodes for training LLMs. 3FS is unique from other file systems thanks largely to its almost singular prioritization of random read speeds above all else, and almost completely ignoring read caching.

When training AI models, compute units need to access random training data constantly, and reading this data is a one-time-only process. Therefore, a read cache is nearly useless and is largely done away with by 3FS. In fact, using the read cache when training LLMs may be potentially harmful; as LLMs are basically just super-tuned inference machines, reading the same data in the same order repeatedly has the potential to link completely different data as a set to the language model.

The team responsible for operating one of DeepSeek's deep learning clusters, Fire-Flyer 2, published this paper last August outlining using 3FS in the custom-built system. In Fire-Flyer 2, DeepSeek utilized 180 storage nodes, each loaded with 16 16TB SSDs and two 200Gbps NUCs. These nodes served 10,000 PCIe Nvidia A100 GPUs, built out in much cheaper servers than Nvidia's proprietary DGX-A100 products.

Across the whole array, DeepSeek claims it benchmarked 3FS's performance at 6.6 TB/s, while also running training tasks in the background that added an additional 1.4TB/s of read throughput. In comparison, competitor file system Ceph only reached speeds of 1.1 TB/s read throughput (on a server with 68 nodes, loaded with 10 16TB SSDs and 2 x 100 Gbps networking) for the first time in early 2024.

3FS was credited as a crucial part of DeepSeek's software stack for training DeepSeek AI in the above paper, as tested on the Fire-Flyer 2 HPC solution that achieved 80% of the performance of Nvidia's DGX-A100 server solution for 50% of the price and 60% of the power draw.

Those curious about trying out the Fire-Flyer File System and its random-read-forward style for AI-HPC solutions can find the full download on DeepSeek's Github page. We'd be surprised if this new open-source system does not become a hit for enthusiasts and enterprise AI-HPC users alike, though it may have to overcome some level of anti-Chinese tech fear to hit blockbuster status.

Byran Huang, a senior high school student from the Phillips Exeter Academy, built a fully open-source laptop as his fall term senior project, and it took him just six months to finish it. He called it anyon_e and it features a 4K AMOLED display, a Cherry MX mechanical keyboard, and around seven hours of battery life. According to Huang, he wanted to build a laptop “that hits as many qualities of a modern commercial thin & light laptop—while trying to do as much from scratch as possible.”

He used a Raspberry Pi alternative, the Rockchip RK3588 SoC, which he says is the “fastest consumer-procurable chip on the market”, built into the FriendlyElec CM3588 system-on-a-module with 16GB of LPDDR4X memory. The specifications for the SoC included a quad-core Cortex-A76 and a quad-core Cortex-A55 CPU, a Mali-G10 GPU, and an NPU that outputs 6TOPS. It could also push out 8K video at 60FPS, plus its I/O ports include an 8K display, dual USB3.1, PCIe 3.0 x4, and HDMI 2.1/eDP 1.4. We wonder if the upcoming Raspberry Pi Compute Module 5 16GB will find itself in a similar build?

For the display, Byran used a 13.3-inch 4K Samsung AMOLED ATNA33TP11 display that’s also found on several Asus laptops. Although it may seem like this screen is plug-and-play, Huang soon found out that it was difficult to drive a 4K display over eDP 1.4 because of signal integrity issues. After a few tests and calculations, he discovered that shortening the length between the CPU and display solved it, which finally allowed his screen to work.

The laptop’s keyboard was built using Cherry MX ULP mechanical switches on a separate removable case that can be used as a separate wireless keyboard. But despite this functionality (which meant the keyboard required its own 200mAh battery and SoC), he was able to maintain a height of under 7mm to keep it nice and compact. As for the mouse, anyon_e is equipped with the Azoteq PXM0057-401 glass surface and multi-touch trackpad that worked over USB; and it only cost Huang $35.

Aside from these things, he also acquired several parts like the batteries from AA Portable Power Corp. in California, and screen hinges from the Framework Laptop 13. He also used Joshua Riek’s ubuntu-rockchip kernel/distro, which made it a lot easier to have a working operating system on the anyon_e because of its great out-of-the-box experience and optimizations. However, Byran still had to build some things from scratch himself. This included the laptop’s mainboard and its CNC aluminum chassis, plus a few small 3D printed plastic structural parts to give the laptop stability.

In the end, Byran was able to create a sleek and polished laptop that could compare against other flagship devices created by Asus and Apple that cost several thousand dollars. The anyon_e laptop could even boot three seconds faster than Huang’s own MacBook Pro, which is already a tightly integrated device. And since this is an open-source project, the laptop is not hiding any secrets underneath—anyone with the knowhow, equipment, and grit to follow his process could recreate the anyon_e.

Earlier this week, an open source, super portable 3D printer dubbed Lemontron made its debut [h/t Hackaday], introducing a new entry-level DIY inverted 3D printer whose parts cost roughly $413 to assemble, mostly from existing components but also requiring a custom PCB print. When assembled, it's quite a compact printer but still has a print plate size comparable to the Prusa MK4S, a 3D printer we've covered favorably in the past.

While entry-level 3D printers can certainly be cheaper than Lemontron, they're considerably bulkier than Lemontron— though seeing as Lemontron requires access to a 3D printer to build anyway, it's more likely that this is better for use when traveling than to be someone's first or primary 3D printer anyway. Compared to other DIY options that cost upward of $699, though, this is still pretty good, particularly considering the ease of transport.

The build process of the Lemontron is pretty involved, but fortunately the debut video gives a fairly good overview of the process. There's also an in-progress written guide to help, though both may be best if you choose to undergo this project for yourself.

Some readers may notice that the 3D printer seems to be inverted— and that's actually no mistake. Following the trend of other Positon-styled 3D printers, Lemontron is designed to perform inverted 3D prints. As Lemontron (the person) explains of inverted prints in an a blog post, "the fundamental principles of 3D printing remain the same. The printer still follows the sliced model data layer by layer. The gravitational effects on the molten material are minimal because it solidifies almost instantly upon extrusion, whether it’s oriented upwards or downwards. You can even try this on any printer, go ahead and flip yours while it prints, you’ll see it doesn’t care!"

While we don't take responsibility for anyone encouraged to try that out with their own 3D printers, the principle seems sound and demonstrated working quite well with the Lemontron.

Otherwise, all we can really remark on is that it's impressive to see an effective DIY 3D printer at a (relatively) low price point for DIY 3D printers— and in such a form factor that it may actually be ideal for certain on-the-go artists and the like. While its utility as a starter or primary 3D printer is obviously questionable when parts of it need to be 3D printed to even begin with, this project is geared toward DIY-minded folk who most likely already have a 3D printer, perhaps one of the best 3D printers, anyway.

The Two Trees SK1 is a good 3D printer, though not a perfect 3D printer. It’s open source, runs Klipper, has smooth linear rails and a nice sturdy build made entirely of metal. It’s a Core XY printer in the style of the Voron Trident, with three independently driven Z screws and the ability to “Z tilt” its way to a trammed bed.

Clones are nothing new in the 3D printing world, and I had to take a moment to figure out what makes this one unique. The Two Trees SK1 is essentially a Voron Trident with a bit of Bambu Lab flavor thrown in for good measure. Its 256 x 256 build plate is the same size, if not the same shape, as a Bambu Lab, and it’s using a Bambu clone hotend. Remarkably, Two Trees does not sell either component, though they can be found on AliExpress.

The SK1 is not a beginner printer, but rather a lazy man’s Voron: a machine for someone who wants the open source freedom of Klipper, but doesn’t have the time to build on from scratch.

The printer has a few flaws that aren’t major hurdles for someone with experience. The documentation that comes with the machine stops abruptly after the printer is assembled. There are a treasure trove of video tutorials for the SK1, but they’re tucked away on the company website under “support” and not included with the printer or referenced in the manual.

I only found them after searching the internet for the SK1’s mystery nozzle. I also found a few spots where the touchscreen’s menu used the wrong language, which makes me fear there are other, more important, quality control issues. 

Retailing at $599, and currently on sale for $499, the Two Trees SK1 competes directly with a single color Bambu Lab P1P or perhaps the slightly smaller Creality K1. Its advantage over these two printers is that it is Open Source. If you want a fast Core XY printer to mod and tinker, this is it. However, the scattered documentation, noisy fans and lack of easily found parts keeps the SK1 off our list of the very best 3D printers.

Specifications: Two Trees SK1

Two Trees SK1: Included in the box

The Two Trees SK1 comes with everything you need to get your printer set up. You get tools to set up and maintain the printer, side cutters, a nozzle cleaning kit, a spare nozzle, and a USB thumb drive. You also get a small spool of white PLA to print your first model.

The included USB Flash drive has a PDF copy of the manual and test files. There are also a copy of both Cura and Prusa Slicer, with configuration files for both.

Design of the Two Trees SK1

The Two Trees SK1 is a basic open frame Core XY with a solid all metal construction. There are holes and slots left open for adding an enclosure later, but these are subtle, and the frame still manages to look finished. The branding is a low-key “Two Trees” printed on one corner. It comes with a few nice touches, like a hidden LED light and an easy to access side mounted spool holder.

It borrows from Voron’s Trident design, with three independently driven Z screws, but comes equipped with a clone of Bambu Lab’s hotend. I was initially baffled by the printer’s tiny spare nozzle, until I figured out someone had copied Bambu’s style of hotend, but made the nozzle a separate piece. The hotend is all metal and capable of handling 300 C. Two Trees doesn’t list any stats on the nozzle itself, which by its color, seems to be long lasting hardened steel.

One problem with the tool head I ran into was the shroud, which is impossible to remove without taking out the tension screw for the extruder. This gave me some difficultly when resembling, especially when I forgot to take note of how tight to reinstall it.

The printer is loaded with sensors for both leveling the bed and performing input shaping. There’s no manual leveling, but it can do a “Z tilt” where it will automatically adjust each Z axis to tram the bed on its own. You’ll still need to use a piece of paper to set the nozzle’s final Z height.

The firmware is pure Klipper, and Two Trees says it can reach speeds of 700 mm/s with an acceleration of 20,000. This is remarkably fast, but in practicality, the slicer slows it to a more reasonable 200-300mm/s. I’ve seen this kind of speed discrepancy a lot in fast 3D printers, where marketing makes bolder statements than the tech team can hold up.

The SK1 use of vanilla Klippers gives the user easy access to files and settings from a computer, and allows you to easily send files over your home Wi-Fi network. You can also send files via a USB flash drive. There is no need to talk to a Cloud or sync a phone app, which is refreshing.

Because the SK1 uses Klipper it should be simple to add a webcam to the printer for monitoring and recording prints. Directions are not included, but a quick Google should point the way.

The touch screen has a custom Two Trees interface that does what it needs too, but has a few typos where words show up in German and Chinese. It didn’t stop me from operating the printer, but did make me concerned that other quality control issues could be present.

Assembling the Two Trees SK1

The Two Trees SK1 is mostly pre-assembled and only needs to have the screen and spool holder installed before you’re ready to switch it on.

You’ll need to flip the power supply switch to match your main household electric, which is 115V in the U.S.

Leveling the Two Trees SK1

The Two Trees SK1 has built-in sensors installed and performs a 36 point leveling routine automatically. You’ll need to set the Z height with a piece of paper, but the machine walks you through this process during set up and is very simple to follow.

The bed is hard mounted, so there are no manual adjustment wheels. I didn’t have any issue with getting the Two Trees SK1 level.

Loading Filament on the Two Trees SK1

To load filament, trim the filament flat and push it through the Bowden tube near the spool holder all the way to the tool head. Give it an extra nudge until it can’t go any further. The SK1’s observation light will flash when you’re in place.

Click the horizontal lines icon on the bottom of the main menu. Press Load/Unload on the top menu. Click Load or Unload as you need. The printer will warm up and feed the filament automatically.

Preparing Files / Software for Two Trees SK1

The Two Trees SK1 comes with a copy of both Cura and PrusaSlicer, along with the start & end gcode and several config files for both programs. The directions for installing the software and creating the printing profiles are located on Two Trees website. It would be better if the directions were included with the printer, as they use a newer “drag and drop” system that is unfamiliar to beginners.

Printing on the Two Trees SK1

The Two Trees SK1 printed great right out of the box. My first print was a Benchy that I sliced myself, then I moved on to a Speed Benchy.  The small sample spool was more generous than most, but you’ll still want to buy a full sized spool right away. Check out our list of the best filaments for 3D printing here.

