Hprc banner tamu.png

Difference between revisions of "HW:GPUs"

From TAMU HPRC
Jump to: navigation, search
(Features)
(Features)
Line 51: Line 51:
 
| Interconnect || Example || Example || Example || Example || Example || Example
 
| Interconnect || Example || Example || Example || Example || Example || Example
 
|-
 
|-
| Compute Capability || CUDA 3.5 || Example || Example || Example || Example || Example
+
| Compute Capability || CUDA 3.5 || CUDA 7.0 || CUDA 3.7 || CUDA 8.0 || CUDA 7.5 || CUDA 7.5
 
|}
 
|}
  

Revision as of 12:35, 25 February 2021

[Need to turn this (and additional info) into a nice table. Just dumping nvidia-smi for now. eg Need to add things like 1)#/cluster 2)# threads/pipes/"tensors"/whatever 3) clock speeds 4) GPURAM 5) "Compute Capabilities" (!!!) - e.g. cuda_compute_capabilities = ['3.5', '3.7'] # for Tesla K20 (ada) and K80 (terra) 6) etc. ]

Overview

GPUs By Cluster

GPU Offerings by Cluster
Ada Tesla K20m
Tesla V100 PCIe 32GB
Terra Tesla K80
Grace Grid A100 PCIe 40GB
Quadro RTX 6000/8000
Tesla T4

Features

Table Showing Specifications for the GPU's available on our cluster. Specs listed as Rows with GPU's as columns. Resources for table data -- CUDA compute capability

Specification Tesla K20m Tesla V100 PCIe 32GB Tesla K80 Grid A100 PCIe 40GB Quadro RTX 6000 Tesla T4
Cluster Ada Ada/Terra Terra Grace Grace Grace
Cores 2496 5120 2496x2 6912 4608 2560
GPU Memory 5GB 32GB 12GB x2 40GB 24GB 16GB
GPU Memory Bandwidth 208.0 GB/s 897.0 GB/s 240.6 GB/s x2 1,555 GB/s 672.0 GB/s 320.0 GB/s
Peak FP64 1,175 GFLOPS 7.066 TFLOPS 1,371 GLOPS x2 9.746 TFLOPS 509.8 GFLOPS 254.4 GFLOPS
Peak FP32 3.524 TFLOPS 14.13 TFLOPS 4.113 TFLOPS x2 19.49 TFLOPS 16.31 TFLOPS 8.141 TFLOPS
Peak Mixed Precision (FP16/FP32) Example Example Example Example Example Example
Peak INT8 Example Example Example Example Example Example
Peak INT4 Example Example Example Example Example Example
Interconnect Example Example Example Example Example Example
Compute Capability CUDA 3.5 CUDA 7.0 CUDA 3.7 CUDA 8.0 CUDA 7.5 CUDA 7.5

Ada

GK110GL [Tesla K20m] (rev a1)

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.95.01    Driver Version: 440.95.01    CUDA Version: 10.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla K20m          On   | 00000000:20:00.0 Off |                    0 |
| N/A   19C    P8    26W / 225W |      0MiB /  4743MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
NVIDIA-SMI 440.95.01 Driver Version: 440.95.01 CUDA Version: 10.2
GPU Name Persistence-M Fan
Temp Perf Pwr:Usage/Cap
Bus-Id Disp.A
Memory-Usage
Volatile Uncorr. ECC
GPU-Util Compute M.
0 Tesla K20m On
N/A 19C P8 26W / 225W
00000000:20:00.0 Off
0MiB / 4743MiB
0
0% Default

GV100GL [Tesla V100 PCIe 32GB] (rev a1)

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.95.01    Driver Version: 440.95.01    CUDA Version: 10.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla V100-PCIE...  Off  | 00000000:3B:00.0 Off |                    0 |
| N/A   62C    P0   194W / 250W |   1839MiB / 32510MiB |    100%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla V100-PCIE...  Off  | 00000000:D8:00.0 Off |                    0 |
| N/A   32C    P0    27W / 250W |     12MiB / 32510MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
NVIDIA-SMI 440.95.01 Driver Version: 440.95.01 CUDA Version: 10.2
GPU Name Persistence-M Fan
Temp Perf Pwr:Usage/Cap
Bus-Id Disp.A
Memory-Usage
Volatile Uncorr. ECC
GPU-Util Compute M.
0 Tesla V100-PCIE... Off
N/A 62C P0 194W / 250W
00000000:3B:00.0 Off
1839MiB / 32510MiB
0
100% Default
1 Tesla V100-PCIE... Off
N/A 32C P0 27W / 250W
00000000:D8:00.0 Off
12MiB / 32510MiB
0
0% Default

Terra

GK210GL [Tesla K80] (rev a1)

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 450.51.06    Driver Version: 450.51.06    CUDA Version: 11.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla K80           On   | 00000000:83:00.0 Off |                  Off |
| N/A   27C    P8    26W / 149W |      0MiB / 12206MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
|   1  Tesla K80           On   | 00000000:84:00.0 Off |                  Off |
| N/A   35C    P8    30W / 160W |      0MiB / 12206MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+

Grace

GA100 [GRID A100 PCIe 40GB] (rev a1)

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.27.04    Driver Version: 460.27.04    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  A100-PCIE-40GB      On   | 00000000:3B:00.0 Off |                    0 |
| N/A   20C    P0    29W / 250W |      0MiB / 40536MiB |      0%      Default |
|                               |                      |             Disabled |
+-------------------------------+----------------------+----------------------+

TU102GL [Quadro RTX 6000/8000] (rev a1)

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.27.04    Driver Version: 460.27.04    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Quadro RTX 6000     On   | 00000000:3B:00.0 Off |                    0 |
| N/A   18C    P8    12W / 250W |      0MiB / 22698MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
|   1  Quadro RTX 6000     On   | 00000000:D8:00.0 Off |                    0 |
| N/A   19C    P8    14W / 250W |      0MiB / 22698MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+

TU104GL [Tesla T4] (rev a1)

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.27.04    Driver Version: 460.27.04    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla T4            On   | 00000000:3B:00.0 Off |                    0 |
| N/A   22C    P8     9W /  70W |      0MiB / 15109MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
|   1  Tesla T4            On   | 00000000:AF:00.0 Off |                    0 |
| N/A   25C    P8     9W /  70W |      0MiB / 15109MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+