GPU model | Barts | GA102 |
---|---|---|
GPU variant | Barts PRO (215-0798006) | GA102-200-KD-A1 |
Architecture | TeraScale 2 | Ampere |
Foundry | TSMC | Samsung |
Manufacturing process | 40 nm | 8 nm |
Number of transistors | 1,700 million | 28,300 million |
Die size | 255 mm² | 628 mm² |
Release date | Oct 21st, 2010 | Sep 1st, 2020 |
---|---|---|
Generation | Northern Islands | GeForce 30 |
Production | End-of-life | Active |
Launch price | 179 USD | 699 USD |
Interface | PCIe 2.0 x16 | PCIe 4.0 x16 |
Reviews | 105 in our database | 99 in our database |
Predecessor | Evergreen | GeForce 20 |
Successor | Southern Islands | — |
gpu.details.availability | — | Sep 17th, 2020 |
Current price | — | Amazon / Newegg |
GPU clock | 775 MHz | — |
---|---|---|
Memory clock | 1000 MHz 4 Gbps effective | 1188 MHz 19 Gbps effective |
Base clock | — | 1440 MHz |
Boost clock | — | 1710 MHz |
Memory size | 1024 MB | 10 GB |
---|---|---|
Memory type | GDDR5 | GDDR6X |
Memory bus width | 256 bit | 320 bit |
Bandwidth | 128.0 GB/s | 760.3 GB/s |
Shading units | 960 | 8704 |
---|---|---|
TMUs | 48 | 272 |
ROPs | 32 | 96 |
Compute units | 12 | — |
Cache L1 | 8 KB (per CU) | 128 KB (per SM) |
Cache L2 | 512 KB | 5 MB |
SM count | — | 68 |
Tensor cores | — | 272 |
RT cores | — | 68 |
Pixel rate | 24.80 GPixel/s | 164.2 GPixel/s |
---|---|---|
Texture rate | 37.20 GTexel/s | 465.1 GTexel/s |
FP32 (float) performance | 1,488 GFLOPS | 29.77 TFLOPS |
FP16 (half) performance | — | 29.77 TFLOPS (1:1) |
FP64 (double) performance | — | 465.1 GFLOPS (1:64) |
Slot width | Dual-slot | Dual-slot |
---|---|---|
Length | 198 mm 7.8 inches | 285 mm 11.2 inches |
TDP | 127 W | 320 W |
Suggested PSU | 300 W | 700 W |
Outputs | 2x DVI1x HDMI2x mini-DisplayPort | 1x HDMI3x DisplayPort |
Power connectors | 1x 6-pin | 1x 12-pin |
Board number | C223 | PG132 SKU 30 |
Width | — | 112 mm 4.4 inches |
Height | — | 40 mm 1.6 inches |
DirectX | 11.2 (11_0) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.4 | 4.6 |
OpenCL | 1.2 | 3.0 |
Vulkan | — | 1.2 |
Shader model | 5.0 | 6.6 |
CUDA | — | 8.6 |