Tuesday, June 21st 2016

NVIDIA Announces a PCI-Express Variant of its Tesla P100 HPC Accelerator

NVIDIA announced a PCI-Express add-on card variant of its Tesla P100 HPC accelerator, at the 2016 International Supercomputing Conference, held in Frankfurt, Germany. The card is about 30 cm long, 2-slot thick, and of standard height, and is designed for PCIe multi-slot servers. The company had introduced the Tesla P100 earlier this year in April, with a dense mezzanine form-factor variant for servers with NVLink.

The PCIe variant of the P100 offers slightly lower performance than the NVLink variant, because of lower clock speeds, although the core-configuration of the GP100 silicon remains unchanged. It offers FP64 (double-precision floating-point) performance of 4.70 TFLOP/s, FP32 (single-precision) performance of 9.30 TFLOP/s, and FP16 performance of 18.7 TFLOP/s, compared to the NVLink variant's 5.3 TFLOP/s, 10.6 TFLOP/s, and 21 TFLOP/s, respectively. The card comes in two sub-variants based on memory, there's a 16 GB variant with 720 GB/s memory bandwidth and 4 MB L3 cache, and a 12 GB variant with 548 GB/s and 3 MB L3 cache. Both sub-variants feature 3,584 CUDA cores based on the "Pascal" architecture, and core clock speed of 1300 MHz.
Add your own comment

6 Comments on NVIDIA Announces a PCI-Express Variant of its Tesla P100 HPC Accelerator

#1
xkm1948
These beasts does protein interactions modeling great. Sadly I am still trying how to configure my FuryX to do similar things with OpenCL. So far I can only get GPUBlast to work.(A type of bioinformatics aligement)
Posted on Reply
#2
matar
I guess this will be the chip for the GTX 1080 Ti
Posted on Reply
#3
robal
matar, post: 3476248, member: 96547"
I guess this will be the chip for the GTX 1080 Ti
Nope. 1080 Ti will be based on smaller chip, without unnecessary FP64.
Posted on Reply
#4
Fluffmeister
The GP102 is a different chip, as robal says the consumer part really doesn't need that level of FP64 support.
Posted on Reply
#5
Slizzo
robal, post: 3476285, member: 57245"
Nope. 1080 Ti will be based on smaller chip, without unnecessary FP64.
Fluffmeister, post: 3476298, member: 101373"
The GP102 is a different chip, as robal says the consumer part really doesn't need that level of FP64 support.
None of that is concrete yet, so I wouldn't post it as fact until nVidia says something about big pascal for consumers.
Posted on Reply
#6
Fluffmeister
Slizzo, post: 3476638, member: 97498"
None of that is concrete yet, so I wouldn't post it as fact until nVidia says something about big pascal for consumers.
Of course, but really the GP100 in it's current configuration would make zero sense as a consumer part.
Posted on Reply