Sign in with
Sign up | Sign in

Stepping Through Tegra 4

Nvidia's Tegra 4 GPU: Doubling Down On Efficiency
By

Tegra 4’s vertex processing architecture is essentially the same as NV40’s—the GPU that powered GeForce 6800 and its derivatives back in 2004.

Geometry data from the vertex processing stage is fed into the triangle setup engine, which can do a visible triangle in five clock cycles.

From there, triangles are turned into pixels. Tegra 4 does a combination of raster and early-Z detection at a rate of eight pixels per clock, discarding data that isn’t going to be visible early on in the pipeline and saving the engine from working on occluded pixels unnecessarily. It should come as no surprise (given Nvidia’s background) that this approach borrows heavily from the desktop GPU space.

How does Nvidia’s immediate-mode renderer differ from the tile-based deferred method employed by Imagination Technologies’ PowerVR IP core, and consequently Apple's Ax and Intel's Atom SoCs? Prior to the rasterization stage in a TBDR architecture, frames are cut up into tiles, and the resulting geometry data is put into a memory buffer, where occluded pixels are resolved. Particularly as the geometric complexity of a scene increases, the hidden surface removal process doesn’t fare as well.

Tegra 4 block diagramTegra 4 block diagram

Back to Tegra 4. Color and Z data are then compressed via a lossless algorithm. This is especially beneficial for enabling anti-aliasing without huge memory bandwidth costs (more on this shortly), since the values contained entirely within a primitive tend to be the same. Thus, they compress away nicely, yielding high compression ratios. Now, a lossless approach means that data is only compressed when it can be, so you still have to allocate space in memory—there are no savings there. But a lot of bandwidth can be conserved.

The raster stage feeds Tegra 4’s fragment pipe, which can do four pixels per clock. As mentioned, each pixel pipe has three ALUs with four multiply-add units, plus one multi-function unit, facilitating a number of VLIW instruction combinations to do different things (normalizes and combines, blends, traditional lighting calculations, and so on). Tegra 4 exposes 24 FP20 registers per pixel, compared to Tegra 3’s 16, allowing more threads in flight at any given time.

The four pipes have their own read- and write-capable L1 cache, and are serviced by a shared L2 texture cache—a new feature to Tegra 4. Naturally, you get better locality for texture filtering, again saving memory bandwidth. According to Nvidia, the cache is also really well-optimized for 2D imaging-style operations, which plays into the company’s work with computational photography.

Even as Nvidia’s engineers emphasize bandwidth-saving across the GPU, balancing a significant increase in texture rate requires a memory subsystem able to keep the SoC’s resources busy. Tegra 3 got by with a single-channel 32-bit interface. Tegra 4 uses two 32-bit channels, along with LPDDR3 memory at up to 1,866 MT/s, to push more than 3x the throughput available from LPDDR2-1066.

Display all 41 comments.
This thread is closed for comments
Top Comments
  • 23 Hide
    Memnarchon , February 24, 2013 9:17 PM
    A Sunday article? :o 
  • 20 Hide
    blazorthon , February 24, 2013 10:02 PM
    deedee2die4Nvidia staying on top, the best of the best!


    Uhh, no... T4 isn't supposed to be out for like six months, yet it's already not as fast as some of Qualcomm's latest. Nvidia is improving, but as usual, they're staying a little behind in technology.
  • 18 Hide
    s3anister , February 24, 2013 8:17 PM
    I'm always amazed with the progress made in strides in this ultra-competitive sector so it's nice to see nvidia finally hit 28mm with Tegra 4. I'm sure some of their performance gains can be attributed to this.
Other Comments
  • 18 Hide
    s3anister , February 24, 2013 8:17 PM
    I'm always amazed with the progress made in strides in this ultra-competitive sector so it's nice to see nvidia finally hit 28mm with Tegra 4. I'm sure some of their performance gains can be attributed to this.
  • -3 Hide
    levin70 , February 24, 2013 8:22 PM
    Charlie at semiaccurate is correct. The Tegra 4 is DOA. Almost no one will be using it. Everyone else is already ahead of where the T4 is today, and it hasn't even launched. How many design wins were noted? 1?

    Yeah, says it all.
  • 23 Hide
    Memnarchon , February 24, 2013 9:17 PM
    A Sunday article? :o 
  • -3 Hide
    deedee2die4 , February 24, 2013 9:25 PM
    Nvidia staying on top, the best of the best!
  • 20 Hide
    blazorthon , February 24, 2013 10:02 PM
    deedee2die4Nvidia staying on top, the best of the best!


