GPU model | TU117 | GP107 |
---|---|---|
GPU variant | TU117-300-A1 | — |
Architecture | Turing | Pascal |
Foundry | TSMC | Samsung |
Manufacturing process | 12 nm | 14 nm |
Number of transistors | 4,700 million | 3,300 million |
Die size | 200 mm² | 132 mm² |
Release date | Apr 23rd, 2019 | Feb 7th, 2017 |
---|---|---|
Generation | GeForce 16 | Quadro |
Predecessor | GeForce 10 | — |
Successor | GeForce 20 | — |
Production | Active | End-of-life |
Launch price | 149 USD | — |
Interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Reviews | 45 in our database | — |
Base clock | 1485 MHz | 1329 MHz |
---|---|---|
Boost clock | 1665 MHz | 1557 MHz |
Memory clock | 2001 MHz 8 Gbps effective | 1002 MHz 4 Gbps effective |
Memory size | 4 GB | 2 GB |
---|---|---|
Memory type | GDDR5 | GDDR5 |
Memory bus width | 128 bit | 128 bit |
Bandwidth | 128.1 GB/s | 64.13 GB/s |
Shading units | 896 | 384 |
---|---|---|
TMUs | 56 | 24 |
ROPs | 32 | 16 |
SM count | 14 | 3 |
Cache L1 | 64 KB (per SM) | 48 KB (per SM) |
Cache L2 | 1024 KB | 1024 KB |
Pixel rate | 53.28 GPixel/s | 24.91 GPixel/s |
---|---|---|
Texture rate | 93.24 GTexel/s | 37.37 GTexel/s |
FP16 (half) performance | 5.967 TFLOPS (2:1) | 18.68 GFLOPS (1:64) |
FP32 (float) performance | 2.984 TFLOPS | 1,196 GFLOPS |
FP64 (double) performance | 93.24 GFLOPS (1:32) | 37.37 GFLOPS (1:32) |
Slot width | Dual-slot | Single-slot |
---|---|---|
Length | 229 mm 9 inches | 150 mm 5.9 inches |
Width | 111 mm 4.4 inches | 69 mm 2.7 inches |
Height | 35 mm 1.4 inches | — |
TDP | 75 W | 40 W |
Suggested PSU | 250 W | 200 W |
Outputs | 1x DVI1x HDMI1x DisplayPort | 4x mini-DisplayPort |
Power connectors | None | None |
Board number | PG174 SKU 0 | PG212 SKU 501 |
DirectX | 12 (12_1) | 12 (12_1) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 3.0 | 3.0 |
Vulkan | 1.2 | 1.2 |
CUDA | 7.5 | 6.1 |
Shader model | 6.6 | 6.4 |