A Speed Benchy needs to follow certain rules: 2 walls, 3 top and bottom layers, 10 % grid infill, a .25 layer height and .5 layer width, and also our treacherous Tom’s Hardware speed settings of 200-300 on the walls and flat out on acceleration.It won’t top our list of the fastest 3D printers,, but it did place in the top five with a clean 17 minute and 13 second boat. 

This boat had some odd bumps, but the hull was beautifully sharp with no cooling issues, no ringing and nice overhangs. This is one of the best speed boats I’ve printed. Remember – when racing a 3D Printer, it's more about printing a “Benchy Shaped Object” than real quality. This was printed in ordinary gray Inland PLA, so none of the defects are hidden.

I also sliced the Maker’s Muse Calibration Castle. I skipped the sample PLA and used a roll of Creality’s Hyper Series PLA, which is formulated for high speed printing. The Castle turned out very good, with excellent bridging and slightly rough layers. This printed in one hour and 47 minutes, using a .2mm layer height and default settings around 200 mm/s.

For its feature print, I gave it something a bit tough: Chelsey Creates Things’ Chinese Dragon. This is a three-part print that required a ton of supports. Any scraps of supports left on the print are more about my slicing than the printer’s ability. 

The dragon was printed in a matte finish red/black dual color PLA from Polymaker. It may not have been the best choice, as it easily highlighted the layer lines. This part printed in 4 hours and 17 minutes using the default setting for a .2mm layer height, with a 270 mm/s inner wall and 60mm/s outer wall. 

The golden slab is Polymaker PLA Pro in Gold. This was printed in with a .16 layer height to try and reveal the details on the edge, but didn’t. This took 2 hours and 18 minutes and showed a bit of layer separation due to the default high speed. 

The “jade” pedestal is Greengate 3D recycled PETG in Emerald. This took 5 hours and 10 minutes to print using a fine .16 mm layer height and a slower 60mm/s speed. This slower print was flawless, proving that sometimes speed isn’t everything.

For TPU, I printed a spiral vase mode rose, using a chunky .28 layer height and a slow 60mm/s print speed. The flowers really worked well, with nice smooth layers. This is printed in Inland’s rainbow TPU.

Bottom Line

The Two Trees SK1 is an open-frame Core XY 3D printer running speedy Klipper firmware. Its 256 x 256mm build plate gives it a little more wiggle room than your average Ender 3 sized machine, but its AliExpress parts make finding replacements an adventure. It would be better if Two Trees stocked nozzles and hotends on their own storefront to make inevitable replacements easier to find. 

The SK1 is truly fast and capable of producing excellent quality, however it still boils down to having good slicer settings. Like a lot of fast 3D printers, the top advertising speed is strictly for travel moves, not while it’s laying down plastic. We found it to be a little faster than the Bambu machines we’ve tested, but the Creality K1 Max still beat it on the speedboat challenge. Still, it gave its best quality out when allowed to run a bit slower. 

Retailing at $599, and currently on sale for $499, the Two Trees SK1 is a good choice for someone looking for a true Klipper printer that isn’t beholden to a Cloud server or proprietary software. This machine can be tinkered with and modded to your heart’s content.

If you’re looking for a superfast budget 3D printer with Klipper pre-installed, you can also check out the Sovol SV07, retailing for $339. If you want a fully enclosed Core XY printer, and don’t mind the Creality fork of Klipper, go with the Creality K1 for $499.  

Members of the OpenWRT project, one of the legendary pioneers of open source networking firmware, are finalizing the hardware features for the 'OpenWrt One' routers. The group conducted a poll with its members and drafted hardware specs, which will be based on a MediaTek SoC (System on Chip) and Wi-Fi chip. The idea of having a router came at the same time the community announced it had been twenty years since its first open-source router firmware was released. 

Once the features and hardware are finalized, OpenWrt intends to work with Banana Pi to manufacture and distribute the routers. This would make it an easier process as OpenWrt can work on the router design and firmware while the manufacturer can seek the necessary FCC/EC/RoHS compliance and produce the hardware for distribution to retail channels. 

OpenWrt hopes that part of the revenue from these sales will be forwarded as a donation towards its Software Freedom Conservancy project and fund various other ventures. 

Proposed Hardware and Features

• SOC: MediaTek MT7981B
• Wi-Fi: MediaTek MT7976C (2x2 2.4 GHz + 3x3/2x2 + zero-wait DFS 5Ghz)
• DRAM: 1 GiB DDR4
• Flash: 128 MiB SPI NAND+ 4 MiB SPI NOR
• Ethernet: 2x RJ45 (2.5 GbE + 1 GbE)
• USB (host): USB 2.0 (Type-A port)
• USB (device, console): Holtek HT42B534-2 UART to USB (USB-C port)
• Storage: M.2 2042 for NVMe SSD (PCIe gen 2 x1)
• Buttons: 2x (reset + user)
• Mechanical switch: 1x for boot selection (recovery, regular)
• LEDs: 2x (PWM driven), 2x ETH Led (GPIO driven)
• External hardware watchdog: EM Microelectronic EM6324 (GPIO driven)
• RTC: NXP PCF8563TS (I2C) with battery backup holder(CR1220)
• Power: USB-PD 12-25V on USB-C port (optional 802.3at/af PoE via RT5400 module)
• Expansion slots: mikroBUS
• Certification: FCC/EC/RoHS compliance
• Case: PCB size is compatible to BPi-R4 and the case design can be re-used
• JTAG for main SOC: 10-pin 1.27 mm pitch (ARM JTAG/SWD)
• Antenna connectors: 3x MMCX for easy usage, assembly and durability
• Schematics: these will be publicly available (license TBD)
• GPL compliance: 3b. "Accompany it with a written offer ... to give any third
  party ... a complete machine-readable copy of the corresponding source code"
• Price: aiming for below 100$

The proposed OpenWRT One specification sees a MediaTek MT7981B SoC and an MT7976C 5GHz Wi-Fi chip, two ethernet ports (2.5G and 1G) with 1 GiB DDR4 memory, and an internal M.2 slot for a 2042 NVMe SSD. Its hardware specifications are taking shape rather nicely, but its features are reportedly not yet finalized. Some members have proposed USB-PD 12-25V and PoE and three MMCX antenna connectors. One of the project goals is to keep the price tag under $100. To keep the device secure, the OpenWrt One will use two flash chips to allow a main loader (NAND) and a (NOR) write-protected firmware recovery.

There's no official release date for such a wireless router yet. However, the process to finalize the features is still underway. Coming up with features for a wireless router within $100 is understandably no easy task. Since this is community-driven, there will be some back and forth between certain ideas and implementations before finalizing the product. Expectations are high; some members have donated to develop the OpenWrt firmware. Other members question the hardware choices, such as having an M.2 2042 NVMe storage.

The Industry's History with Open-Source Router Firmwares

OpenWrt's firmware implementation has not always been smooth sailing. In the past, companies, including Cisco, for example, have released routers whose hardware was restricted to deter its firmware from being replaced with a Linux-based alternative. Many users at the time were not happy with Cisco's decision to force its buyers to stick with its proprietary firmware, forcing them to re-consider and market the community-friendly WRT54GL router. Some were less accommodating, such as TP-Link, which blocked OpenWrt installation. Belkin-owned Linksys so far allows open-source firmware like OpenWrt and Tomato.

Once the hardware manufacturing and distribution begin, OpenWRT One will still have some challenges. OpenWrt and Banana Pi wouldn't be the first to make open-source routers. Turris is a company that makes open-source routers. A few years ago, they crowd-sourced funds for its Omnia routers.

DD-WRT sells routers with pre-flashed firmware and two license packs. Asus also developed custom-made firmware exclusive for its routers while retaining some of its in-house features like AiCloud and AiProtection. Regardless, it wouldn't have happened without the foundation put there by OpenWrt. Some would argue this is overdue, but many welcome OpenWrt's decision to branch out hardware.

OpenWrt firmware was first made for WRT54G and continued to be made for a few other routers, with some clashes with the FCC and some router manufacturers along the way. 

AMD and GPU Open released the FidelityFX Super Resolution 3 (FSR 3) source code today. AMD promised to make the FSR 3 source code available when the first games featuring the latest version of its super resolution technology were released in September. However, we’ve known about the main feature of FSR 3, its AFMF frame generation technology since November 2022.

If you are a developer, tinkerer, or just curious, the new code can be found on the FSR 3 branch of the FidelityFX-SDK on GitHub. The AMD FSR 3 integration guide PDF published by GPU Open might also be useful. The full source code for integrating FSR 3 features into DirectX 12 and Unreal Engine 5 apps, two of the most popular modern PC gaming APIs, is provided in the download. The diagram below clearly and concisely illustrates how FSR 3 interfaces with a game engine.

Of course, FSR 3 isn’t just about frame generation; it brings forth all the FSR 2 temporal upscaling technology and more. A bullet point list of the key features of FSR 3 is provided below:

• Frame generation, with interpolated frames for smoother motion
• Resolution upscaling, claimed to be near-native quality in scaling factors from 1x to 9x
• Latency reduction, delivering reduced latency with improved frame pacing
• Anti-aliasing, providing high quality AA, with FSR 3 capacle of replacing TAA in any game frame and offering a new ‘Native AA’ mode
• Optimized, for good performance on mid-to high-end GPUs and either GPU or CPU limited situations
• Cross platform, working on GPUs across brands, and multiple platforms (e.g. PC, Xbox).

Moreover, AMD and GPU Open assert that FSR 3 is easy and intuitive to integrate into a game or app, especially one that previously used FSR 2. The open source (MIT License) also facilitates easy development, with the provided DX12 and UE5 code samples.

GPU Open also highlights some important things to consider when using FSR 3, and this information is helpful to gamers so they can understand how / when to use FSR 3.

Firstly, for gamers who want to use FSR 3 for upscaling without any frame generation tech – this is OK, and it can function just like FSR 2. Secondly, it is advised to use FSR 3 technology only when it can achieve 60 fps+. As a guide, GPU Open says that if your PC can’t hit 30 fps native at the selected resolution, you should avoid using FSR 3 upscale. Lastly, gamers should use an in-game frame limiter if one is available and/or use V-Sync On.

AMD’s FSR 3 launched with just two supported games, Forspoken and Immortals of Aveum, but a newly updated list has been provided alongside this source code release (see above). Hopefully, this open-source release will also build momentum with FSR 3 adoption.

Developer Sergey Kiselev has released version 1.0.0 of his open source BIOS code for Intel 8088 devices (h/t Liliputing). After a dozen years in development the code is now available on GitHub in a milestone release, and has confirmed support for a number of interesting devices.

Intel 8088-based computers have traditionally shipped with proprietary BIOS firmware, so this open source BIOS might be a great resource for retro-modern makers and tinkerers. The list of hardware explicitly supported by this release of the BIOS includes the Micro 8088, Xi 8088, Sergey’s XT, the IBM XT, and the Book 8088.

The Book 8088 made news back in May as it was released on AliExpress alongside the Hand 386. These portables come in mini form factors and deliver retro computing on the go. The clamshell Book 8088 could be used for handheld Intel 8088-powered activities, but it is sad to hear that it was launched with a pirated copy of Kiselev’s 8088 BIOS installed. Though this BIOS has always been open source, the Book 8088 makers apparently violated the license by stripping mentions of Kiselev and GPL.

If you acquired a Book 8088 with the pirate BIOS, the latest genuine GPL-compliant v1.0.0 release might hold some appeal. A replacement plug-and-play BIOS chip is thus being sold by Kiselev’s Tindie store and it is pretty accessible at just $10. These ROMs are reprogrammable and ready for future upgrades if you have access to the necessary tools (EPROM programmer).

In addition to offering official support for the Book 8088, Kiselev’s 8088 BIOS v1.0.0 features a handful of bug fixes. Check the GitHub changelog to find out what’s new and see the fixes delivered over the history of the development – there are quite a lot.

The v1.0.0 release of the BIOS was obviously a milestone, but there are still some major works and wrinkles to iron out on the official ‘to do’ list. It looks like the biggest task for the dev is to work up some technical documentation. After that, there are a handful of bugs to tackle, addressing things like input peripherals, sound, and a few more things.

Stable Diffusion is a deep learning application that creates images from text prompts. Released in 2022, it requires considerably more computing power than a Raspberry Pi. This impressive feat, put together by maker and developer Vita Plantamura who has managed to successfully run a 1 billion parameter Stable Diffusion model on a Raspberry Pi Zero 2 W using just 260MB of RAM.

This is an impressive feat, largely because the Raspberry Pi Zero 2 W is known for its compact design rather than computing performance. But not only is it physically small, it has hardware limitations that restrict its performance. In this case, the Pi Zero 2 W has just 512MB of RAM. Stable Diffusion prefers at least 8GB of RAM for optimal performance.

