GPU model | Vega 10 | GA106 |
---|---|---|
GPU variant | Vega 10 XL (215-0894216) | — |
Architecture | GCN 5.0 | Ampere |
Foundry | GlobalFoundries | Samsung |
Manufacturing process | 14 nm | 8 nm |
Number of transistors | 12,500 million | 13,250 million |
Die size | 495 mm² | 276 mm² |
Release date | Aug 14th, 2017 | Apr 12th, 2021 |
---|---|---|
Generation | Radeon Pro Mac (Vega Series) | Quadro Mobile (Ax000) |
Production | Active | Active |
Interface | PCIe 3.0 x16 | PCIe 4.0 x16 |
Base clock | 1138 MHz | 893 MHz |
---|---|---|
Boost clock | 1250 MHz | 1358 MHz |
Memory clock | 786 MHz 1572 Mbps effective | 1375 MHz 11 Gbps effective |
Memory size | 8 GB | 4 GB |
---|---|---|
Memory type | HBM2 | GDDR6 |
Memory bus width | 2048 bit | 128 bit |
Bandwidth | 402.4 GB/s | 176.0 GB/s |
Shading units | 3584 | 2560 |
---|---|---|
TMUs | 224 | 80 |
ROPs | 64 | 48 |
Compute units | 56 | — |
Cache L1 | 16 KB (per CU) | 128 KB (per SM) |
Cache L2 | 4 MB | 2 MB |
SM count | — | 20 |
Tensor cores | — | 80 |
RT cores | — | 20 |
Pixel rate | 80.00 GPixel/s | 65.18 GPixel/s |
---|---|---|
Texture rate | 280.0 GTexel/s | 108.6 GTexel/s |
FP16 (half) performance | 17.92 TFLOPS (2:1) | 6.953 TFLOPS (1:1) |
FP32 (float) performance | 8.960 TFLOPS | 6.953 TFLOPS |
FP64 (double) performance | 560.0 GFLOPS (1:16) | 108.6 GFLOPS (1:64) |
Slot width | IGP | IGP |
---|---|---|
TDP | 210 W | 95 W |
Outputs | 1x HDMI3x DisplayPort | No outputs |
Power connectors | None | None |
DirectX | 12 (12_1) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 2.1 | 3.0 |
Vulkan | 1.2 | 1.2 |
Shader model | 6.4 | 6.6 |
CUDA | — | 8.6 |