GPU model | TU102 | GF106 |
---|---|---|
GPU variant | TU102-875-A1 | N12P-Q3 |
Architecture | Turing | Fermi |
Foundry | TSMC | TSMC |
Manufacturing process | 12 nm | 40 nm |
Number of transistors | 18,600 million | 1,170 million |
Die size | 754 mm² | 238 mm² |
Release date | Unknown | — |
---|---|---|
Generation | GRID | — |
Production | Active | — |
Interface | PCIe 3.0 x16 | — |
Base clock | 1065 MHz | — |
---|---|---|
Boost clock | 1395 MHz | — |
Memory clock | 1750 MHz 14 Gbps effective | 900 MHz 1800 Mbps effective |
GPU clock | — | 550 MHz |
Shader clock | — | 1100 MHz |
Memory size | 4 GB | 2 GB |
---|---|---|
Memory type | GDDR6 | DDR3 |
Memory bus width | 384 bit | 128 bit |
Bandwidth | 672.0 GB/s | 28.80 GB/s |
Shading units | 4608 | 192 |
---|---|---|
TMUs | 288 | 32 |
ROPs | 96 | 16 |
SM count | 72 | 4 |
Tensor cores | 576 | — |
RT cores | 72 | — |
Cache L1 | 64 KB (per SM) | 64 KB (per SM) |
Cache L2 | 6 MB | 256 KB |
Pixel rate | 133.9 GPixel/s | 4.400 GPixel/s |
---|---|---|
Texture rate | 401.8 GTexel/s | 17.60 GTexel/s |
FP16 (half) performance | 25.71 TFLOPS (2:1) | — |
FP32 (float) performance | 12.86 TFLOPS | 422.4 GFLOPS |
FP64 (double) performance | 401.8 GFLOPS (1:32) | 35.20 GFLOPS (1:12) |
Slot width | Dual-slot | MXM Module |
---|---|---|
Length | 267 mm 10.5 inches | — |
TDP | 260 W | 55 W |
Suggested PSU | 600 W | — |
Outputs | No outputs | No outputs |
Power connectors | 1x 6-pin + 1x 8-pin | — |
Board number | PG150 SKU 215 | — |
DirectX | 12 Ultimate (12_2) | 12 (11_0) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 3.0 | 1.1 |
Vulkan | 1.2 | — |
CUDA | 7.5 | 2.1 |
Shader model | 6.6 | 5.1 |
Release date | — | Jan 13th, 2011 |
---|---|---|
Generation | — | Quadro Mobile (x000M) |
Production | — | End-of-life |
Interface | — | MXM-A (3.0) |
Reviews | — | 26 in our database |
Predecessor | — | Quadro FX Mobile |