GPU model | G84 | TU106 |
---|---|---|
GPU variant | NB8P-GS | N19E-Q1-KA-K1 |
Architecture | Tesla | Turing |
Foundry | TSMC | TSMC |
Manufacturing process | 80 nm | 12 nm |
Number of transistors | 289 million | 10,800 million |
Die size | 169 mm² | 445 mm² |
Release date | Oct 5th, 2007 | May 27th, 2019 |
---|---|---|
Generation | GeForce 8M (8000M) | Quadro Mobile (Tx000) |
Production | End-of-life | Active |
Interface | MXM-II | PCIe 3.0 x16 |
Predecessor | GeForce Go 7 | — |
Successor | GeForce 9M | — |
GPU clock | 470 MHz | — |
---|---|---|
Shader clock | 940 MHz | — |
Memory clock | 635 MHz 1270 Mbps effective | 1750 MHz 14 Gbps effective |
Base clock | — | 945 MHz |
Boost clock | — | 1380 MHz |
Memory size | 512 MB | 6 GB |
---|---|---|
Memory type | GDDR3 | GDDR6 |
Memory bus width | 128 bit | 192 bit |
Bandwidth | 20.32 GB/s | 336.0 GB/s |
Shading units | 32 | 1920 |
---|---|---|
TMUs | 16 | 120 |
ROPs | 8 | 64 |
SM count | 4 | 30 |
Cache L2 | 32 KB | 3 MB |
Tensor cores | — | 240 |
RT cores | — | 30 |
Cache L1 | — | 64 KB (per SM) |
Pixel rate | 3.760 GPixel/s | 88.32 GPixel/s |
---|---|---|
Texture rate | 7.520 GTexel/s | 165.6 GTexel/s |
FP32 (float) performance | 60.16 GFLOPS | 5.299 TFLOPS |
FP16 (half) performance | — | 10.60 TFLOPS (2:1) |
FP64 (double) performance | — | 165.6 GFLOPS (1:32) |
Slot width | MXM Module | IGP |
---|---|---|
TDP | 20 W | 80 W |
Outputs | No outputs | No outputs |
Power connectors | — | None |
DirectX | 11.1 (10_0) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 3.3 | 4.6 |
OpenCL | 1.1 | 3.0 |
Vulkan | — | 1.2 |
CUDA | 1.1 | 7.5 |
Shader model | 4.0 | 6.6 |