GPU model | GA102 | TU104 |
---|---|---|
GPU variant | GA102-890-A1 | TU104-400A-A1 |
Architecture | Ampere | Turing |
Foundry | Samsung | TSMC |
Manufacturing process | 8 nm | 12 nm |
Number of transistors | 28,300 million | 13,600 million |
Die size | 628 mm² | 545 mm² |
Release date | Apr 12th, 2021 | Sep 20th, 2018 |
---|---|---|
gpu.details.availability | 2021 | — |
Generation | Tesla | GeForce 20 |
Production | Active | Active |
Interface | PCIe 4.0 x16 | PCIe 3.0 x16 |
Predecessor | — | GeForce 10 |
Launch price | — | 699 USD |
Current price | — | Amazon / Newegg |
Reviews | — | 66 in our database |
Successor | — | GeForce 30 |
Base clock | 885 MHz | 1515 MHz |
---|---|---|
Boost clock | 1695 MHz | 1710 MHz |
Memory clock | 1563 MHz 12.5 Gbps effective | 1750 MHz 14 Gbps effective |
Memory size | 24 GB | 8 GB |
---|---|---|
Memory type | GDDR6 | GDDR6 |
Memory bus width | 384 bit | 256 bit |
Bandwidth | 600.2 GB/s | 448.0 GB/s |
Shading units | 9216 | 2944 |
---|---|---|
TMUs | 288 | 184 |
ROPs | 96 | 64 |
SM count | 72 | 46 |
Tensor cores | 288 | 368 |
RT cores | 72 | 46 |
Cache L1 | 128 KB (per SM) | 64 KB (per SM) |
Cache L2 | 6 MB | 4 MB |
Pixel rate | 162.7 GPixel/s | 109.4 GPixel/s |
---|---|---|
Texture rate | 488.2 GTexel/s | 314.6 GTexel/s |
FP16 (half) performance | 31.24 TFLOPS (1:1) | 20.14 TFLOPS (2:1) |
FP32 (float) performance | 31.24 TFLOPS | 10.07 TFLOPS |
FP64 (double) performance | 976.3 GFLOPS (1:32) | 314.6 GFLOPS (1:32) |
Slot width | Single-slot | Dual-slot |
---|---|---|
Length | 267 mm 10.5 inches | 267 mm 10.5 inches |
Width | 112 mm 4.4 inches | 116 mm 4.6 inches |
TDP | 150 W | 215 W |
Suggested PSU | 450 W | 550 W |
Outputs | No outputs | 1x HDMI3x DisplayPort1x USB Type-C |
Power connectors | 8-pin EPS | 1x 6-pin + 1x 8-pin |
Board number | PG133 SKU 215 | PG180 SKU 2 |
Height | — | 35 mm 1.4 inches |
DirectX | 12 Ultimate (12_2) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 3.0 | 3.0 |
Vulkan | 1.2 | 1.2 |
CUDA | 8.6 | 7.5 |
Shader model | 6.6 | 6.6 |