GPU model | G96 | GA106 |
---|---|---|
GPU variant | NB9E-GE | GA106-150-A1 |
Architecture | Tesla | Ampere |
Foundry | UMC | Samsung |
Manufacturing process | 65 nm | 8 nm |
Number of transistors | 314 million | 13,250 million |
Die size | 144 mm² | 276 mm² |
Release date | Jul 29th, 2008 | — |
---|---|---|
Generation | GeForce 9M (9700M) | — |
Production | End-of-life | — |
Interface | MXM-II | — |
Predecessor | GeForce 8M | — |
Successor | GeForce 100M | — |
GPU clock | 625 MHz | — |
---|---|---|
Shader clock | 1550 MHz | — |
Memory clock | 800 MHz 1600 Mbps effective | 1750 MHz 14 Gbps effective |
Base clock | — | 1552 MHz |
Boost clock | — | 1777 MHz |
Memory size | 512 MB | 8 GB |
---|---|---|
Memory type | GDDR3 | GDDR6 |
Memory bus width | 128 bit | 128 bit |
Bandwidth | 25.60 GB/s | 224.0 GB/s |
Shading units | 32 | 2560 |
---|---|---|
TMUs | 16 | 80 |
ROPs | 8 | 48 |
SM count | 4 | 20 |
Cache L2 | 32 KB | 2 MB |
Tensor cores | — | 80 |
RT cores | — | 20 |
Cache L1 | — | 128 KB (per SM) |
Pixel rate | 5.000 GPixel/s | 85.30 GPixel/s |
---|---|---|
Texture rate | 10.00 GTexel/s | 142.2 GTexel/s |
FP32 (float) performance | 99.20 GFLOPS | 9.098 TFLOPS |
FP16 (half) performance | — | 9.098 TFLOPS (1:1) |
FP64 (double) performance | — | 142.2 GFLOPS (1:64) |
Slot width | MXM Module | Dual-slot |
---|---|---|
TDP | 45 W | 90 W |
Outputs | No outputs | 1x HDMI3x DisplayPort |
Length | — | 242 mm 9.5 inches |
Width | — | 112 mm 4.4 inches |
Suggested PSU | — | 250 W |
Power connectors | — | 1x 6-pin |
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 | 8.6 |
Shader model | 4.0 | 6.6 |
Release date | — | 2022 |
---|---|---|
gpu.details.availability | — | 2022 |
Generation | — | GeForce 30 |
Predecessor | — | GeForce 20 |
Production | — | Unreleased |
Interface | — | PCIe 4.0 x16 |