Loongson: Next-Gen Quad-Core Chinese CPU Matches Intel's Tiger Lake

Chinese chipmaker Loongson has long promised that its next-generation 3A6000-series processors based on the LoongArch microarchitecture will match AMD's Zen 3 CPUs in terms of instructions per clock (IPC) performance. This week the company disclosed (via @9550pro) some of the actual test results of its quad-core 3A6000 processor and said that they were comparable to Intel's quad-core '10th-Gen Core' processor from 2020. But there's a catch.

"Based on the relevant test results, the overall performance of the Loongson 3A6000 processor is comparable to that of Intel's 10th generation Core quad-core processor launched in 2020," a statement by Loongson reads.

Loongson's quad-core 3A6000 processor has a 2.50 GHz clock rate. The chip hit 43.1/54.6 points in the SPECint_base2006/SPECfp_base2006 benchmarks, as well as 155/140 in SPECint_rate_base2006 (8 copies)/SPECfp_rate_base2006 (8 copies), at least according to the tests conducted by Saixi Laboratory of the China Institute of Electronic Technology Standardization.

However, SPEC discontinued its SPEC 2006 CPU benchmark in 2018, making it impossible to compare Loonson's performance numbers to independently obtained results approved by SPEC. Meanwhile, the new 3A6000 is noticeably faster than its predecessor, based on previously published test results of the quad-core 3A5000 at 2.50 GHz.

The same institute also tested the CPU in UnixBench v5.1.3 and got a 2284.5 single-thread score and 7438.4 eight-thread score. If Intel Core i7-10750H's UnixBench performance numbers published in a Reddit post are to be believed (1713 single-thread, 8248 12-thread), then the 3A6000 indeed outpaces Intel's six-core Comet Lake at 2.60 GHz when it comes to single-thread workloads at around the same clock, but falls behind a six-core CPU with SMT enabled. It's noteworthy that Intel's Comet Lake uses Skylake cores from 2015.

To that end, it is too early to draw any conclusions about the performance of Loongson's upcoming processors. On the one hand, it looks like the 3A6000 is faster than the 3A5000 at the same clock in the discontinued SPEC CPU 2006 benchmark, but that's based on results published by the company itself. The chip also seems to be faster than its predecessor and Intel's Skylake in UnixBench at 2.50 – 2.60 GHz.

TOPICS

Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.

13 Comments Comment from the forums

ekio

It matches tiger lake...?

You really think people are learning the marketing nickname of each generation?
Just say gen 11000 or something, you know, like the real product name.
Reply
bit_user

Based on the relevant test results, the overall performance of the Loongson 3A6000 processor is comparable to that of Intel's 10th generation Core quad-core processor launched in 2020
That would be either Ice Lake (mobile) or Comet Lake (desktop). Not Tiger Lake (Gen 11, mobile). BTW, Ice Lake launched at the end of 2019.

SPEC discontinued its SPEC 2006 CPU benchmark in 2018, making it impossible to compare Loonson's performance numbers to independently obtained results approved by SPEC.
That might be true for official figures, but there's plenty of SPEC2006 data out there, if you look. Anandtech continued using it through 2020. These are single-threaded:
Source: https://www.anandtech.com/show/16252/mac-mini-apple-m1-tested/4
So, their CPU's int score of 43.1 puts it just below Tiger Lake @ 15 W, but the fp score of 54.6 falls well short of Tiger Lake and lands just below Ryzen 2700X.
Reply
bit_user

ekio said:
It matches tiger lake...?

You really think people are learning the marketing nickname of each generation?
Just say gen 11000 or something, you know, like the real product name.
Tiger Lake is the codename of Intel's Gen 11 mobile CPUs that were at the cutting-edge from late 2020 to late 2021 (?).