Plantamura was able to make this work by developing what he calls OnnxStream. This is a specially-made inference library designed to reduce memory consumption so that it’s able to generate images on a Raspberry Pi Zero 2 W. Is it practical? Not exactly seeing as it takes much longer to create images. Images created using a VAE decoders in W8A8 precision can take three hours to generate. However, it does work and it’s quite impressive in that regard. According to Plantamura, OnnxStream consumes 55 times less memory than OnnxRuntime but is only at the most twice as slow.

OnnxStream has plenty of useful features for those looking to create with the tool. A few examples include attention slicing, both dynamic and static quantization, FP16 support and even an inference engine decoupled from the WeightsProvider. You can find a complete list of features over at the official project page.

The software is completely open source and available for anyone to explore or modify. While it can run on a Raspberry Pi Zero 2, users can also implement it on other more capable systems with the added benefit of better memory consumption. There are instructions over at the project GitHub page detailing how to set it up on Windows, Mac and Linux machines. 

If you want to learn more about this Raspberry Pi project or just get a better idea of how it goes together, check out the full details over at the OnnxStream GitHub and be sure to follow Vito Plantamura for future updates.

Have you ever been relaxing in a garden and wondered what bird is making that lovely song or what sort of feathered creature is squawking like a laser gun? Either way, this Raspberry Pi project will answer that question for you! Jumpin_jon, as they’re known over at Reddit, is using a Raspberry Pi to listen to bird calls and even recognize what bird it is using an open-source tool called BirdNET-Pi.

BirdNET-Pi is a tool designed to run on the Raspberry Pi that uses TFLite to process audio input and listen to bird calls. It was put together by Patrick McGuire, who forked the project from Stefan Kah’s BirdNET-Lite. BirdNET-Pi is optimized for the Raspberry Pi and can run on the Raspberry Pi 4B, Raspberry Pi 400, Raspberry Pi 3B, and even the Raspberry Pi Zero 2 W.

The system works by listening to bird calls using a USB microphone. The audio input is parsed through BirdNET-Pi and processed to identify what bird was likely making the sound. According to the GitHub page, BirdNET-Pi can identify hundreds of birds from different countries around the world.

Setting up this project doesn’t require much hardware; you mostly need a compatible Pi and a microphone. In this case, Jumpin_jon uses a Raspberry Pi 4 B 4GB model, but older and smaller Pis will work in its stead. Jumpin_jon used a quality mic setup and implemented an Agptek Lavalier omnidirectional mic that connects via 3.5mm using a USB adapter with the Pi.

The official BirdNET-Pi GitHub page has many details about the application, including various features you can use, like real-time effects and more. The dev recommends using the 64-bit Raspberry Pi OS Lite version—specifically Bullseye. Even though you can use other Pis, you will likely see the best performance from a Pi 4 with plenty of RAM.

If you want to make this Raspberry Pi project at home, check out the original thread shared to Reddit by Jumpin_jon and dig into the GitHub for BirdNET-Pi to get a closer look at how this system operates.

There are some problems in life that are just more fun to solve with a Raspberry Pi. Today we’ve got a clever project to share with our nautical fans created by a maker named Tommy Joad. Joad is the captain of a sailboat and has great familiarity with modern autopilot systems. However, he’s taken the bold step to create his own autopiloting system using none other than our favorite SBC, the Raspberry Pi.

According to Joad, there are several problems that he needed to overcome in order to design a more efficient system. Joad explains that many autopiloting systems are expensive and when they break, they can be pricey to fix — often requiring a specialist to repair. By creating his own autopiloting system with a Raspberry Pi, the price is cut significantly, and being the creator of the project means it’s much easier to troubleshoot.

The main system behind this Raspberry Pi autopilot project is called Pypilot. It’s an open-source application designed by Sean d’Epagnier. Pypilot is included in the Open Plotter package, this is a chart plotting system designed to work with ARM computers. In this case, it works great with the Raspberry Pi but it technically is compatible with other ARM devices.

Joad opted to use a Raspberry Pi 4 for this project but you could get away with using a Raspberry Pi 3 B if that’s all you had on hand. In addition to the Pi, Joad is using a Pypilot HAT designed just for integration with this sort of project. It also requires a motor controller that moves a belt attached to the steering wheel of the boat.

The software side of things relies on a custom OS image. Joad provides instructions on how to set up OpenPlotter using the Raspberry Pi imager. Once it’s installed, you can set up Pypilot. You’ll also need to familiarize with Signal K. This is an open source open data format intended for marine use. Altogether, these systems allow you to plot courses and keep yourself busy with other things while your boat safely pilots you where you want to go.

If you want to get a closer look at this Raspberry Pi project in action, check out the video Joad shared to YouTube and dig into his tutorial for a more in-depth look at its construction.

If you’ve dreamt of the ultimate mecha-powered future with giant robots and mecha suits controlled by the human body, you’re sure to get excited about this project from Ultimate Robots. The team has created a robotic arm that can be controlled using muscle movement thanks to their EMG signal sensor PCB, the uMyo. It also leverages one of our favorite microelectronics boards -- the Arduino.

This project was designed as a simple demonstration of what the uMyo sensor module can achieve. It’s fitted with three separate uMyo PCBs to detect movement from the wearer accurately. Each finger on the robotic arm has two tendons. These are connected to a wheel that is operated by a servo. The servo determines whether or not to curl or uncurl the fingers.

The shining gem of this creation is the uMyo sensor. It’s an open-source device designed to be worn for user input. It can transmit data wirelessly, so the wearer shouldn’t expect to be bogged down with cables tethering them to the output device. According to Ultimate Robotics, the uMyo can detect signals from various muscle groups, including arms, like in this project, legs, face muscles, and torso muscles.

Two uMyo sensors are placed at the elbow to monitor finger muscle signals. A third sensor is used at the wrist to monitor thumb muscle movement. The signals are transmitted to an Arduino, which uses an nRF24 module to receive the wireless signal. The Arduino then processes the input to send commands to the servos via a PCA9685 driver board, causing the robotic arm to move in response.

Not only is the uMyo sensor open source, but so is the software used in this robotic arm project. The team was kind enough to share everything on GitHub for anyone interested in perusing the source code.

To get a closer look at this project, check out the official uMyo breakdown uploaded by Ultimate Robotics at Hackaday. The team shared plenty of details about how it works and what goes into the PCB. You can find more information on the robotic arm on Reddit and see it in action via YouTube.

We’ve been excited to see what the Asus ROG Ally is capable of since we first reviewed the unit at launch. Today we’re taking a closer look at how well the Asus ROG Ally Extreme runs ChimeraOS, an open source Linux-based operating system based on Arch that simulates the Steam Deck OS. This comes to us from a video shared by ETA Prime, who has taken to pushing the Asus ROG Ally to its limits lately through all sorts of emulation tests.

ChimeraOS is extremely similar to Steam Deck OS and enables you to play many of your Steam games on the new handheld. If you’re looking for a best of both worlds alternative to the Steam Deck, ChimeraOS on the Asus ROG Ally seems to be an excellent compromise. ETA Prime indicates that ChimeraOS has many features you’ll find in SteamOS 3, so the interface and compatibility should be remarkably similar.

The OS loaded just fine on the first try and was responsive and provided access to ETA Prime's library through its user interface. The biggest issue was a lack of audio support. To circumvent this, ETA Prime needed to use its 3.5mm audio jack for audio output with an external speaker. The WiFi worked periodically but wasn't consistent — hopefully this can be tweaked in future upgrades. The last drawback was a lack of TDP control. That said, there are plenty of other performance settings available so the graphics settings can be tweaked as needed.

ETA Prime took the time to test several popular titles including Forza Horizon 5, Horizon Zero Dawn, and Doom Eternal at 1080p. However, the resolution did not appear as dense as it was supposedly set to during some of the tests. Other games like Left 4 Dead 2, Skyrim, and Cyberpunk 2077 tested great with video outputting around 120 Hz. It’s worth noting the Asus ROG Ally Extreme display is AMD FreeSync certified.

Overall the Asus ROG Ally appears to run ChimeraOS quite well. There are a few adjustments that might make for a better experience, such as integrated audio support, wireless reliability, and TDP control. While the Steam Deck stands out with better battery life, the Asus ROG Ally has proven to be rather versatile when running this Linux OS. If you want to see ChimeraOS in action, check out ETA Prime’s latest review over at YouTube.

If you’re ready to race, you should check out Donkey Car. This open-source Python library makes it super easy to get off the ground with setting up your own driveable car project. Because it’s Python-focused, you can use it with a wide variety of boards ranging from Raspberry Pi to even the Nvidia Jetson range of boards.

This platform makes it easy to not only to drive cars remotely but also integrate them with AI systems so they can drive themselves. You can interface with it using a separate computer or device (like a smartphone or tablet) and make the RC car as complex as you want. Long gone are the days of building your own tool from scratch, Donkey Car has plenty of ready-to-go features out of the box that you can build off of to customize your project the way you want.

What can you do with Donkey Car

• Build your own toy car that can drive itself.
• Drive your car with your phone or laptop.
• Record images, steering angles & throttles.
• Train neural net pilots to drive your car on different tracks.

Not only can you use the SBC of your choice, you can build the RC car of your dreams. It works with plenty of existing car kits you can find online but you can also make one from scratch. The team at Donkey Car recommends using their Donkey2 setup. It costs around $250 to get all of the components and can be assembled in just a couple of hours.

The standard Donkey Car kit is available on the Donkey Car website for $92. It includes a camera, servo driver, 3D printed frame components, as well as screws and jumper wires for assembly. You’ll still need an RC car component, an SBC (like the Raspberry Pi), an SD card, and a battery. While it doesn’t have everything you need, it’s a great starting point. There are also a few chassis made by the brand Exceed that work well with Donkey Car. You can read more about the hardware requirements in the build guide.

If you want to read more about this project or maybe even make one yourself, check out the official Donkey Car website. If you’ve already got an RC car project lying around, it might be worth tinkering with to see how well it handles the Donkey Car library.

The Raspberry Pi is a popular go-to board for building gaming platforms but not so often is the Raspberry Pi Pico chosen. We’ve seen some Pico handhelds before, like Pimoroni’s PicoSystem, but today’s project, created by a maker who goes by HalloSpaceBoy, has created one that you can make yourself at home called the PicoBoy.

Like many projects we love to share, the PicoBoy is completely open source from the code it uses to its custom 3D-printed housing. It’s not exactly plug-and-play. You will need to solder a few components together to complete the build. Thankfully, HalloSpaceBoy was kind enough to share a build guide complete with detailed instructions on how to put it together.

The PicoBoy console has a custom operating system and user interface. When it boots, you can cycle through a list of games on the main menu and choose the game you want to play. There is a user guide available, as well, that better explains how to use the interface. You can find replicas of classic games like Breakout, Flappy Bird, The Game of Life, and even a beta version of a Paint application.

The PicoBoy uses a custom PCB with all of the buttons needed to navigate the system put into place. It’s attached to a 1.5-inch LCD module and has a speaker for audio output. For power, it relies on three AAA batteries. The case is 3D printed and can be downloaded for free at Printables or ordered in a kit with everything you need.

In addition to the PicoBoy operating system, HalloSpaceBoy also created what they call the PicoBoy Communication Software. This is an application you run on your PC that allows you to interact with your PicoBoy to make modifications. You can also use it to add or remove games, as well as format the system. You can find more information about the software-side of the project over at the PicoBoy releases page at GitHub.

If you want to get a closer look at this Raspberry Pi project and maybe even recreate it yourself, check out the official project page over at GitHub. And be sure to follow HalloSpaceBoy for new projects as well as any updates on this one.

TL;DR How to Download YouTube Videos to PC for Free (in Up to 4K)

• YouTube videos can be legally downloaded if you pay for YouTube Premium.
• These videos cannot be used with AI voice-to-text transcription services.
• Tools like youtube-dl can be used to download content from YouTube, which can be used with AI voice-to-text transcription services.
• GUI tools like ClipGrab make youtube-dl easier to use.
• Web-based tool cnvmp3.com provides a zero install method to download YouTube videos and audio.
• Downloading copyrighted work without permission is illegal and downloading files from YouTube violates its terms of service and may lead to termination of your account.

YouTube is, by far, the most popular video streaming service in the world, hosting everything from children's songs to live streaming tech shows. To use it though, you need a constant Internet connection. But you can download YouTube videos for offline viewing. The official way is to use YouTube Premium which costs $14 per month, but there are free tools to download any YouTube video as an MP4 file, which you can transfer to any device or modify or transcribe as necessary. Note that downloading copyrighted work without permission is illegal and downloading files from YouTube violates its terms of service and may lead to termination of your account.

