GPU model | Polaris 22 | NV18 |
---|---|---|
GPU variant | VegaM MGL XL | — |
Architecture | GCN 4.0 | Celsius |
Foundry | GlobalFoundries | TSMC |
Manufacturing process | 14 nm | 150 nm |
Number of transistors | 5,000 million | 29 million |
Die size | 208 mm² | 65 mm² |
Release date | Apr 24th, 2018 | — |
---|---|---|
Generation | Vega (Vega M) | — |
Production | Active | — |
Interface | IGP | — |
Base clock | 931 MHz | — |
---|---|---|
Boost clock | 1011 MHz | — |
Memory clock | 700 MHz 1400 Mbps effective | 200 MHz 400 Mbps effective |
GPU clock | — | 300 MHz |
Memory size | 4 GB | 64 MB |
---|---|---|
Memory type | HBM2 | DDR |
Memory bus width | 1024 bit | 128 bit |
Bandwidth | 179.2 GB/s | 6.400 GB/s |
Shading units | 1280 | — |
---|---|---|
TMUs | 80 | 4 |
ROPs | 32 | 4 |
Compute units | 20 | — |
Cache L1 | 16 KB (per CU) | — |
Cache L2 | 1024 KB | — |
Pixel shaders | — | 2 |
Vertex shaders | — | 0 |
Pixel rate | 32.35 GPixel/s | 1.200 GPixel/s |
---|---|---|
Texture rate | 80.88 GTexel/s | 1.200 GTexel/s |
FP16 (half) performance | 2.588 TFLOPS (1:1) | — |
FP32 (float) performance | 2.588 TFLOPS | — |
FP64 (double) performance | 161.8 GFLOPS (1:16) | — |
Slot width | IGP | Single-slot |
---|---|---|
TDP | 65 W | unknown |
Outputs | No outputs | 1x LFH60 |
Board number | D136 | — |
Suggested PSU | — | 200 W |
Power connectors | — | None |
DirectX | 12 (12_0) | 7.0 |
---|---|---|
OpenGL | 4.6 | 1.5 |
OpenCL | 2.1 | — |
Vulkan | 1.2 | — |
Shader model | 6.4 | — |
Pixel shader | — | 1.1 |
Vertex shader | — | 1.1 |
Release date | — | Nov 12th, 2002 |
---|---|---|
Generation | — | Quadro4 XGL |
Production | — | End-of-life |
Interface | — | AGP 8x |
Reviews | — | 26 in our database |