GPU model | GA100 | GK20A |
---|---|---|
Architecture | Ampere | Kepler 2.0 |
Foundry | TSMC | TSMC |
Manufacturing process | 7 nm | 28 nm |
Number of transistors | 54,200 million | unknown |
Die size | 826 mm² | unknown |
Release date | Jun 28th, 2021 | — |
---|---|---|
Generation | Tesla | — |
Production | Active | — |
Interface | PCIe 4.0 x16 | — |
Base clock | 1065 MHz | 756 MHz |
---|---|---|
Boost clock | 1410 MHz | 951 MHz |
Memory clock | 1593 MHz 3.2 Gbps effective | System Shared |
Memory size | 80 GB | System Shared |
---|---|---|
Memory type | HBM2e | System Shared |
Memory bus width | 5120 bit | System Shared |
Bandwidth | 2,039 GB/s | System Dependent |
Shading units | 6912 | 192 |
---|---|---|
TMUs | 432 | 8 |
ROPs | 160 | 4 |
SM count | 108 | — |
Tensor cores | 432 | — |
Cache L1 | 192 KB (per SM) | 16 KB (per SMX) |
Cache L2 | 80 MB | 128 KB |
SMX count | — | 1 |
Pixel rate | 225.6 GPixel/s | 3.804 GPixel/s |
---|---|---|
Texture rate | 609.1 GTexel/s | 7.608 GTexel/s |
FP16 (half) performance | 77.97 TFLOPS (4:1) | — |
FP32 (float) performance | 19.49 TFLOPS | 365.2 GFLOPS |
FP64 (double) performance | 9.746 TFLOPS (1:2) | 11.41 GFLOPS (1:32) |
Slot width | Dual-slot | IGP |
---|---|---|
Length | 267 mm 10.5 inches | — |
TDP | 250 W | 8 W |
Suggested PSU | 600 W | — |
Outputs | No outputs | No outputs |
Power connectors | 8-pin EPS | — |
Board number | P1001 SKU 200 | — |
DirectX | — | 12 (11_0) |
---|---|---|
OpenGL | — | ES 3.1 |
OpenCL | 3.0 | — |
Vulkan | — | 1.1 |
CUDA | 8.0 | 3.2 |
Shader model | — | 5.1 |
Release date | — | Oct 15th, 2014 |
---|---|---|
Generation | — | Tegra |
Production | — | End-of-life |
Interface | — | IGP |