Below, we'll show you three different methods of downloading YouTube videos to your PC. The first method involves using a free, open-source command-line tool called YouTube-DL.

The second method involves using a freeware app called ClipGrab, which has the advantage of allowing you to choose the download resolution -- grabbing 4K videos at 4K if you want -- and file format, but it can be sluggish. The third method of downloading YouTube videos to a PC is to use a free web tool called CnvMP3. Note that our previous recommended method, using VLC player, no longer works.

How to Download YouTube Videos with YouTube-DL

These instructions are for the Windows version of YouTube-DL, but you there are are also instructions for macoS and Linux here.

1. Download youtube-dl for your OS. In windows, it’s yt-dlp.exe.

2. Copy the file into a folder that’s in your file path. In Windows, you can view your path by entering:

echo %PATH%

at the command prompt.

You can also add new folders to the path by searching for "environment variables,” and clicking Edit Environment variables. Do not put the file in C:\Windows\system32.

3. Install ffmpeg in a path folder by entering winget install ffmpeg at the command prompt from within that folder.

4. Locate and copy the YouTube URL of the video you wish to download.

5. Open a command prompt on your computer and navigate to a folder where you wish the download to appear. You can get to the downloads folder by typing:

cd %UserProfile%\Downloads

6. Enter the following command to download the YouTube video at low resolution (640 x 360).

yt-dlp [YOUTUBE URL]

If you want to download the video at its highest possible resolution (4K if YouTube has it in 4K) use the following command:

yt-dlp --merge-output-format mp4 -f "bestvideo+bestaudio" [YOUTUBE URL]

How to Download YouTube Videos With ClipGrab

ClipGrab is a free YouTube downloading app that has the advantage of letting you choose which resolution you want. The reason why this is not our top choice method is that it asks you to install unwanted extra apps such as McAfee WebAdvisor and I find out that the app could be a bit sluggish.

1. Download and install ClipGrab. You can get it from ClipGrab's site. The installer tries to get you to install other apps you may not want such as McAfee WebAdvisor. Click the Skip All button to avoid this.

2. Launch ClipGrab. If it asks you for permission to download youtube-dlp, allow it.

3. Navigate to the Downloads tab.

4. Copy and paste the YouTube URL you want to download from into the top text field.

The program will take a couple of seconds to get the list of available resolutions from YouTube.

5. Choose the file format and resolution you want from the Format and Quality fields or leave them at the defaults. If you leave the format as "original," most videos should download as MP4s.

6. Click Grab this clip! and then select a file location.

Your clip will then be listed in the Current Downloads field, along with its status.

When the download is complete, you'll find the file in the location where you set it to go.

How to Download YouTube Videos or Audio Using CNVMP3

If you want a YouTube download process that doesn't involve installing anything at all, CNVMP3 is a good choice. It also allows you to download YouTube videos at MP3 files so you can just listen to them as audio files.

1. Navigate to cnvmp3.com.

2. Paste the YouTube URL you want into the video field.

3. Select MP4 if you want to download as a video or leave it as MP3 if you want audio.

4. Select the resolution you want to download.

5. Click Convert.

The YouTube video will now download directly to your PC.

It's easy to build your own desktop PC but, if you want a phone, you usually have to settle for a sealed box that's made by one of a handful of large corporations. Maker Evan Robison wants to change all that as he posted instructions for an open-source, Raspberry Pi-powered called OURPhone with the acronym OURS standing for Open-source, Upgradable, Repairable Smartphone. 

According to Robinson, the idea was to create a smartphone alternative for people looking to control their privacy. He also wanted to make a smartphone that could be easily modified and repaired, so an open-source solution was the perfect fit. The OURphone project has quite a few specs that you’ll find on many smartphones including 4G LTE internet support, GPS support, Bluetooth and WiFi capability, as well as basic phone operations like the ability to call, text and save contacts in an address book.

However, instead of running on Android or iOS, the phone uses Raspberry Pi OS, the Linux-based native operating system for Raspberry Pis. This means that you have very fine control over what software you run on it, but the UI (as pictured) is not particularly touch or phone friendly. You can find all of the code used in the project (as well as detailed instructions) at GitHub.

In his build guide, Robinson is using a Raspberry Pi 3 B+ but there’s no reason you couldn’t upgrade it to a Pi 4. It’s accompanied by a 4G HAT with GSM and GPS antennas. It has a Waveshare touchscreen for video output and user input. A camera can be attached but it isn’t necessary for the build. A pair of headphones with a built-in microphone is used for call support. To keep the unit mobile, it operates off of a couple of 18650 batteries. 

The housing, ade out of 3mm MDF board, is a bit bulky but necessary to contain all of the hardware. It’s laser cut with port access made available all around the edges.

There are already plans in the works to add more features like a Pi upgrade, a capacitive touch screen, and a monitoring system to check the battery level. Robinson asks that anyone interested feel free to improve upon the design and share any changes they implement. He’s also made himself available at LinkedIn for anyone that has questions along the way.

If you want to get a closer look at this Raspberry Pi project, check out the official project page over at GitHub and be sure to follow Evan Robinson for future projects as well as any updates on this one.

When it comes to protecting data, you can never be too careful. While there are plenty of tools available for virtual protection, this project is focused on the physical protection of your data. BusKill is known for creating and selling computer kill cords. These trigger safety measures on your PC when unexpectedly disconnected. Today we’re sharing a development from one of their creators, Melanie Allen, who has created a 3D-printable BusKill cable that you can create at home.

This project is still technically a work in progress but enough work has been completed to issue a working prototype. According to BusKill, the project is open source and the files are available for anyone who wants to experiment with it and help BusKill improve their design. You will need a 3D printer, or at least access to a service that can print the components for you, and soldering experience to get started but there are plenty of detailed instructions to guide you through the build process.

When the team was designing this project, they had a few specific goals in mind. They wanted to ensure the case for the module was small. They also wanted to ensure it could be easily disassembled to check for tampering. With these goals in mind, Allen went to work leading to the prototype we have available today.

The project consists of a 3D printed case comprised of six individual pieces. This case was designed using OpenSCAD. There are eight magnets used to hold the unit together. A USB to USB port cord is used, as well, requiring the use of four pogo pins and receptors. Allen opted to use E360 glue to mount the magnets in place as they’re so strong, they can accidentally pop out of the case. A carabiner is also included on one end of the cable.

The OpenSCAD files and STL files are available for anyone that wants to experiment with the design. If you don’t want to build your own breakaway BusKill cable, you can always buy one directly from the website. Visit the BusKill website for more details about this project and to find detailed instructions on how to assemble the cable from scratch.

Josef Průša, the Founder of Prusa Research, dropped a bombshell on the 3D printing community Wednesday morning. After teasing a video on Twitter of a fast printing Benchy on Tuesday, he announced the long awaited Original Prusa MK4 was ready to ship. Several hundred prebuilt MK4s were in stock and ready to purchase for $1099 with a $799 DIY kit on the way. 

The announcement took Twitter by storm. Fans of the popular orange printers have suffered through a drought of upgrades, with no sign of relief despite newcomers AnkerMake and Bambu Lab appearing on the scene in 2022 with eye-popping speed. The last update to the flagship series was the MK3S+ in late 2020, which is still one of the best 3D printers. The original MK3 was released in 2017.

I was invited to chat with Průša to follow up on the launch of the MK4. “The release was awesome, better than any Kickstarter you’ve seen,” he said. Průša declined to state how many units are ready for shipping, but said thousands have been ordered. “No one was expecting it this close to shipping the XL.”

“It’s surreal for us,” he added. Průša said they’ve been using the MK4 for the past year, putting the machines through practical testing around the office and within their print farm. When asked how they’ve managed to keep the coveted new printer under wraps, he smiled and said, “we’re good at hiding.”

Průša explained that the MK4 needed to be released on the heels of the more expensive Original Prusa XL tool changer because they share a lot of hardware and firmware. He was concerned that his competition would tear into the XL looking for secrets and possibly copycat the MK4 before it could get out the door.

It’s widely known that Průša is a staunch supporter of the Open Source movement, but he has concerns. Companies like his practice open source to share ideas and innovations, helping each other grow by combining their knowledge. He wrote a blog post about his thoughts on March 29, hoping to spur communication and ultimately devise a new license for the 3D printing community.  

Prusa lamented the recent lack of innovation in 3D printing, stating that the community was more vibrant just five years ago. “Everyone is doing a rehash of the i3; no one is trying new things,” he said. He dismissed his competitor’s speedy Core XY machines as “just Voron with a cloak.”

He admits his new MK4 might look very similar to the MK3, which has become an industry standard. He explained that most of the changes are under the hood. Průša is most proud of the 32-bit STM32-powered xBuddy board which handles all the printer’s functions, including Prusa Research’s new firmware which handles Input Shaping and Pressure Advance. Other fast 3D printers, like the Biqu Hurakan, require two separate boards and an awkward internal placement to do the same job.

The xBuddy board also includes built-in Ethernet, Wi-Fi, a USB-C port and ports for an optional accelerometer and multi-material unit.

A “hidden gem” that Průša feels users will really enjoy is the new load cell sensor, which his engineers have also installed on the recently released Original Prusa XL. The analog sensor helps the printer lay down a perfect first layer automatically. It measures the distance between the nozzle and the print sheet and calculates any required Z offset. He said that other companies have been using similar technology, but not as well as they have. It requires no input from the user, and completely eliminates the need for Live Adjust Z.

To speed up prep time, the sensor only examines the portion of the bed that will be printed on. 

Customers who have pre-ordered the massive Original Prusa XL tool changer can avoid buyer’s remorse and switch to a MK4 instead. If you already have an MK3S+ you can purchase a variety of upgrade options.

“The upgrade is a nice thing to do for customers,” he said, “it will save on e-waste. And you can re-program the old board like an Arduino for making projects.” 

Průša recommends the MK3.5 upgrade as the “best bang for your buck.” This upgrade is $249 and contains the xBuddy board and a new LCD screen, giving your old printer access to Input Shaper and Pressure Advance for high speed printing. 

He doesn’t recommend the full upgrade kit, which also includes the Nextruder and new stepper motors, but felt the need to include the option because someone will ask for it. The full upgrade kit is $589, roughly half the price of a complete MK4.

Nvidia is one of the leading designers of chips used for artificial intelligence (AI) and machine learning (ML) acceleration. Therefore it is apt that it looks set to be one of the pioneers in applying AI to chip design. Today, it published a paper and blog post revealing how its AutoDMP system can accelerate modern chip floor-planning using GPU-accelerated AI/ML optimization, resulting in a 30X speedup over previous methods. 

AutoDMP is short for Automated DREAMPlace-based Macro Placement. It is designed to plug into an Electronic Design Automation (EDA) system used by chip designers, to accelerate and optimize the time-consuming process of finding optimal placements for the building blocks of processors. In one of Nvidia’s examples of AutoDMP at work, the tool leveraged its AI on the problem of determining an optimal layout of 256 RSIC-V cores, accounting for 2.7 million standard cells and 320 memory macros. AutoDMP took 3.5 hours to come up with an optimal layout on a single Nvidia DGX Station A100. 

Macro placement has a significant impact on the landscape of the chip, “directly affecting many design metrics, such as area and power consumption,” notes Nvidia. Optimizing placements is a key design task in optimizing the chip performance and efficiency, which directly affects the customer.

On the topic of how AutoDMP works, Nvidia says that its analytical placer “formulates the placement problem as a wire length optimization problem under a placement density constraint and solves it numerically.” GPU-accelerated algorithms deliver up to 30x speedup compared to previous methods of placement. Moreover, AutoDMP supports mixed-sized cells. In the top animation, you can see AutoDMP placing macros (red) and standard cells (gray) to minimize wire length in a constrained area.

We have talked about the design speed benefits of using AutoDMP, but have not yet touched upon qualitative benefits. In the figure above, you can see that compared to seven alternative existing designs for a test chip, the AutoDMP-optimized chip offers clear benefits in wire length, power, worst negative slack (WNS), and total negative slack (TNS). Results above the line are a win by AutoDMP vs the various rival designs.

AutoDMP is open source, with the code published on GitHub.

Nvidia isn’t the first chip designer to leverage AI for optimal layouts; back in February we reported on Synopsys and its DSO.ai automation tool, which has already been used for 100 commercial tape-outs. Synopsys described its solution as an “expert engineer in a box.” It added that DSO.ai was a great fit for on-trend multi-die silicon designs, and its use would free engineers from dull iterative work, so they could bend their talents towards more innovation.

