NVIDIA Tesla V100 FHHL vs NVIDIA A100 PCIe
AI GPU
We compared a Professional market GPU: 16GB VRAM Tesla V100 FHHL and a GPU: 40GB VRAM A100 PCIe to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA A100 PCIe 's Advantages
Released 2 years and 3 months late
Boost Clock has increased by 9% (1410MHz vs 1290MHz)
More VRAM (40GB vs 16GB)
Larger VRAM bandwidth (1555GB/s vs 829.4GB/s)
1792 additional rendering cores
Score
Benchmark
FP32 (float)
Tesla V100 FHHL
13.21 TFLOPS
A100 PCIe
+47%
19.49 TFLOPS
Graphics Card
Mar 2018
Release Date
Jun 2020
Tesla
Generation
Tesla
Professional
Type
AI GPU
PCIe 3.0 x16
Bus Interface
PCIe 4.0 x16
Clock Speeds
937 MHz
Base Clock
765 MHz
1290 MHz
Boost Clock
1410 MHz
810 MHz
Memory Clock
1215 MHz
Memory
16GB
Memory Size
40GB
HBM2
Memory Type
HBM2e
4096bit
Memory Bus
5120bit
829.4GB/s
Bandwidth
1555GB/s
Render Config
-
-
-
80
SM Count
108
5120
Shading Units
6912
320
TMUs
432
128
ROPs
160
640
Tensor Cores
432
-
-
-
128 KB (per SM)
L1 Cache
192 KB (per SM)
6 MB
L2 Cache
40 MB
-
-
-
Theoretical Performance
165.1 GPixel/s
Pixel Rate
225.6 GPixel/s
412.8 GTexel/s
Texture Rate
609.1 GTexel/s
26.42 TFLOPS
FP16 (half)
77.97 TFLOPS
13.21 TFLOPS
FP32 (float)
19.49 TFLOPS
6.605 TFLOPS
FP64 (double)
9.746 TFLOPS
Board Design
250W
TDP
250W
600 W
Suggested PSU
600 W
No outputs
Outputs
No outputs
1x 8-pin
Power Connectors
8-pin EPS
Graphics Processor
GV100
GPU Name
GA100
-
-
-
Volta
Architecture
Ampere
TSMC
Foundry
TSMC
12 nm
Process Size
7 nm
21.1 billion
Transistors
54.2 billion
815 mm²
Die Size
826 mm²
Graphics Features
12 (12_1)
DirectX
N/A
4.6
OpenGL
N/A
3.0
OpenCL
3.0
1.3
Vulkan
N/A
7.0
CUDA
8.0
6.6
Shader Model
N/A
