GPU model | GA102 | GA102 |
---|---|---|
GPU variant | GA102-300-A1 | — |
Architecture | Ampere | Ampere |
Foundry | TSMC | Samsung |
Manufacturing process | 7 nm | 8 nm |
Number of transistors | 40,000 million | 28,300 million |
Die size | 627 mm² | 628 mm² |
Release date | 2020 | Nov 23rd, 2021 |
---|---|---|
Generation | GeForce 30 | Quadro |
Production | Unreleased | Active |
Launch price | 1,399 USD | — |
Interface | PCIe 4.0 x16 | PCIe 4.0 x16 |
Base clock | 1410 MHz | 1050 MHz |
---|---|---|
Boost clock | 1740 MHz | 1650 MHz |
Memory clock | 2375 MHz 19000 MHz effective | 2000 MHz 16 Gbps effective |
Memory size | 12 GB | 20 GB |
---|---|---|
Memory type | GDDR6X | GDDR6 |
Memory bus width | 384 bit | 320 bit |
Bandwidth | 912.0 GB/s | 640.0 GB/s |
Shading units | 5120 | 7168 |
---|---|---|
TMUs | 320 | 224 |
ROPs | 96 | 96 |
SM count | 80 | 56 |
Tensor cores | 640 | 224 |
RT cores | 160 | 56 |
Cache L1 | 192 KB (per SM) | 128 KB (per SM) |
Cache L2 | 6 MB | 6 MB |
Pixel rate | 167.0 GPixel/s | 158.4 GPixel/s |
---|---|---|
Texture rate | 556.8 GTexel/s | 369.6 GTexel/s |
FP16 (half) performance | 35.64 TFLOPS (2:1) | 23.65 TFLOPS (1:1) |
FP32 (float) performance | 17.82 TFLOPS | 23.65 TFLOPS |
FP64 (double) performance | 556.8 GFLOPS (1:32) | 739.2 GFLOPS (1:32) |
Slot width | Dual-slot | Dual-slot |
---|---|---|
Length | 267 mm 10.5 inches | 267 mm 10.5 inches |
TDP | 320 W | 200 W |
Suggested PSU | 700 W | 550 W |
Outputs | 1x HDMI3x DisplayPort1x USB Type-C | 4x DisplayPort |
Power connectors | 1x 12-pin | 1x 8-pin |
Board number | PG132 SKU 20 | — |
Width | — | 112 mm 4.4 inches |
DirectX | 12 Ultimate (12_2) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 2.0 | 3.0 |
Vulkan | 1.2.140 | 1.2 |
CUDA | 8.5 | 8.6 |
Shader model | 6.5 | 6.6 |