A new RISC-V concept laptop design is in the works, known as the Balthazar Personal Computing Device. This laptop design is designed from the ground up to be a completely open-source laptop, that is cheap to buy and can be directly upgraded by the user themself.  The laptop is not being sold directly by its creators. Instead, it's a concept design people or companies can use to build real versions of the device.

The goal of the Balthazar laptop is to give users complete control over their computing experience with a device that is capable of using hardware and software that is built with open and secure standards and is inexpensive to make. The Balthazar laptop project wants to lead by example, showing hardware manufacturers that an open-source future is the way to go, abandoning the closed hardware architectures so many manufacturers use today.

As a result, the Balthazar laptop is one of the most unique concept designs to date and goes much further than other projects similar to it, such as the Framework laptop. From head to toe, the entire device is built with open-standards and is designed to be user-maintained and upgradeable at a very low cost (though the exact cost has not been specified). The laptop also abandons any intention of using Windows operating systems, being entirely dedicated to running Linux operating systems designed for RISC-V hardware.

Since the laptop is focused more on function than form, it is anything but sleek or aesthetically pleasing. The device features a chunky 13.3-inch form factor, with a very thick body, featuring a white, and green color aesthetic. 

If anything the laptop's design language is reminiscent of rugged business laptops such as the Dell Latitude 7330 Rugged Extreme Laptop. It features a rugged outer casing using either polycarbonate, aluminum composite or recyclable plastic, hot-swappable bays for the battery and storage drives, a waterproof Cherry MX low-profile keyboard, fanless cooling and Tempest-shielded internal cabling.

Currently, the concept design specifications include a specialized SoC based on the RISC-V and ISA CPU architectures, this SoC runs on a SoM (or System on a Module) card that is replaceable and upgradeable. RAM is also housed on this card.

For the GPU the design team is looking at using the ARM Cortex A7x, but the team is waiting on open documentation that will enable open-source drivers and software to be written for the GPU. Alternatively, the team is looking into using Nvidia GPU options, that could be used in conjunction with open-source documentation from Nvidia's side.

For storage, the concept is designed to use a SATA SSD that can be installed and taken out very quickly thanks to a user-accessible bay on the side of the laptop. Alternatively, the laptop also features an eSATA connector for external storage connectivity. Sadly you can't use modern storage solutions like M.2 NVMe drives.

Other specifications include a 13.3-inch, 16:10 full HD screen with an LED backlight, a 10,000+ mAh recyclable battery, a removable camera module and an ergonomic keyboard. For wired connectivity, it has USB ports, an OTG port, an HDMI port, a GPIO port and 3.5mm audio.

The operating systems the Balthazar laptop is designed to run include 64-bit versions of NixOS, Trisquel GNU, Guix, Debian, and Ubuntu variants for RISC-V.

When Will It Be Ready

Unfortunately, there is no due date for the Balthazar laptop design to be complete, but the good news is the team is already working on its first prototype and is actively working towards the finalized system. For more details be sure to check out the projects news feed here.

Many Tom’s Hardware readers are familiar with building PCs, customizing or assembling mechanical keyboards and even with 3D printing the odd component. Canada’s Ploopy would like to catch a ride on this adventurous tech DIY trend, and for audiophiles to enjoy headphones based on its open-source designs promising “studio monitor frequency response, out of the box.”

The amusingly named cans are available now in a number of forms from the Ploopy shop. You can get a full kit of components for CAD $150, or you can get a fully assembled set of headphones for CAD $300. There is even a headphone refresh kit with new padding parts which will be useful for maintenance. (All headphone products sold are pre-order only at the time of writing.)

Importantly, as befits an open source community driven project like this, you don’t nee­d to buy anything. Ploopy provides links to source code, documentation, and all instructions – so you can build upon the Ploopy design and firmly embrace the DIY and maker philosophy.

That will involve some patience. Several steps have to be done right the first time, and you have to be handy with a soldering iron.

So, what do you get if you make or buy these headphones? Ploopy is quite confident in its audiophile claims, and this is likely driven by the backing of its audiophile Reddit community which drove the project forward from concept to creation. It says that the planar magnetic design open-back headphones with built-in DAC and EQ can offer listeners studio quality audio. Behind this claim is the use of 3D-printed metamaterials and digital signal processing provided by the PCM3060 DAC. The DAC and RP2040 controller work in harmony to deliver up to 192kHz sampling and 24-bit audio with up to six digital filters of any kind.

Is Ploopy’s first headphones product the audio equivalent of the IBM PC standard? Probably not, but it isn’t a bad path to wander down with a 3D printer. Moreover, if you like the Ploopy’s implementation of this open source headphones project, it may be worth a look at its multiple trackball and mouse projects.

HandBrake, the popular free and open source video transcoder, has been updated to version 1.6.0. This major point upgrade is notable for facilitating AV1 video encoding for the first time in a general release. Moreover, those with Intel Quick Sync Video (QSV) enabled processors, and those with Intel Arc GPUs will be able to encode AV1 video with hardware acceleration.

HandBrake 1.6.0 can encode AV1 videos on any of its supported systems. In the current release its SVT-AV1 encoder offers the widest support, encoding on your processor through software. However, those with Intel QSV supporting CPUs or discrete Arc graphics can use the QSV-AV1 encoder for hardware accelerated processing. QSV isn't supported if your CPU is an ‘F’ suffixed model (i.e. it doesn't have an iGPU), or it is older than the Skylake generation. If you are lucky enough to have multiple QSV accelerators in your system, support for Intel Deep Link Hyper Encode should accelerate processing further. While AMD and Nvidia have AV1 encoders available for their latest GPUs, they currently aren’t integrated with HandBrake.

AV1 video is set to become the dominant codec across app-based streaming services and the wider internet, offering attractions such as; an open and royalty-free architecture, improved compression enabling efficient 8K video streaming, and support for the newest HDR standards. Developed by the Alliance for Open Media, the AV1 standard is expected to usurp the likes of H.264/AVC and HEVC, and it looks like a sure-fire winner with the support of tech giants like Amazon, Apple, ARM, Facebook, Google, Intel, Microsoft, Netflix, Nvidia, and Samsung.

Alongside the new AV1 transcoding support, the HandBrake developers have put together several 10-bit encoder profiles, and a handful of presets for dealing with typical AV1 encoding tasks. For those still interested in H.264 and H.265 encoding there are new profiles too. Meanwhile, a multitude of the app’s built in filters are updated in v1.6.0, with many of the updates implemented to support >8-bit color depths. The latest release notes on GitHub details all the above changes, as well as the updated libraries, and tweaks such as bug fixes on various platforms.

If you are interested in grabbing the latest HandBrake release to dabble in AV1 encoding, earlier versions of the transcoding utility are currently not seeing the update as available. Instead of waiting, you can head on over to the official downloads page and download and install or upgrade your existing version. HandBrake is available for Windows 10 or later, MacOS, and Linux.

Linux graphics library Mesa 3D has released an update, version 22.3.0, that adds a number of optimizations and new features to the open-source library. The biggest of these updates is support for AMD's RDNA3 graphics architecture within AMD's own Radeon Vulkan Driver. 

This will provide Linux gamers with support for AMD's latest RX 7000 series graphics cards running on the RDNA3 GPU architecture when running titles that use the Vulkan API. This support should also extend to compatibility layers such as Proton and Wine, which are designed to run Windows DirectX-based titles on Linux through Vulkan.

Along with RDNA3 support, the new Mesa 3D update also adds a boatload of other additions and optimizations: including Ray Tracing in the RADV driver, and the addition of the Radeon Raytracing Analyzer for analyzing potential bottlenecks in an application's ray tracing pipeline.

For now, we don't know which Linux distros will be updating to version 22.3.0 automatically, so you may have to update to this version manually yourself, if you want to check out the new features.

Here is the full list of patch notes:

• GL_ARB_shader_clock on llvmpipe
• VK_KHR_shader_clock on lavapipe
• Mesa-DB, the new single file cache type
• VK_EXT_attachment_feedback_loop_layout on RADV, lavapipe
• VK_KHR_global_priority on RADV
• GL_KHR_blend_equation_advanced_coherent on zink
• VK_EXT_load_store_op_none on RADV
• VK_EXT_mutable_descriptor_type on RADV
• VK_EXT_shader_atomic_float on lvp
• VK_EXT_shader_atomic_float2 on lvp
• GL_NV_shader_atomic_float on llvmpipe
• VK_EXT_image_robustness on v3dv
• VK_EXT_extended_dynamic_state3 on lavapipe
• VK_EXT_extended_dynamic_state3 on RADV & anv
• VK_EXT_pipeline_robustness on v3dv
• Mali T620 on panfrost
• Shader disk cache on Panfrost
• support for R8G8B8, B8G8R8, R16G16B16 and 64-bit vertex buffer formats
   on RADV
• initial GFX11/RDNA3 support on RADV
• various ray tracing optimizations on RADV
• extendedDynamicState2PatchControlPoints on RADV
   (VK_EXT_extended_dynamic_state2 feature)
• Radeon Raytracing Analyzer integration (using RADV_RRA_* environment
   variables)
• OpenGL 4.5 on freedreno/a6xx (up from 3.3)
• VK_EXT_mesh_shader on ANV

The concept of quadruped robots isn’t new, but it’s as cool and fascinating today as it was a few decades ago when four-legged robots first surfaced. While there have been great advancements in the field of robotics since, it’s important to continue to push forward and inspire the next generation of engineers. Petoi carries on this mission by developing products like Bittle that teaches young robot enthusiasts the mechanics of building and programming their inventions.

Bittle is a programmable robot dog targeted at robotic beginners ages 14 years and up, or anyone who wants to have fun learning and playing with robots. Priced at $339 (assembled), Bittle comes programmed with a few starter tricks like walk and trot that learners can try out to familiarize them with what the robot can do, and then later expand to more complex types of behavior. This may seem like an expensive STEM kit, especially in today’s economy, but it’s fairly affordable compared to similar quadruped products on the market right now. In fact, some servo-based quadruped robots like the Xgo Mini or the PuppyPi retail for closer to a thousand dollars. Petoi can keep its production costs down thanks to its open hardware framework and open-source software platform.

Bittle’s open-source platform also allows you to add a Raspberry Pi or attach Grove sensors to extend its capabilities, explore AI machine learning projects, or try out various STEM experiments. These add-ons, however, are not included in the basic package, so you’ll need to purchase them separately. Unfortunately, Raspberry Pis remains in short supply, so you might be hard-pressed to find a good deal on it, much less find one in stock.

Set Up and How it Works

Bittle ships in two models, a pre-assembled kit and a DIY construction kit. It comes with three color options: black and yellow, blue and yellow or red and yellow.

 Bittle Basic Kit 

Our Bittle sample unit came pre-assembled with the black and yellow color combo. It required minimal setup since the main body and legs were already put together. All that was needed was to snap the head in (which also came pre-assembled), add the tail, charge the batteries and Bittle was good to go out of the box. Almost everyone could enjoy this model, even younger makers (kids under 14), who just want to play with Bittle as a regular pet toy and get some hands-on experience on movement manipulation or adding on to its built-in routines.

There are three ways you can control your Bittle: using the IR remote control that comes with the kit, downloading the mobile app on your phone, or installing the desktop app on your computer.

The pre-assembled robot also came pre-calibrated and preloaded with a set of basic tricks like walk, sit, stand and trot, which you can control to speed up or slow down using the controller of your choice. It also had some more fun tricks that my 9-year-old was excited to try right away, like “Say Hi,” “Pee,” “Do pushups” and “Play Dead.” 

I could see my daughter's eyes light up with how naturally Bittle could move and trot around. It was very agile and maintained good balance even as she changed the direction of its movements from left to right and back and forth on our living room floor. However, Bittle wasn’t as graceful moving on the carpet as it was on hardwood or other smooth surfaces. 

She thoroughly enjoyed manipulating the speed for each movement as well, which led to discovering a cool feature — Bittle could actually flip itself back to a standing position if it tripped up or somehow fell due to an obstacle in its path. She was amazed at this fail-safe maneuver because it was autonomous and looked like something a real pet would do.

Standing six inches tall and weighing less than two pounds, Bittle is the perfect size for her to pick up and place in different play settings. However, she wasn’t a fan of its head constantly falling off. Though it was easy enough to snap back in place each time it fell, after a while, she found herself avoiding certain movements because she did not want to ‘hurt’ Bittle. Oddly enough, it turned out to be an actual feature of the robot rather than a product defect. The head also acts as a clip that can hold onto tiny objects and is also where the add-on sensors are placed for exploring fun STEM projects.

