GPU model | Polaris 20 | GV100 |
---|---|---|
GPU variant | Polaris 20M XT | — |
Architecture | GCN 4.0 | Volta |
Foundry | GlobalFoundries | TSMC |
Manufacturing process | 14 nm | 12 nm |
Number of transistors | 5,700 million | 21,100 million |
Die size | 232 mm² | 815 mm² |
Release date | Apr 11th, 2018 | — |
---|---|---|
Generation | Mobility Radeon (RX M500X) | — |
Production | Active | — |
Interface | MXM-B (3.0) | — |
Predecessor | Crystal System | — |
Base clock | 1000 MHz | 1245 MHz |
---|---|---|
Boost clock | 1077 MHz | 1380 MHz |
Memory clock | 2000 MHz 8 Gbps effective | 876 MHz 1752 Mbps effective |
Memory size | 8 GB | 16 GB |
---|---|---|
Memory type | GDDR5 | HBM2 |
Memory bus width | 256 bit | 4096 bit |
Bandwidth | 256.0 GB/s | 897.0 GB/s |
Shading units | 2304 | 5120 |
---|---|---|
TMUs | 144 | 320 |
ROPs | 32 | 128 |
Compute units | 36 | — |
Cache L1 | 16 KB (per CU) | 128 KB (per SM) |
Cache L2 | 2 MB | 6 MB |
SM count | — | 80 |
Tensor cores | — | 640 |
Pixel rate | 34.46 GPixel/s | 176.6 GPixel/s |
---|---|---|
Texture rate | 155.1 GTexel/s | 441.6 GTexel/s |
FP16 (half) performance | 4.963 TFLOPS (1:1) | 28.26 TFLOPS (2:1) |
FP32 (float) performance | 4.963 TFLOPS | 14.13 TFLOPS |
FP64 (double) performance | 310.2 GFLOPS (1:16) | 7.066 TFLOPS (1:2) |
Slot width | MXM Module | Dual-slot |
---|---|---|
TDP | 100 W | 300 W |
Outputs | No outputs | No outputs |
Power connectors | None | 2x 8-pin |
Suggested PSU | — | 700 W |
Board number | — | PG500 SKU 200 |
DirectX | 12 (12_0) | 12 (12_1) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 2.1 | 3.0 |
Vulkan | 1.2 | 1.2 |
Shader model | 6.4 | 6.6 |
CUDA | — | 7.0 |
Release date | — | Jun 21st, 2017 |
---|---|---|
Generation | — | Tesla |
Production | — | End-of-life |
Interface | — | PCIe 3.0 x16 |