    Uhh, no... T4 isn't supposed to be out for like six months, yet it's already not as fast as some of Qualcomm's latest. Nvidia is improving, but as usual, they're staying a little behind in technology.
  • 6 Hide
    aicom , February 24, 2013 10:05 PM
    levin70Charlie at semiaccurate is correct. The Tegra 4 is DOA. Almost no one will be using it. Everyone else is already ahead of where the T4 is today, and it hasn't even launched. How many design wins were noted? 1?Yeah, says it all.

    Nobody is ahead of Tegra's four Cortex A15 cores. Krait is at less performance than A15 (until the refresh at least). Samsung's got Exynos 5 Octa, but that's not out yet either and T4 will probably still top it in the GPU performance department. Speaking of which, Tegra 4 has the most powerful GPU in floating-point of anyone (including the iPad 4) with 74.8 TFLOPS @ 672 MHz. It only takes a 825 MHz Cortex A15 to match a 1.6 GHz A9, and Tegra 4 is supposed to ship at 1.9 GHz. Unfortunately, TDP does go up in the process.

    You also have to look at where these parts are targeted. Krait is really gunning for phone design wins and they have many. It's a very power efficient chip that found its way into some very nice phones. Tegra 4 is not aimed at that market; Tegra 4i is. Tegra 4 will have a much higher TDP than 4i (and Krait) and will get substantially higher performance as a result.
  • 0 Hide
    tjosborne , February 24, 2013 10:33 PM
    Hey guys, I am considering getting a Asus transformer prime tablet with the tegra 3. Would it be best to wait till this processor ends up in a tablet to get one?
  • 4 Hide
    Anonymous , February 24, 2013 10:39 PM
    So at 1.3Gpix/s, Nvidia has just admitted to 10x overdraw...per second? So we're looking at 9~10 frames per second oh high res displays. Lag lives on.
  • 4 Hide
    PreferLinux , February 24, 2013 11:35 PM
    aicomNobody is ahead of Tegra's four Cortex A15 cores. Krait is at less performance than A15 (until the refresh at least). Samsung's got Exynos 5 Octa, but that's not out yet either and T4 will probably still top it in the GPU performance department. Speaking of which, Tegra 4 has the most powerful GPU in floating-point of anyone (including the iPad 4) with 74.8 TFLOPS @ 672 MHz. It only takes a 825 MHz Cortex A15 to match a 1.6 GHz A9, and Tegra 4 is supposed to ship at 1.9 GHz. Unfortunately, TDP does go up in the process.You also have to look at where these parts are targeted. Krait is really gunning for phone design wins and they have many. It's a very power efficient chip that found its way into some very nice phones. Tegra 4 is not aimed at that market; Tegra 4i is. Tegra 4 will have a much higher TDP than 4i (and Krait) and will get substantially higher performance as a result.

    You mean Gigaflops, not Teraflops.
  • -1 Hide
    blazorthon , February 24, 2013 11:43 PM
    aicomNobody is ahead of Tegra's four Cortex A15 cores. Krait is at less performance than A15 (until the refresh at least). Samsung's got Exynos 5 Octa, but that's not out yet either and T4 will probably still top it in the GPU performance department. Speaking of which, Tegra 4 has the most powerful GPU in floating-point of anyone (including the iPad 4) with 74.8 TFLOPS @ 672 MHz. It only takes a 825 MHz Cortex A15 to match a 1.6 GHz A9, and Tegra 4 is supposed to ship at 1.9 GHz. Unfortunately, TDP does go up in the process.You also have to look at where these parts are targeted. Krait is really gunning for phone design wins and they have many. It's a very power efficient chip that found its way into some very nice phones. Tegra 4 is not aimed at that market; Tegra 4i is. Tegra 4 will have a much higher TDP than 4i (and Krait) and will get substantially higher performance as a result.


    S4 Pro is a faster CPU IIRC. IDK about how the graphics compares and won't comment about it.

    Nvidia, like I said, is getting better, but they're still going to be a little behind. They're making up a lot of ground here, especially with how they're making Tegra 4 and Tegra 4i instead of a single SoC to take both places, but they seem like they'l still have a little room to make up, at least in CPU performance, to be the best. Like I said before (at least in other articles about it), they'll still be near the top either way.
  • 1 Hide
    blazorthon , February 25, 2013 12:38 AM
    ankit0x1can this play crysis? ('console port' on arm)


    Just keep the texture settings low and the resolution at minimum and it might be almost playable if someone manages to get it to work on ARM ;) 
  • -8 Hide
    sanilmahambre , February 25, 2013 1:00 AM
    Tegra 4 in Razer gaming tab is best example of NVIDIA's superiority.
  • 2 Hide
    mayankleoboy1 , February 25, 2013 2:04 AM
    blazorthonUhh, no... T4 isn't supposed to be out for like six months, yet it's already not as fast as some of Qualcomm's latest. Nvidia is improving, but as usual, they're staying a little behind in technology.