Bittle’s short battery life quickly became an issue. As it turns out, an hour isn’t enough time for play and exploration at all. 😊 Once the warning light turned on and Bittle slowed down or stopped, my daughter knew it was time to plug him back in for charging. It might be a good idea to purchase extra batteries (sold separately or as an add-on) so you can just switch them out and limit the disruption to the flow of fun and learning.

Bittle Construction Kit

Bittle also comes in a do-it-yourself construction kit model, which ships unassembled so you can have the experience of putting together the entire robot yourself — a worthy and most enjoyable activity for anyone interested in robotics. You’ll also learn circuitry and cable management. It's even $10 cheaper than the one that comes prebuilt.

The main body frame and legs are made of hard, durable plastic material. It’s cleverly designed with an interlocking mechanism, so parts easily snap in place. Assembly videos and tutorials are available online if you need instruction or help. There are also diagrams that you can refer to in the online manual. When you assemble Bittle yourself, you’ll need to upload code to the NyBoard and make sure your connectors are connected to the right circuit.

After calibration, then you can attach the flat-end screws. The kit comes with a mini screwdriver, but to speed up the build process, you can use an electric screwdriver if you have one handy. It should take about 40 minutes to assemble, calibrate and upload the software.

I would say that although our review unit did not initially require assembly, we got the experience of disassembling Bittle’s legs and putting them back together in an effort to troubleshoot some calibration issues that came up after extended time testing Bittle’s maneuvering capabilities.

One thing we appreciated was the time and support we received from Petoi in this process. They were quick to respond and very helpful when we reached out with issues. First, we needed to break down and unscrew all the legs, remove the servos from their sockets and then put everything back in place. After that we had to go through manual calibration of each servo motor. The calibration joint tuner included in the kit came in handy, but the provided screwdriver was hard to use. We recommend using a different one if you have one in your toolbox, or even using an electric screwdriver if you have one. Honestly, it took some time to get Bittle calibrated properly. There was a lot of trial and error involved, and we had to redo the process a few times before we got Bittle back to an acceptable workable state.

Joint Calibration Tip: Try to center the holes when aligning the corresponding holes on the calibration tuner tool. 

Bittle Hardware

Servos

Bittle uses nine customized PS1 servos: two on each leg that act as joints and one for the head. The kits ship with an extra servo as a spare.

Note: When building the robot, it’s important to make sure you position the servos correctly to ensure the legs will work properly. The direction of the motors is important, and the cables must be properly connected to the NyBoard to ensure robot operations run smoothly.

NyBoard

Bittle is based on a customized Arduino board that comes pre-programmed with a dozen neat tricks or routines that the robot can execute (for the assembled model). Our review model came with the first version of the board, though Petoi has come out with version 2 already.

Bittle Adapters / Connectors

Bittle comes with three adaptors that you can use to charge, code, calibrate or load new firmware to the robot.

• USB
• WIFI
• Bluetooth

Bittle Controllers / Apps

There are three ways you can control Bittle:

IR Remote Control

Bittle kits ship with an IR remote control (battery not included). Each button shows an icon that corresponds to a pre-programmed maneuver that you can try out. Our remote control stopped working after some time, even after we added a new battery.

Mobile App

You can install the Petoi Robot Controller App on iOS or Android and connect to Bittle via Bluetooth. From here, you can calibrate each servo, use the main control pad to engage with Bittle, or upload new capabilities.

Desktop App

Use the Petoi Desktop App for new firmware uploads, joint calibration and composing new skills. It’s compatible with Windows PC or Mac. You will need the Micro-USB dongle to connect to Bittle through the USB adapter. We found that this is not the most reliable method because there are times when the app could not find a port and would show errors. We preferred connecting to Bittle via Bluetooth.

Programming Bittle

To start coding Bittle, there are two applications available: using Codecraft (Petoi’s custom Scratch-based app) and the Petoi Desktop App.

Codecraft

To use Codecraft to program Bittle, you must make sure you have OpenCat 1.0, as version 2.0 is not yet compatible with Codecraft. First, you need to connect to your Bittle to check what version of OpenCat is currently installed. Use the USB adapter and dongle, WiFi or Bluetooth to connect, then run the Petoi Desktop App. In the main menu window, click on Firmware Uploader. The next window will show you what version of OpenCat you have. You can easily change between versions using the pulldown menu option available. Just choose the version you need, click the Upload button, and follow the prompts provided.

You can download the Codecraft app on your computer or go online to https://ide.tinkergen.com/ and select Bittle to start experimenting and programming different routines for Bittle. When done, click the Upload button on the left side menu. It takes a few seconds before Bittle executes your script. My daughter, who is already familiar with Scratch from use in school, dove right into Codecraft. She particularly enjoyed seeing Bittle trot to the tune of Jingle Bells and combining multiple skills and looping them. Unfortunately, we did not have any of the Grove modules available, but she expressed the desire to use those in the future.

In addition to needing to downgrade your OpenCat version to use Codecraft and use the USB adapter to connect to Bittle, you may experience issues with port detection. We often had to close and re-open the app because the software kept throwing connection errors. Thank goodness for the Bluetooth option!

Skill Composer

To program new skills for Bittle using the Desktop App, connect to Bittle, run the Desktop App, and Click Skill Composer. If you downgraded to the 1.0 version of OpenCat for Codecraft, we recommend that you now upgrade or go back to version 2.0 or the latest version of OpenCat available before you use the Skill Editor. We had issues when using the older version for the Desktop App.

Under Skill Editor, click on the Add button to add another line and choose one of the Preset Postures. You can use the Preset Postures as-is or adjust the movable sliders (in yellow) on the left-hand side to customize its movements. Save and hit Play to run the script. You can add a loop or program the number of times you want the script to repeat.

We experienced the same port detection errors when using the USB connection to use Skill Composer and had to re-try a few times before the software found the available port.

There are 16 online courses available for beginners and advanced learners. Bittle also supports Python and C programming languages.

Bottom Line

Who wouldn’t want a pet robot like Bittle? It’s cool, it’s fun and it’s entertaining. Moreover, it’s an amazing piece of tech that will keep young enthusiasts and hobbyists engaged while introducing them to the basic concepts of robotics. Granted, Bittle isn’t the friendliest looking pet robot out there — my daughter mentioned Bittle reminded her of a guard dog, and it reminded me of Alpha, the Doberman Pinscher from the movie Up. But even that could be a great challenge for young makers because they can dig into their own creativity to come up with fresh ways to mimic real-life dog interactions. One idea we had was to play around with its head movements to inject more character and personality into Bittle — there are lots of possibilities to explore through the Skill Composer.

In terms of its value, compared to Xgo Mini ($799) or PuppyPi ($423-$940), Bittle at $339 is in the lower price range and is currently the more affordable option. While that's a steep price to pay if you or your child are only interested in using Bittle as a toy for entertainment, the true value of Bittle comes in the education it brings. The hands-on coding experience from the apps, free online coding courses, and further expansion projects and experiments definitely make Bittle a worthwhile investment. Not to mention the support of the open-source online community.

While Bittle's out-of-the-box agility was highly impressive, it’s important to reiterate the calibration process was not easy and Bittle’s short battery life might leave young learners hanging. However, like many STEM robot kits, Petoi's Bittle is designed to provide an entryway to robotics and AI, which was certainly achieved during our time with this robot. Overall, the technical hiccups are outweighed by the educational experience and the fun Bittle provides when operating smoothly. Perhaps the next model will offer upgrades to correct the issues we found in our testing.

SiFive announced a pair of new high-performance RISC-V processors aimed at what it calls "next-generation wearables and smart consumer devices." Known as the P670 and P470, the processors offer new features and improved performance compared to previous CPUs based on the popular open-source architecture.

There's support for virtualization, including a separate IOMMU for accelerating virtualized device I/O, and a full out-of-order vector implementation based on the RISC-V Vector v1.0 spec that was ratified last year. The chips claim to be the first on the market to support the new RISC-V vector cryptography extensions. They also exhibit enhanced scalability, with clusters of up to 16 cores able to work together, though the company has talked up 128 cores in the past.

SiFive’s 600-series is performance-focused — the P670’s predecessor the P650 was expected to match the Arm Cortex A-77 for performance — while the 400-series is more of an efficiency chip. The P550 was the chip of choice for Intel’s Horse Creek development board, which paired RISC-V with DDR5 RAM and PCIe 5.0. Intel attempted to buy SiFive for $2 billion this time last year, but the deal fell apart after the companies failed to agree terms.

The P670 is built on a 5nm process and can achieve a maximum clockspeed of more than 3.4GHz. SiFive’s description makes it sound like a P650 with two added vector units, and the new chip aims to beat Arm designs in the performance-per-mm metric that measures the space a chip takes up against its processing capability.

The P470 is designed as a companion chip for the P670, and we expect it to appear in big.LITTLE configurations in exactly the same way Arm chips do in smartphones and tablets. It’s also designed on a 5nm node and hits the same clockspeed as its bigger brother. A P450 also exists, which is a P470 without its single vector unit, and it has been area-optimized to fit in particularly small places.

“The P670 and P470 are specifically designed for and capable of handling the most demanding workloads for wearables and other advanced consumer applications. These new products offer powerful performance and compute density for companies looking to upgrade from legacy ISAs,” said Chris Jones, SiFive VP of product. “We have optimized these new RISC-V Vector enabled products to deliver the performance and efficiency improvements the industry has long been asking for.”

There's no news of when the P670 and P470 will be available, nor which companies will be using them.

Carrier boards are important for users of the Raspberry Pi Compute Module 4, because in its natural state it lacks any I/O options beyond onboard wireless, and some boards don’t even have that. If you want USB, or PCIe, or even some storage beyond any on-board flash you may have specified, you need to connect to a carrier board. Raspberry Pi offer its own Carrier Board, which breaks out everything you can think of. It’s unlikely, however, that you’ll use every connector on a board, so a new release from Mirko Electronics on GitHub, the PicoBerry, adds just a USB-C port and GPIO pins to the module, for minimalist projects.

You’ll need one of the CM4 variants with some on-board eMMC storage to get any use out of the PicoBerry, as you’re not going to be able to add any storage using the carrier board. The USB-C port is only for power, like the Raspberry Pi 4, so all the I/O for your project is going to go through the GPIO array, which has 40 pins in the standard Raspberry Pi style. 

This means you can connect the CM4/PicoBerry combo to things like LCD screens, HATs, audio DAC boards, and whatever else you can think of that uses SPI, DPI, RS232/485, or any of the other interfaces supported on the Pi’s GPIO. Measuring just 20 x 70mm (0.8x2.75in)  the board takes up very little space, and might be suitable for fitting in a tight casing, assuming you don’t need extras like USB ports. If you do need these, you could add something like the USB and Ethernet HAT available from The Pi Hut, but now you’re adding two boards where a single more capable one might do. CM4 compatible boards, like BigTreeTech's CB1, may be able to use the board too, but the CB1 comes with its own carrier, which turns it into a fully fledged RP4-a-like, so again why bother?

There are two user-addressable LEDs on-board, one red and the other green, plus activity and power lights, but that’s really all you get. The PicoBerry is being released as open-source hardware, and there are manufacturing and source files available under the CERN-OHL-1.2 license at Mirko Electronics’ GitHub page. The boards aren’t produced for retail sale, so you’ll need to know what you’re doing to create one.

RISC-V processors have been gaining traction and software support for some time, so it’s good to see the open-source architecture making its way into actual products. The Alibaba Roma RISC-V laptop, announced back in the summer and spotted by CNX Software, is finally available, and contains a quad-core processor plus plenty of the features we’ve become used to from Intel and AMD computers.

The Roma is based on a computing platform known as Wujian 600 aimed at cost-effective edge computing. The CPU in the laptop is an Alibaba T-Head TH1520 quad-core Xuantie C910 processor that’s clocked at up to 2.5GHz with a 4 TOPS NPU and an Imagination Technologies GPU on the side. It can support up to 16GB LPDDR4 or 4X RAM at up to 4266 MT/s, and has 256GB of SSD storage. 

The display is a 14.1-inch 1080p panel, a resolution matched by the webcam. Should that not be enough screen real e state, there's an HDMI port for hooking up an external monitor. For networking, there's Wi-Fi 5 (yes, 5) and an Ethernet port. There's also Bluetooth 5, for headphones and other peripherals.

You can charge over USB Type-C port, and there are a couple of Type-A ports, too. Battery life is rated at 10 hours. There's a dedicated security chip onboard, an Arm SC300 Cortex-M3 security enclave processor with Trusted Execution Environment security certification, and the laptop runs Alibaba’s own Linux-based OS OpenAnolis.

