GPU model | GA102 | GA104 |
---|---|---|
GPU variant | GA102-890-A1 | GA104-875-A1 |
Architecture | Ampere | Ampere |
Foundry | Samsung | Samsung |
Manufacturing process | 8 nm | 8 nm |
Number of transistors | 28,300 million | 17,400 million |
Die size | 628 mm² | 392 mm² |
Release date | Apr 12th, 2021 | Apr 12th, 2021 |
---|---|---|
gpu.details.availability | 2021 | — |
Generation | Tesla | Quadro |
Production | Active | Active |
Interface | PCIe 4.0 x16 | PCIe 4.0 x16 |
Base clock | 885 MHz | 735 MHz |
---|---|---|
Boost clock | 1695 MHz | 1560 MHz |
Memory clock | 1563 MHz 12.5 Gbps effective | 1750 MHz 14 Gbps effective |
Memory size | 24 GB | 16 GB |
---|---|---|
Memory type | GDDR6 | GDDR6 |
Memory bus width | 384 bit | 256 bit |
Bandwidth | 600.2 GB/s | 448.0 GB/s |
Shading units | 9216 | 6144 |
---|---|---|
TMUs | 288 | 192 |
ROPs | 96 | 96 |
SM count | 72 | 48 |
Tensor cores | 288 | 192 |
RT cores | 72 | 48 |
Cache L1 | 128 KB (per SM) | 128 KB (per SM) |
Cache L2 | 6 MB | 4 MB |
Pixel rate | 162.7 GPixel/s | 149.8 GPixel/s |
---|---|---|
Texture rate | 488.2 GTexel/s | 299.5 GTexel/s |
FP16 (half) performance | 31.24 TFLOPS (1:1) | 19.17 TFLOPS (1:1) |
FP32 (float) performance | 31.24 TFLOPS | 19.17 TFLOPS |
FP64 (double) performance | 976.3 GFLOPS (1:32) | 599.0 GFLOPS (1:32) |
Slot width | Single-slot | Single-slot |
---|---|---|
Length | 267 mm 10.5 inches | 241 mm 9.5 inches |
Width | 112 mm 4.4 inches | 112 mm 4.4 inches |
TDP | 150 W | 140 W |
Suggested PSU | 450 W | 300 W |
Outputs | No outputs | 4x DisplayPort |
Power connectors | 8-pin EPS | 1x 6-pin |
Board number | PG133 SKU 215 | — |
DirectX | 12 Ultimate (12_2) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 3.0 | 3.0 |
Vulkan | 1.2 | 1.2 |
CUDA | 8.6 | 8.6 |
Shader model | 6.6 | 6.6 |