Today at their 2016 GPU Technology Conference, NVIDIA announced the first of their Pascal architecture powered Tesla cards, the Tesla P100. The P100 is the first major update to the Tesla HPC family since the launch of the first Kepler cards in late 2012, and represents a very sizable performance increase for the Tesla family thanks to the combination of the smaller 16nm manufacturing process and the Pascal architecture.

NVIDIA Tesla Family Specification Comparison
  Tesla P100 Tesla K80 Tesla K40 Tesla M40
Stream Processors 3584 2 x 2496 2880 3072
Core Clock 1328MHz 562MHz 745MHz 948MHz
Boost Clock(s) 1480MHz 875MHz 810MHz, 875MHz 1114MHz
Memory Clock 1.4Gbps HBM2 5Gbps GDDR5 6Gbps GDDR5 6Gbps GDDR5
Memory Bus Width 4096-bit 2 x 384-bit 384-bit 384-bit
Memory Bandwidth 720GB/sec 2 x 240GB/sec 288GB/sec 288GB/sec
VRAM 16GB 2 x 12GB 12GB 12GB
Half Precision 21.2 TFLOPS 8.74 TFLOPS 4.29 TFLOPS 6.8 TFLOPS
Single Precision 10.6 TFLOPS 8.74 TFLOPS 4.29 TFLOPS 6.8 TFLOPS
Double Precision 5.3 TFLOPS
(1/2 rate)
2.91 TFLOPS
(1/3 rate)
1.43 TFLOPS
(1/3 rate)
213 GFLOPS
(1/32 rate)
GPU GP100
(610mm2)
GK210 GK110B GM200
Transistor Count 15.3B 2 x 7.1B(?) 7.1B 8B
TDP 300W 300W 235W 250W
Cooling N/A Passive Active/Passive Passive
Manufacturing Process TSMC 16nm FinFET TSMC 28nm TSMC 28nm TSMC 28nm
Architecture Pascal Kepler Kepler Maxwell 2

Powering the Tesla P100 is a partially disabled version of NVIDIA's new GP100 GPU, with 56 of 60 SMs enabled. GP100 is a whale of a GPU, measuring 610mm2 in die size on TSMC's 16nm FinFET process and composed of 15.3B transistors. It is remarkable in and of itself that NVIDIA and TSMC are in volume production of such a large 16nm die at this time, as everything else we've seen with a confirmed size is at best one-quarter of this size. GP100 is being produced on TSMC's Chip-On-Wafer-On-Substrate technology, with this line apparently being used for putting the GPU and HBM2 DRAM stacks on the same interposer.

We'll dive into the full Pascal architecture (as implemented by GP100) at a later time, but it's worth noting that Pascal here is 64 FP32 CUDA cores per SM, versus 128 on Maxwell. Each of those SMs also contains 32 FP64 CUDA cores - giving us the 1/2 rate for FP64 - and new to the Pascal architecture is the ability to pack 2 FP16 operations inside a single FP32 CUDA core under the right circumstances. With a boost clock of 1.48GHz, altogether Tesla P100 will offer 10.6 TFLOPS of FP32 performance or 5.3 TFLOPS of FP64 performance, more than doubling and tripling Tesla K40's rated throughput on these metrics respectively. NVIDIA has been happy to crow about the performance of Tesla P100, and for good reason, as this stands to be a very powerful processor.

Paired with the GP100 GPU on Tesla P100 is 16GB of HBM2 VRAM, laid out in 4 stacks for a 4096-bit memory bus. NVIDIA quotes P100 as offering 720GB/sec of memory bandwidth, which works out to a memory clock of 1.4Gbps. As we've seen with other HBM products, this marks a significant increase in memory bandwidth, more than doubling NVIDIA's last generation of cards.

In their announcement, NVIDIA also confirmed that Tesla P100 will support NVLink, with 4 NVLink controllers. Previously announced, NVLink will allow GPUs to connect to either each other or to supporting CPUs (OpenPOWER), offering a higher bandwidth cache coherent link than what PCIe 3 offers. This link will be important for NVIDIA for a number of reasons, as their scalability and unified memory plans are built around its functionality.

Speaking of functionality, Tesla P100 and the underlying GP100 GPU is a full-featured HPC GPU. It supports all of the HPC-centric functionality that the Tesla K20/40/80 embodied, including ECC memory protection for the register file, caches, and HBM2 DRAM. Coupled with the very high FP64 rate, and it's clear that this is the successor of the GK110/GK210 GPU.

NVIDIA's pictures also confirm that this is using their new mezzanine connector, with flat boards no longer on perpendicular cards. This is a very HPC-centric design (I'd expect to see plenty of PCIe cards in time as well), but again was previously announced and is well suited for the market NVIDIA is going after, where these cards will be installed in a manner very similar to LGA CPUs. The P100 is rated for a TDP of 300W, so the cooling requirements are a bit higher than last-generation cards, most of which were in the 230W-250W range.

Finally, in its initial implementation NVIDIA is focusing on customers that need extreme scaling capabilities, and I wouldn't be too surprised if this was in part due to the margins of that market and how these initial cards will be in demand. NVLink of course plays a big part here, with NVIDIA able to go up to 8-way configurations thanks to it.

Source: NVIDIA

Comments Locked

93 Comments

View All Comments

  • nathanddrews - Tuesday, April 5, 2016 - link

    Does the GP100 confirm that this is the full Pascal die? 16GB of HBM2... good God! I wonder what the 1080/1080Ti will come with?
  • Ryan Smith - Tuesday, April 5, 2016 - link

    This is 36 of 40 SMXes for GP100 enabled.
  • T1beriu - Tuesday, April 5, 2016 - link

    *This is 56(!) of 60(!) SMXes for GP100 enabled.
  • Ryan Smith - Tuesday, April 5, 2016 - link

    Yes. I can't count today. Thanks!
  • Samus - Wednesday, April 6, 2016 - link

    Jesus those are some crazy specs for a not even fully enabled die. 16nm is going to make GPU's ridiculous.
  • ImSpartacus - Wednesday, April 6, 2016 - link

    But we only get it once then that's it.

    we're already at reticle with this one, so we won't even see bigger chips either.

    The only trick is a maxwell-esque efficiency strategy, but I doubt they can get materially more efficient.

    Low dp would buy time though. But that's it.

    Let's hope we don't get stuck on the same process for five years again.
  • SunnyNW - Wednesday, April 6, 2016 - link

    Could they not possibly remove the FP64 cuda cores from the SMs and use that space for more FP32 cores?
  • Flunk - Wednesday, April 6, 2016 - link

    Maybe, but I doubt they will.
  • CrazyElf - Wednesday, April 6, 2016 - link

    It would not be surprising if we got stuck at 16/20nm (it's really a hybrid process) for longer than we did at 28nm.

    Die shrinks are getting harder as we approach the limits of physics.
  • FunBunny2 - Wednesday, April 6, 2016 - link

    -- Die shrinks are getting harder as we approach the limits of physics.

    why do so many ignore this? reality bites, yeah. but there's not a thing you can do about it.

Log in

Don't have an account? Sign up now