The keyboard is backlit, and at only 20 mm (0.7 inches) thick and weighing 1.7 kg (3.7 pounds), it’s definitely at the thinner and lighter end of the market, even if it can’t take on the featherlike HP Pavilion Aero. 

The Roma is available from Alibaba as a basic package for $1,499, which comes with a warranty of ‘more than’ five years and free spare parts. There’s also a $4,999 ‘premium’ package that sees extras like headphones and a smartwatch added to the deal, along with the chance to have your name engraved on the laptop’s casing.

The Roma is available for pre-order from Alibaba, in black or gray, and the first 100 orders aim to deliver in Q4 2022, while a further 1,000 will arrive in Q1 2023.

Google and the U.S. Department of Commerce’s National Institute of Standards and Technology (NIST) on Wednesday signed a joint research and development pact under which Google will finance production of open-source chips that could be used by academic and small business researchers to build a variety of emerging applications.

The majority of newly developed devices require chips to enable their core or 'smart' functionality. But startups usually formed in universities and small businesses typically do not have money to develop their own chips, order custom chips from design houses, or procure large quantities of of-the-shelf components. But scholars and small companies tend to have some bright ideas when it comes to emerging applications, which is why to NIST decided to team up with Google to help them turn their ideas into products.

From Lab to Factory, Quickly

Under the terms of the deal, NIST and its university research partners* will design 40 open-source chips optimized for different applications, while Google will finance initial cost of setting up production and will fund the first production run. Those chips as well as their open-source designs will be available to academic and small business researchers who work on various projects. Some of those chips could be used to power various innovative devices without paying licensing fees, whereas other could be customized further to build ICs aimed at a specific application. 

The key idea of NIST's initiative is to provide essential building blocks to interested parties in a bit to speed up prototyping and then product integration, thus cutting down time-to-market for new products. Startups creating the most promising products could end up being acquired by Google or other high-tech giants.

NIST does not specify what kind of open-source chips it plans to design, but it says that it is looking to address new kinds of memory devices, nanosensors, bioelectronics, and advanced devices needed for artificial intelligence and quantum computing.

"Moving to an open-source framework fosters reproducibility, which helps researchers from public and private institutions iterate on each other's work," said Will Grannis, CEO of Google Public Sector. It also democratizes innovation in nanotechnology and semiconductor research. 

200-mm Wafers, 130nm Node

The open-source ICs designed by NIST and its partners will be produced by SkyWater Technology at its 200-mm fab in Bloomington, Minnesota. 

For now, NIST and Google do not mention which process technologies will be used, but SkyWater and Google already have an open-source 130nm process design kit (PDK) that can be used to design new chips or customize other open-source designs. Meanwhile, SkyWater offers 90nm and 130nm process technologies as well as has a qualified 65nm production node. 

Interestingly, SkyWater will supply those open-source chips on wafers, which will require universities and other purchasers to dice them into individual dies, assemble them into appropriate packages, and test the final chips at their own processing facilities.

*University of Michigan, the University of Maryland, George Washington University, Brown University, and Carnegie Mellon University.

AMD has announced the immediate availability of version 2.1 of its popular FidelityFX Super Resolution (FSR) algorithm on GPUOpen, bringing an even bigger knife to the fight against Nvidia's proprietary DLSS technology. Version 2.1 of FSR delivers a number of performance and visual quality improvements to the FSR 2.0 package, with improvements to some of its more glaring issues, such as ghosting and shimmering - artifacts of temporal-based solutions that look at previous frames in order to reconstruct the ones that come after.

The new update is already being distributed in the latest version of "Farming Simulator," released early this week - and AMD shared a couple of screenshots that showcase the differences in implementation. The improvements to ghosting - one of the more prevalent issues in the new age of temporal supersampling - are readily apparent.

There are a number of changes under the hood, but all of them appear to deal with increased stability of temporal images, less artifacting and even improved image quality - specifically, the release notes indicate that AMD has turned some half-precision computations to full precision ones. By itself, this change has the potential to reduce performance, but AMD didn't mention any performance reduction as part of the new update, so it's likely other areas saw performance improvements that could outweigh the penalty. Remember that in graphics technology, image quality improvements always impact the performance equation. Quashing some bugs, as AMD also says it did, could be a way to recoup more performance still.

AMD is clearly committed to its FSR support, which saw a relative defeat in its first iteration when compared to the more mature and computationally-heavy DLSS. But AMD has persevered, and its continuous improvements to the algorithm showcase that it's up to the task of maintaining its software solutions - especially amidst the expected competition from Intel, which is gearing ever closer to releasing its Arc Alchemist graphics cards with their own upscaling algorithm, XeSS (Xe Super Sampling).

AMD's release includes an update to its FSR 2 Unreal engine plugins - an extremely welcome tool for Unreal Engine developers to streamline their FSR integration. According to AMD, changes from FSR 2.0 to 2.1 should be easy for game developers to implement - a potential boon for the 45 available or upcoming games that make use of AMD's open-source and vendor agnostic upscale technology.

As part of its release of FSR 2.1, AMD also announced a cadre of games that will feature support for its latest supersampling tech:

• Choo-Choo Charles
• Cyberpunk 2077
• Deep Rock Galactic
• Destroy All Humans! 2 – Reprobed
• Miasma Chronicles
• No One Survived
• Project Haven
• Red Dead Redemption 2
• Scathe
• Scorn
• VEIN.

FSR support on "Red Dead Redemption 2' has certainly been a long time coming for AMD gamers. If you want to see the new FSR 2.1 algorithm at work, AMD has provided a comparison video that should make improvements clearer (especially when it comes to shimmering, which is nigh impossible to capture in a still frame).

If you can’t find the device you want, there’s no reason you can’t make it yourself from scratch—at least, that seems to be the trend in the Raspberry Pi community. Today we’ve got a project to share from a maker named Anil, aka one_free_man_ as he’s known as over at Reddit, who’s created a portable handheld device known as the Octapod.

The Octapod is built around a Raspberry Pi Zero 2 W and runs DietPi, a low resource Linux operating system designed to create appliances. Backed up with a touchscreen interface, it works as a custom music player with the help of an application called Lollypop. Anil has programmed the Octapod so that it boots straight into the music-playing interface so no extra steps are necessary to load up your favorite songs.

In addition to the touchscreen, there are also a few buttons that can be used to control settings. A power button is used to safely turn the Pi on and off while a sync button is included to wake up the device, launch services, and close the music application.

As mentioned above, it’s built around a Raspberry Pi Zero 2 W and also uses a 2.8-inch Waveshare touchscreen along with a 3500mah battery as well as a 512GB micro SD card for storage. Holding everything together is a case made by Anil that he constructed using epoxy, rather than something laser cut or 3D printed.

Like most Pi projects shared with the community, this one is totally open source. Anil has made the source code available for anyone who wants to either make their own or just see how this one goes together. You can explore the code over at the original project thread at Reddit.

If you want to recreate this Raspberry Pi project, check out the full post and look for comments posted by Anil to get an idea of what hardware you’ll need and how to go about programming the handheld. Be sure to follow Anil, or one_free_man_, for more cool projects and any future updates on this one.

Finding a stockist of Rockchip-powered boards just became a bit easier, as UK based (but worldwide-delivering) online store OKdo, part of the RS Group, has announced via eeNews Europe, a partnership with Chinese open source hardware specialist Radxa which will see Raspberry Pi alternative, Rockchip powered boards manufactured and distributed under the Rock moniker.

The news comes less than a week after news broke of RS Group's decade long Raspberry Pi license came to an end.

Co-Founder and CTO of OKdo, Richard Curtin, said: “We’ve listened to our customers and understand the challenges that the current supply chain dynamic creates. Our mission is to solve customer problems, and we’re delighted to announce our partnership with Radxa which allows us to provide leading technology to our customers worldwide."

The process starts with the Rock 4 SE, a £61.13 ($72.21) version of Radxa’s Rock 4C+ that’s been tweaked to reduce its cost. The RK3399-T CPU sports six Arm cores, split between two Cortex A72 and four Cortex A53, and adds a Mali T860MP4 GPU plus 4GB of RAM. There's a tiny Micro SD slot, and an undetermined amount of eMMC can be specified. There's also an M.2 slot for an NVMe SSD.

Expansion comes via four USB ports, two 2.0 and two 3.0, a full-size 4K HDMI port, 3.5mm audio out, plus Wi-Fi and Bluetooth. Power is delivered over USB-C, and there's a 40-pin GPIO array, said to be “extensively compatible” with SBC accessories.  What remains to be seen is how extensive the compatibility with Raspberry Pi accessories is. It is common to see boards being electrically compatible but software support is often lacking.

Volume purchasers can customize the board by adding or removing components (what you can add isn’t detailed, though a full-size PCIe 4.0 x16 slot down the middle seems unlikely) to tailor the board to their precise needs. This has been common for industrial Raspberry Pi applications, such as EV charging stations.

Operating system support comes in the form of Debian or Ubuntu Linux and Android from 7.1 to 11. GPU-enabled AI stacks such as Caffe can be used, and there's a hardware access and control library for Linux and Android.

At the time of writing, the Rock 4 SE is out of stock on the OKdo site, but this could just be early teething trouble, as the site says it’s ‘coming very soon’. “Tackling the shortage of semiconductors on the market, OKdo, in collaboration with Radxa, ensured long-term stock availability of ROCK 4 SE, sparing you from stock hunting,” reads the ‘Availability’ section of the product page.

Framework is changing the shape of the power button on its open source, 3D-printed cases for its mainboards after it received a letter from Lenovo's legal team suggesting the button looks too much like Lenovo's Legion "O" trademark, Framework tweeted today.

The clip of the letter can be seen in the tweet below, which you may have to expand.

The letter, directed to Framework Computer CEO Nirav Patel, is only shown in a partial screenshot and is on letterhead from Lenovo's lawyers at Taft Stettinius & Hollister LLP. "We believe that the Broken O Case" may infringe Lenovo's trademark rights in the LEGION Trademarks, and may ultimately lead to consumer confusion," it reads in part, further suggesting that Framework's case or mainboard may appear to be affiliated with Lenovo. While we can't see the rest of the letter's contents, it may be a cease and desist letter.

Framework doesn't seem interested in fighting the challenge. Instead, it's starting a community contest for the new design. Entries must be submitted by 11:59 Pacific Time on Aug. 25, with a winner "judged subjectively by our CEO"  on Aug. 26.

No date is shown on the snippet of the letter that Framework tweeted, so it's unclear when Lenovo's lawyers sent it. Framework first started offering the 3D printing schematics in April when it began selling mainboards separately from its computers. It doesn't appear that Framework ever directly sold the case in its marketplace.

Lenovo did not respond to a request for comment in time for publication. We'll update this story if it responds. Framework acknowledged a request but didn't return with a statement in time for publication.

Lenovo's Legion lineup consists of the company's premium gaming laptops, desktops and monitors. Some lower-end laptops are also under Lenovo's Ideapad branding. Lenovo's "O" design is typically used within the word "Legion" on its products, though previous versions of the design were used alone. Framework doesn't currently offer any products other than its Framework Laptop and components for it and has yet to step into gaming.

The "O" shaped power button on the 3D case isn't on the Framework Laptop, so its flagship product should be unaffected.

Linux support with webcams has reportedly become a major issue in regards to 12th Gen Alder Lake laptops right now, as showcased by a recent Phoronix article. Linux kernel maintainer, Greh Kroah-Hartman, says causal users should steer clear of Alder Lake laptops using Linux if they want to make video calls on their machines. Issues surrounding mainline kernel support for webcams have made webcam support almost nonexistent.

Hartman says - ironically, that Linux support with Alder Lake has been really good. But for some reason, webcams are the only exception. The lack of mainline kernel support has forced manufacturers to create proprietary drivers for their webcams.

The primary example of this is the new MIPI IPU6 webcams found in Alder Lake laptops like the Dell XPS 13 and Lenovo ThinkPad X1 Carbon. The Linux driver supporting IPU6 is completely proprietary and only supports specific kernel versions of Linux. That makes Linux kernel updates impossible, unless you want bricked webcam functionality.

This is just one example of limited Linux webcam support, with Hartman himself commenting that you shouldn't "...buy these notebooks if the vendor isn't willing to get their drivers upstreamed properly."

The main issue is that IPU6 drivers don't make use of the Video 4 Linux 2 interface, which is a common API used by webcams drivers for Linux. Instead, the IPU6 drivers rely on a different API, since the camera is more complex compared to other solutions, and vendors don't want to share any of their imaging algorithms with the public.

Thankfully, Phoronix says there is a plan for upstreaming Intel's IPU6 driver to the CAM kernel API, which appears to be a new open-source webcam API for Linux. But that work is still far and away out, with a timeline of at least a year of development - if not two years.