A lot of regular readers of Toms would know that, so I think it's okay to put in the headline and a lot less awkward than "Intel's 11th Gen Mobile CPUs". If you want the details, you could read the article. Unfortunately, the article references Gen 10, which should be Ice Lake.
Reply
eric79xxl

Even if you just go by the older test program results, this is the closest China (or Russia) has been to matching Intel or AMD performance.... Very worrisome.
It'd be a shame if something happened to their fab and R&D facilities.
Reply
bit_user

ericpeacock79 said:
Even if you just go by the older test program results, this is the closest China (or Russia) has been to matching Intel or AMD performance.... Very worrisome.
Whether it's worrisome is a matter of geopolitics and global trade policy. We might guess how those will play out, but we don't know for certain. I try to set that stuff aside and just look at these developments from a technical standpoint.

I've even chatted a little with a software developer working on supporting the platform, and they're not much different than us. I think it helps to remember that.
Reply
artk2219

bit_user said:
Whether it's worrisome is a matter of geopolitics and global trade policy. We might guess how those will play out, but we don't know for certain. I try to set that stuff aside and just look at these developments from a technical standpoint.

I've even chatted a little with a software developer working on supporting the platform, and they're not much different than us. I think it helps to remember that.
Not different at all, governments are not often representative of their people. At the end of the day we all basically want the same things, shelter, food, clothing, and the ability to support and keep those things for ourselves and others we are close to. Everything else is basically icing on top. People tend to gravitate towards tribalism because its typically easier to "blame the other guy" than to take a step back and look at the bigger picture to realize we're all in this crap together.
Reply
Geef

artk2219 said:
Not different at all, governments are not often representative of their people. At the end of the day we all basically want the same things, shelter, food, clothing, and the ability to support and keep those things for ourselves and others we are close to.
Only difference is I have a feeling the hardware will have strange things added. Not sure what. Maybe it forces your PC's startup screen to have a pic of Xi Jinping or something?
Reply
bit_user

Geef said:
Only difference is I have a feeling the hardware will have strange things added. Not sure what.
To me, the obvious thing would be if it only allows signed microcode, BIOS, and boot images, which must be signed by a government entity. That lets them ensure backdoors or spyware is in place and can't be circumvented.

People talk about backdoors in the hardware, but if you just limit the software & firmware to government-approved releases, then there don't even need to be any backdoors in the hardware itself. In fact, you don't really even need to design a custom CPU, either (except for maybe the microcode-signing part) and you sure don't need a custom ISA. I think the custom ISA is really just about ensuring nobody else controls what they want to do with the platform.

BTW, a good argument not to put backdoors in the hardware itself is if a foreign adversary (or just your run-of-the-mill cybercriminal) finds it, then you're stuck with millions or billions of vulnerable machines in the wild. Whereas, if you do the backdoors in software and one of them is compromised, you can just issue a software update to close it and add a different one in its place.
Reply
bit_user

artk2219 said:
At the end of the day we all basically want the same things, shelter, food, clothing, and the ability to support and keep those things for ourselves and others we are close to.
No, we didn't discuss anything like that. They just seemed like a reasonable developer, capable of seeing things from different perspectives.

Their take on friction between some of the other the Loongson developers and the projects they were trying to upstream their patches to was that a lot of it was due to communication problems and language issues. I suggested they might hire some liaisons to help smooth over some of these interactions, as they didn't have anyone tasked with that sort of thing.

artk2219 said:
Everything else is basically icing on top. People tend to gravitate towards tribalism because its typically easier to "blame the other guy" than to take a step back and look at the bigger picture to realize we're all in this crap together.
The problem for a project maintainer is that you get these patches of varying quality and levels of documentation that seem to come out of nowhere, from people without a track record, and with no or minimal prior discussion on the mailing list. This creates stress for the maintainer, who already has enough on their plate and it creates stress for the submitter, who's under pressure to get the patches accepted and doesn't understand why they're getting so much pushback. As we all know, stressful situations rarely bring out the best in people.

Also, it didn't help that some of the architecture (mostly system architecture rather than ISA, as far as I understand) was forked off of MIPS, yet they insisted it's not MIPS. From a maintainer's perspective, the distinction seems artificial and unnecessary. So, it just creates yet another point of friction.
Reply
nookoool

This might be a fun hobbyist machine outside of China if they can get the board+cpu for a reasonable price and have a english version of a loongson opitimize linux
Reply

Show more comments