NVIDIA DGX A100 Professional Computing Solution
8x NVIDIA Tesla A100
Product status: Official | Last Update: 2020-05-19 | Report Error
Graphics Processor
GA100 ×8
Cores : TMUs : ROPs
55296 : 3456 : 0
Memory Size
320 GB
Memory Bus
40960 bit
NVIDIA DGX A100
Overview
Manufacturer
NVIDIA
Original Series
Data Center Ampere
Announcement Date
May 14th, 2020
Release Date
May 14th, 2020
Launch Price
$199,000 USD
Graphics Processing Unit
GPU Model
8× GA100
Architecture
Ampere
Fabrication Process
7 nm FinFET
Die Size
8× 826 mm2
Transistors Count
8× 54.2B
CUDAs
8× 6912 (55296)
Tensor Cores
8× 432 (3456)
TMUs
8× 432 (3456)
Clocks
Base Clock
1410 MHz
Boost Clock
1410 MHz
Memory Clock
1215 MHz
Effective Memory Clock
2430 Mbps
Memory Configuration
Memory Size
8× 40960 (327680) MB
Memory Type
HBM2
Memory Bus Width
8× 5120 (40960)-bit
Memory Bandwidth
8× 1,555.2 (8) GB/s
Physical
Interface
SXM 4.0
Power Connectors
-
TDP/TBP
6500 W
API Support
DirectX
12.2
Vulkan
1.2
OpenGL
4.6
OpenCL
2.1
Performance
Texture Fillrate
4.9 TTexel/s
FP16 Performance
623.7 TFLOPS
FP32 Performance
155.9 TFLOPS
FP64 Performance
78 TFLOPS
FP32 Perf. per Watt
24 GFLOPS/W
FP32 Perf. per mm2
23.6 GFLOPS/mm2
| Model | Cores | Boost Clock | Memory Clock | Memory Config. | |
|---|---|---|---|---|---|
| NVIDIA HGX A100 (16-GPU) | |||||
| NVIDIA DGX A100 | |||||
| NVIDIA HGX A100 (8-GPU) | |||||
| NVIDIA HGX A100 (4-GPU) | |||||
| NVIDIA A40 | TBC | TBC | TBC | TBC | |
 | NVIDIA A100 PCIe | ||||
 | NVIDIA A100 SXM | ||||
| NVIDIA EGX A100 | TBC | TBC | TBC | TBC |
| Model | Cores | Boost Clock | Memory Clock | Memory Config. | |
|---|---|---|---|---|---|
 | NVIDIA HGX A100 (16-GPU) | ||||
| | NVIDIA DGX A100 | ||||
| | NVIDIA HGX A100 (8-GPU) | ||||
| | NVIDIA HGX A100 (4-GPU) | ||||
| | NVIDIA A100 PCIe | ||||
| | NVIDIA A100 SXM | ||||
| | NVIDIA EGX A100 | TBC | TBC | TBC | TBC |