GPU model | GP104 | TU104 |
---|---|---|
GPU variant | GP104-400-A1 | TU104-875-A1 |
Architecture | Pascal | Turing |
Foundry | TSMC | TSMC |
Manufacturing process | 16 nm | 12 nm |
Number of transistors | 7,200 million | 13,600 million |
Die size | 314 mm² | 545 mm² |
Release date | May 27th, 2016 | Aug 13th, 2018 |
---|---|---|
Generation | GeForce 10 | Quadro |
Predecessor | GeForce 900 | — |
Successor | GeForce 20 | — |
Production | End-of-life | Active |
Launch price | 599 USD | 2,299 USD |
Interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Reviews | 183 in our database | — |
Base clock | 1607 MHz | 1620 MHz |
---|---|---|
Boost clock | 1733 MHz | 1815 MHz |
Memory clock | 1251 MHz 10 Gbps effective | 1750 MHz 14 Gbps effective |
Memory size | 8 GB | 16 GB |
---|---|---|
Memory type | GDDR5X | GDDR6 |
Memory bus width | 256 bit | 256 bit |
Bandwidth | 320.3 GB/s | 448.0 GB/s |
Shading units | 2560 | 3072 |
---|---|---|
TMUs | 160 | 192 |
ROPs | 64 | 64 |
SM count | 20 | 48 |
Cache L1 | 48 KB (per SM) | 64 KB (per SM) |
Cache L2 | 2 MB | 4 MB |
Tensor cores | — | 384 |
RT cores | — | 48 |
Pixel rate | 110.9 GPixel/s | 116.2 GPixel/s |
---|---|---|
Texture rate | 277.3 GTexel/s | 348.5 GTexel/s |
FP16 (half) performance | 138.6 GFLOPS (1:64) | 22.30 TFLOPS (2:1) |
FP32 (float) performance | 8.873 TFLOPS | 11.15 TFLOPS |
FP64 (double) performance | 277.3 GFLOPS (1:32) | 348.5 GFLOPS (1:32) |
Slot width | Dual-slot | Dual-slot |
---|---|---|
Length | 267 mm 10.5 inches | 267 mm 10.5 inches |
Width | 112 mm 4.4 inches | 111 mm 4.4 inches |
Height | 40 mm 1.6 inches | — |
TDP | 180 W | 230 W |
Suggested PSU | 450 W | 550 W |
Outputs | 1x DVI1x HDMI3x DisplayPort | 4x DisplayPort1x USB Type-C |
Power connectors | 1x 8-pin | 1x 6-pin + 1x 8-pin |
Board number | PG413 SKU 0 | — |
DirectX | 12 (12_1) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 3.0 | 3.0 |
Vulkan | 1.2 | 1.2 |
CUDA | 6.1 | 7.5 |
Shader model | 6.4 | 6.6 |