    +1
    In this hyper competitive market, every SoC company needs atleast one product that is best in one area over every other SoC. Middling in all the areas is what T3 did, which made it a failure.
  • 2 Hide
    blazorthon , February 25, 2013 2:31 AM
    mayankleoboy1+1In this hyper competitive market, every SoC company needs atleast one product that is best in one area over every other SoC. Middling in all the areas is what T3 did, which made it a failure.


    To be fair, although in many ways not an exceptional SoC, T3 was a very successful product, wasn't it? They managed to get Tegra 3 in one form or another around to a lot of well-selling devices.
  • 3 Hide
    mayankleoboy1 , February 25, 2013 2:41 AM
    blazorthonTo be fair, although in many ways not an exceptional SoC, T3 was a very successful product, wasn't it? They managed to get Tegra 3 in one form or another around to a lot of well-selling devices.


    I would say that T3 came with a bang and lots of fanfair, but as soon as the Krait4 came, nobody wanted to be seen with a T3.
    After that T3 pretty quickly started appearing in the Tier2 phones. All T1 phones used either a KraitS4 or the Exynos4. I would even say that Ms's Surface was a semi-fail in performance because of the T3. If MS had went with a dual core KraitS4, i am pretty sure Surface would have received much better reviews.
  • 1 Hide
    blazorthon , February 25, 2013 2:43 AM
    mayankleoboy1I would say that T3 came with a bang and lots of fanfair, but as soon as the Krait4 came, nobody wanted to be seen with a T3.After that T3 pretty quickly started appearing in the Tier2 phones. All T1 phones used either a KraitS4 or the Exynos4. I would even say that Ms's Surface was a semi-fail in performance because of the T3. If MS had went with a dual core KraitS4, i am pretty sure Surface would have received much better reviews.


    +1 I can agree with that :) 
  • -4 Hide
    mayankleoboy1 , February 25, 2013 2:46 AM
    As Nvidia is not a market leader in SoC's, they cant do anything revolutionary in the T4 SoC . Like increasing the die area for adding more GPU muscle. Because they dont know if the OEM's will accept that or not. So whatever they do, it will always be conservative.
    Apple and Samsung have no such problem. THey can afford be revolutionary, without much consequences.
  • 5 Hide
    aicom , February 25, 2013 2:47 AM
    Not sure where people are getting the idea that Krait is faster than A15. S4 Pro's max clock is 1.7 GHz, while Tegra 4's is 1.9 GHz, so A15 already wins the clock speed match. In addition, Krait is a touch slower than A15 (3.3 DMIPS/MHz vs 3.5 DMIPS/MHz).

    http://en.wikipedia.org/wiki/Krait_(CPU)
    http://en.wikipedia.org/wiki/Snapdragon_(system_on_chip)#Snapdragon_S4
  • 1 Hide
    mayankleoboy1 , February 25, 2013 2:57 AM
    ^
    Krait S4 is a modified A9 architecture, that performs between A9 and A15 archs, but closer to A9.
    Apples Swift architecture is also a modified architecture that performs between A9 and A15, but much closer to the A15 arch.
  • 1 Hide
    blazorthon , February 25, 2013 3:10 AM
    Quote:
    Not sure where people are getting the idea that Krait is faster than A15. S4 Pro's max clock is 1.7 GHz, while Tegra 4's is 1.9 GHz, so A15 already wins the clock speed match. In addition, Krait is a touch slower than A15 (3.3 DMIPS/MHz vs 3.5 DMIPS/MHz).

    http://en.wikipedia.org/wiki/Krait_(CPU)
    http://en.wikipedia.org/wiki/Snapdragon_(system_on_chip)#Snapdragon_S4


    Quote:
    ^
    Krait S4 is a modified A9 architecture, that performs between A9 and A15 archs, but closer to A9.
    Apples Swift architecture is also a modified architecture that performs between A9 and A15, but much closer to the A15 arch.


    If we're going by the links in the above post, Krait is much closer to A15 than to A9. This is still ignoring the fact that those types of measurements are extremely general and oftentimes only loosely related to performance in the real world.

    My bad, it's the newer top Snapdragon 800 version that Tegra 4 doesn't beat in CPU performance, thanks for the clarification. Regardless, I seem to have given Nvidia too little credit with Tegra 4.
Display more comments