NVIDIA Tesla K40st vs NVIDIA GRID A100A
GPU · Specs Compare
Graphics Card
Release Date
Nov 2013
May 2020
Generation
Tesla
GRID
Type
Professional
Professional
Bus Interface
PCIe 3.0 x16
PCIe 4.0 x16
Clock Speeds
Memory Clock
1502 MHz
1215 MHz
Base Clock
—
1095 MHz
Boost Clock
—
1410 MHz
Memory
Memory Size
12GB
32GB
Memory Type
GDDR5
HBM2e
Memory Bus
384bit
6144bit
Bandwidth
288.4GB/s
1.87TB/s
Render Config
Shading Units
2880
6912
TMUs
240
432
ROPs
48
192
L1 Cache
16 KB (per SMX)
192 KB (per SM)
L2 Cache
1536 KB
32 MB
SM Count
—
108
Tensor Cores
—
432
Theoretical Performance
Pixel Rate
34.50 GPixel/s
270.7 GPixel/s
Texture Rate
138.0 GTexel/s
609.1 GTexel/s
FP32 (float)
3.312 TFLOPS
19.49 TFLOPS
FP64 (double)
1104 GFLOPS
9.746 TFLOPS
FP16 (half)
—
77.97 TFLOPS
Graphics Processor
GPU Name
GK110B
GA100
Architecture
Kepler
Ampere
Foundry
TSMC
TSMC
Process Size
28 nm
7 nm
Transistors
7.08 billion
54.2 billion
Die Size
561 mm²
826 mm²
Board Design
TDP
245W
400W
Suggested PSU
550 W
800 W
Outputs
No outputs
No outputs
Power Connectors
—
None
Graphics Features
DirectX
12 (11_1)
N/A
OpenGL
4.6
N/A
OpenCL
3.0
3.0
Vulkan
1.1
N/A
CUDA
3.5
8.0
Shader Model
5.1
N/A
▲ Specs Compare
Data sourced from official specifications and benchmark tests.