Designing GP104: Running Up the Clocks

So if GP104’s per-unit throughput is identical to GM204, and the SM count has only been increased from 2048 to 2560 (25%), then what makes GTX 1080 60-70% faster than GTX 980? The answer there is that instead of vastly increasing the number of functional units for GP104 or increasing per-unit throughput, NVIDIA has instead opted to significantly raise the GPU clockspeed. And this in turn goes back to the earlier discussion on TSMC’s 16nm FinFET process.

With every advancement in fab technology, chip designers have been able to increase their clockspeeds thanks to the basic physics at play. However because TSMC’s 16nm node adds FinFETs for the first time, it’s extra special. What’s happening here is a confluence of multiple factors, but at the most basic level the introduction of FinFETs means that the entire voltage/frequency curve gets shifted. The reduced leakage and overall “stronger” FinFET transistors can run at higher clockspeeds at lower voltages, allowing for higher overall clockspeeds at the same (or similar) power consumption. We see this effect to some degree with every node shift, but it’s especially potent when making the shift from planar to FinFET, as has been the case for the jump from 28nm to 16nm.

Given the already significant one-off benefits of such a large jump in the voltage/frequency curve, for Pascal NVIDIA has decided to fully embrace the idea and run up the clocks as much as is reasonably possible. At an architectural level this meant going through the design to identify bottlenecks in the critical paths – logic sections that couldn’t run at as high a frequency as NVIDIA would have liked – and reworking them to operate at higher frequencies. As GPUs typically (and still are) relatively low clocked, there’s not as much of a need to optimize critical paths in this matter, but with NVIDIA’s loftier clockspeed goals for Pascal, this changed things.

From an implementation point of view this isn’t the first time that NVIDIA has pushed for high clockspeeds, as most recently the 40nm Fermi architecture incorporated a double-pumped shader clock. However this is the first time NVIDIA has attempted something similar since they reined in their power consumption with Kepler (and later Maxwell). Having learned their lesson the hard way with Fermi, I’m told a lot more care went into matters with Pascal in order to avoid the power penalties NVIDIA paid with Fermi, exemplified by things such as only adding flip-flops where truly necessary.

Meanwhile when it comes to the architectural impact of designing for high clockspeeds, the results seem minimal. While NVIDIA does not divulge full information on the pipeline of a CUDA core, all of the testing I’ve run indicates that the latency (in clock cycles) of the CUDA cores is identical to Maxwell. Which goes hand in hand with earlier observations about throughput. So although optimizations were made to the architecture to improve clockspeeds, it doesn’t look like NVIDIA has made any more extreme optimizations (e.g. pipeline lengthening) that detectably reduces Pascal’s per-clock performance.

Beyond3D Suite - Estimated MADD Latency

Finally, more broadly speaking, while this is essentially a one-time trick for NVIDIA, it’s an interesting route for them to go. By cranking up their clockspeeds in this fashion, they avoid any real scale-out issues, at least for the time being. Although graphics are the traditional embarrassingly parallel problem, even a graphical workload is subject to some degree of diminishing returns as GPUs scale farther out. A larger number of SMs is more difficult to fill, not every aspect of the rendering process is massively parallel (shadow maps being a good example), and ever-increasing pixel shader lengths compound the problem. Admittedly NVIDIA’s not seeing significant scale-out issues quite yet, but this is why GTX 980 isn’t quite twice as fast as GTX 960, for example.

Just increasing the clockspeed, comparatively speaking, means that the entire GPU gets proportionally faster without shifting the resource balance; the CUDA cores are 43% faster, the geometry frontends are 43% faster, the ROPs are 43% faster, etc. The only real limitation in this regard isn’t the GPU itself, but whether you can adequately feed it. And this is where GDDR5X comes into play.

FP16 Throughput on GP104: Good for Compatibility (and Not Much Else) Feeding Pascal: GDDR5X
Comments Locked

200 Comments

View All Comments

  • Ryan Smith - Wednesday, July 20, 2016 - link

    The prices listed are based on checking Newegg daily for the past week. The sub-$429 cards have never been in stock when I've checked.
  • powerarmour - Wednesday, July 20, 2016 - link

    Lulz how late?
  • jabbadap - Wednesday, July 20, 2016 - link

    Great thorough review.

    You have a funny typo on the fast sync page:
    " Fast Sync doesn’t replace either Adaptive Sync or G-Sync"
    I think you meant Adaptive Vsync, not that VESA standard which nvidia does not support.
  • Ryan Smith - Wednesday, July 20, 2016 - link

    Right you are. Thanks!
  • bill44 - Wednesday, July 20, 2016 - link

    Where can I find a FULL review please?

    Thanks
  • Ryan Smith - Wednesday, July 20, 2016 - link

    "Where can I find a FULL review please?"

    This article was an excerpt of the full review. The complete, 5 volume set on GTX 1080 will be available at your local book store in October. ;-)
  • catavalon21 - Wednesday, July 20, 2016 - link

    Awesome
  • AndrewJacksonZA - Thursday, July 21, 2016 - link

    Cool! Thanks Ryan! Did you reach your stretch goals on your Kickstarter campaign so that you can afford the scanning electron microscope for great photos?
  • bill44 - Thursday, July 21, 2016 - link

    Thanks Ryan.
    I was on about audio. Not for games, but for my music playback (SACD, DVDA) with different sampling rates that includes 88.2 and 176.4Khz. I can't find FULL specifications for the 100 series GTX cards that include this information, or any decoding capabilities. I was looking for the GTX 1060 for HTPC use with madVR.

    Where can I find out about audio and sampling rates? MadVR capabilities?
  • Ryan Smith - Friday, July 22, 2016 - link

    As far as I know, audio capabilities are the same as on the 900 series. Unfortunately I don't have more detail than that right now.

Log in

Don't have an account? Sign up now