Basically, it appears that most of the 12th Gen laptop models sport this new MIPI IPU6 webcam, and as a result, can suffer from Linux compatibility issues due to mediocre support via proprietary drivers. 

If you manage to find an Alder Lake laptop with a different webcam, there's a good chance it won't have these issues. You could always attach one of the best external webcams and it will probably have Linux support. Or, alternatively, you could have Windows as your host OS and run Linux in a virtual environment instead

Earlier this week, Google and GlobalFoundries (GloFo, or GF) joined forces to push forward the former’s open silicon initiative. In addition to announcing the ‘milestone’ open-source chip-making venture, the partners released their Process Design Kit (PDK) for the GF 180MCU technology platform. The heady brew of open-source ingenuity and foundry access hopes to spark creativity and innovation in the hardware space.

Given the GF 180MCU codename, the PDK and the associated project provide access to manufacturing using GloFo’s 180nm process technology. In this way, the partners are pioneering the provision of open-source access to affordable high-volume manufacturing to a broader clientele.

As PC enthusiasts reading about tech heavyweights advancing from single-digit nanometer process chips into the angstrom era, it might be easy to sneer at the GF180 process. However, Google expects 180nm to continue to be popular for chips like motor controllers, general purpose MCUs and PMIC, IoT Sensors, Dual Frequency RFID, and Motor Drives. Furthermore, the project should precipitate innovation and confirmation that the open source model for the foundry ecosystem is viable.

Leading up to the announcement, Google spent two years working with SkyWater Technologies after releasing one of their PDKs under the Apache 2.0 license. Google completed six shuttle runs during this time, filtering through 350 unique chip designs from the community into 240 projects that were manufactured “at no cost.”

Post-pandemic, Google reckons we are moving into a “New Golden Age” of semiconductors, with diverse applications such as mobile, IoT, and automotive all growing strongly. Moreover, its open-source chipmaking venture with GloFo will help give this movement critical momentum.

GloFo is currently the fourth largest chip foundry and manufactures about 6% of the world’s semiconductors. It can output 16+ million wafers at 180nm yearly and plans to increase capacity to 22+ million wafers by 2026.

Google ended its blog post by telling software developers and hardware engineers, researchers and undergrad students, hobbyists and industry veterans, new startups and industry players alike that “we need you.” In some ways, this appeal makes this project look like a gamble by Google, with it betting on the chance that some unknown genius makes something special to kick start the open source silicon ecosystem meaningfully.

If you’re looking for a cool Raspberry Pi project to jam out with this summer, take a close look at this open-source MIDI controller and sequencer created by maker and musician Niisse. It’s 100% driven by a Raspberry Pi 4 and features handmade modules that handle special effects as well as input controls for real-time jam sessions.

This is a work in progress but so much progress has been made so far that we think it’s worth sharing. In the demo, Niisse is able to program a beat as well as play instruments with MIDI effects in accompaniment including electric guitar as well as keyboard. Effects can be controlled in real-time using modules and pre-programmed modulations.

The operation is driven by a Raspberry Pi 4 connected to a series of note control modules created by Niisse using SN74 type logic chips on a series of breadboards (at least while it’s in prototype phase). These breadboards feature a series of buttons, LEDs, and display modules to help him keep track of what note is being played as well as any numbers necessary to track the effect details.

The code created for this project was written by Niisse in Python and, like the rest of the project, is open source for anyone who wants to use the code or just take a closer look at how it works. This Python script used by the Pi is responsible for driving the sequencer application and handling input from the individual modules. Visit the RPiMidiSC project page at GitHub to check it out yourself.

According to his GitHub profile, Niisse is a software development student with an affinity for music and the technology that goes along with it. It’s no surprise he’s combined his experience into this open source platform that’s concert ready—even if it’s not totally finished just yet.

There are already plans in the works to upgrade the project. Niisse expressed interest in making the sequencer totally portable and documenting the build process over at YouTube. In the meantime, you can recreate this Raspberry Pi project in its current state by checking out his official YouTube channel where you’ll also find a groovy live demo of it in action.

The netbook age may have come and gone, but there is no denying that sometimes all we need is small device for a quick dose of computing. The tiny, modular Pocket Reform just announced by MNT Research is one of those devices. What first attracted us to the machine was one of its modules includes a Raspberry Pi Compute Module 4, but we stayed for the cute purple styling and chunky good looks.

The idea of the Pocket Reform is to take “the best features” of the original Reform, a laptop designed and built in Berlin, Germany, to be open source in both hardware and software terms. It is sold as a kit, has already received some upgrades, and can be ordered from Crowd Supply in the US.

The Pocket Reform is its little brother, and will be entering a program of beta testing soon. It’s purple like the Nintendo GameCube, and made to be a writing companion, retro gaming deck, or ‘sofa computer’ to chill out with, and is made from a recyclable case and reusable parts. There's even a compact, backlit mechanical keyboard, on Kailh Choc White switches, to make using it even more fun, with a trackball nestled in a gap within the space bar.

The default configuration is built around the NXP i.MX8M Plus chipset (4x ARM Cortex-A53 cores @ 1.8GHz, 4 or 8GB DDR4, Vivante GC7000UL GPU and an NPU). There are various options to replace this, however, including the Compute Module 4 (8GB version), Pine’s SOQuartz RK3566 (via Adapter, 4× ARM Cortex-A55, 8GB DDR4, Mali G52 GPU), the NXP Layerscape LS1028A (2x ARM Cortex-A72 cores with 8 or 16GB DDR4 and the Vivante GC7000UL GPU), and even an FPGA that can become a RISC-V SoC (though this option is marked as for industrial use only). 

The display is 1080p in a 7 inch diagonal for a pixel density of 310ppi, and there's a micro-HDMI output capable of 4K. Wi-Fi 5 and Bluetooth 5.0 are built in by default, and there's an option for a 4G/5G modem with a micro SIM card slot. Storage is taken care of by up to 128GB of eMMC flash, while an M.2 slot awaits an SSD up to 2TB. There's a MicroSD card slot, and a pair of USB-C ports, one of which supports PD charging. Software and OS support is dominated by Unix, with Debian, Arch, Ubuntu and Void Linux all supported, along with the exceptionally nerdy Plan 9 (a research operating system from the same group that created UNIX at Bell Labs) and Genode operating systems. Support for OpenBSD is in development.

It’s not quite clear whether the Pocket Reform will be crowdfunding after its beta program is complete, or will go directly on sale, but we expect further news to come via the MNT website or its CEO’s Twitter account.

Any device sporting a chip and some form of communications protocol can be hacked - that's almost as strong a law as those governing gravity. And sometimes, threats come from where you'd least suspect them. An electronics project by user kingkevin - uploaded on source code repository CuVoodoo git - now aims to help consumers and businesses to protect themselves from an attack vector that's frequently forgotten: a simple HDMI port. Enter the HDMI Firewall.

The HDMI firewall is little more than a physical dongle that can be directly connected to the HDMI port you want to secure. It works by blocking all data that's not related to video and audio streaming. One dongle is required per port to be protected, but they can be repurposed across different devices with a bit of work.

While we may think of HDMI connectors as mere passthroughs for audio and video data, there are a number of parallel protocols that transmit more information than we'd expect. One of the more obvious ones is the addition of HDCP (High-bandwidth Digital Content Protection), a DRM (Digital Rights Management) encryption solution that ensures the transmitted feed isn't copied as it travels towards its destination, and the main culprit of HDMI-related headaches.

Other essential yet often-forgotten HDMI features that aren't specifically related to video and audio include HPD (Hot-Plug Detection) and CEC (Consumer Electronics Control), the later of which allows HDMI devices to be controlled via your remote control - just to name a few.

Interestingly, HDMI also supports I2C - a synchronous, serial communication bus invented forty years ago by Philips Semiconductors. This particular protocol has been shown - time and again - to be vulnerable to intrusion. Programmable controller boards such as an Arduino or Raspberry Pi have been used as means to hack it. But in the HDMI Firewall, it's used to bolster your protection.

The HDMI Firewall needs to be programmed according to the device it's being installed on. Users will need to copy the Extended Display Identification Data (EDID) information of the equipment to protect, which includes information such as supported resolutions, for the passthrough to work. This data can be read by using the I2C-based Display Data Channel (DDC) interface from the device.

This EDID data is then written into the EEPROM of the HDMI firewall, which features a breakable write-protection tab that locks the connector's EEPROM from any further writes. If the Firewall needs to be moved to a new device, users can re-disable write protection by putting a solder blob across two pads on the device, rewrite it with the new devices' EDID data, and then re-enable protection by removing the solder connecting the two pads.

The original post by the creator features a pretty straightforward description of the methods used. Being in an open-source repository, users can freely inspect and fork the code. Healthy discussion on the topic is also taking place at HackerNews, if you're looking for further details on the project and the intricacies of HDMI vulnerabilities. 

We often hear about RISC-V, its competitive open-source architecture, and how it is making inroads into the CPU industry. However, last year we reported on an open source GPU under development, which would seriously bolster the development of open-source graphics engines. This week Think Silicon announced that it is ready to "showcase the industry's first RISC-V-based GPU."

The new RISC-V 3D GPUs from Think Silicon will form the foundation of its Neox series. Two branches are already being readied to expand Neox offerings; the Neox G series for graphics acceleration and the Neox A series for accelerating deep learning tasks.

Why is Think Silicon purposing these two separate Neox branches? It is reasonable to assume that, as Think Silicon is a self-proclaimed "leader in ultra-low power graphics IP," that it designs lean products strictly for targeted tasks. These won't be flabby general-purpose GPUs.

Think Silicon's press release gives us a few more clues to the type of embedded products and applications the respective Neox GPUs will target. It says that the new GPUs will feature customized SoCs for graphics, machine learning, vision/video processing, and general-purpose compute workloads. Mobile devices and platforms appear to be the most obvious, but not exclusive, target for this technology. Think Silicon reckons Neox G and A cores will be used in; next-generation smartwatches, augmented reality (AR) eyewear, video for surveillance and entertainment, and smart displays for point-of-sale/point-of-interaction terminals.

"Unveiling the first RISC-V-based GPU is a significant milestone for the graphics industry and for Think Silicon," said Ulli Mueller, Director of IP Licensing, Sales, and Marketing at Think Silicon. Mueller added that he hoped developers coming to the exhibition would be inspired by his firm's new 3D GPUs. In particular, he hopes that some will see the Neox GPU's potential to create "exceptional user experiences" while being thrifty with power.

Think Silicon will demonstrate its RISC-V 3D GPU family at Embedded World 2022. This trade show takes place in Nuremberg, Germany, from Tuesday, June 21 to Thursday, June 23. It will be interesting to see the level of performance these early RISC-V GPUs can muster, albeit in the low-power and embedded segments.

Before we go, it is worth pointing out that the PC graphics industry big hitters like AMD are also looking at using RISC-V technologies in their product portfolios.

This Raspberry Pi digital audio workstation (DAW), created by maker and musician Stone Preston, takes the Raspberry Pi 4-powered synth idea to a whole new level. The project, known as LMN 3, is entirely open-source and packed with tons of cool features to make custom jams just the way you want them—with plenty of input options and tons of flexibility when it comes to sound wave manipulation.

The Raspberry Pi is a community-driven board and this project wouldn’t be in the same spirit if wasn't open source. Thankfully, Preston has shared all of the juicy details about how it was made, how it works, and what you need to do to make your own at home. All of the repositories are available over at his Fundamental Frequency GitHub profile for any curious party to look through.

Preston considers the LN3 project a DAW-in-a-box with tools to function as a synth, audio sampler and sequencer. This is one of many projects he’s created, past works include an Arduino-based weather application and a custom Arduino library for a 7-segment display module. No doubt this experience culminated into the clever project we see today.

A full build guide is available and includes an exact list of all of the components used in its construction. It’s built around a Raspberry Pi 4 and is assisted by a Teensy 4.1 microcontroller. A Pimoroni Hyperpixel 4 screen is used for video output while input is handled by rotary encoders, a joystick and a series of mechanical keyboard keys. Everything is housed inside of a shell made from an acrylic sheet.

The software used leaves plenty of wiggle room for drawing up custom effects and implementing original audio samples. The interface, firmware and everything necessary to drive the DAW is available over at GitHub for anyone to download.

If you want to recreate this Raspberry Pi project, check out the LMN 3 build guide for the full instructions. There’s also a demo video of the LMN 3 in action over at YouTube if you just want to see what it’s capable of. Be sure to follow Fundamental Frequency for more cool projects and any future updates on this one.