GPU model | Fiji | GA106 |
---|---|---|
GPU variant | Fiji XT CA (215-0862120) | GA106-300-A1 |
Architecture | GCN 3.0 | Ampere |
Foundry | TSMC | Samsung |
Manufacturing process | 28 nm | 8 nm |
Number of transistors | 8,900 million | 13,250 million |
Die size | 596 mm² | 276 mm² |
Release date | Aug 27th, 2015 | Jan 12th, 2021 |
---|---|---|
Generation | Pirate Islands | GeForce 30 |
Production | End-of-life | Active |
Launch price | 649 USD | 329 USD |
Interface | PCIe 3.0 x16 | PCIe 4.0 x16 |
Predecessor | Volcanic Islands | GeForce 20 |
Successor | Arctic Islands | — |
gpu.details.availability | — | Feb 25th, 2021 |
Reviews | — | 60 in our database |
GPU clock | 1000 MHz | — |
---|---|---|
Memory clock | 500 MHz 1000 Mbps effective | 1875 MHz 15 Gbps effective |
Base clock | — | 1320 MHz |
Boost clock | — | 1777 MHz |
Memory size | 4 GB | 12 GB |
---|---|---|
Memory type | HBM | GDDR6 |
Memory bus width | 4096 bit | 192 bit |
Bandwidth | 512.0 GB/s | 360.0 GB/s |
Shading units | 4096 | 3584 |
---|---|---|
TMUs | 256 | 112 |
ROPs | 64 | 48 |
Compute units | 64 | — |
Cache L1 | 16 KB (per CU) | 128 KB (per SM) |
Cache L2 | 2 MB | 3 MB |
SM count | — | 28 |
Tensor cores | — | 112 |
RT cores | — | 28 |
Pixel rate | 64.00 GPixel/s | 85.30 GPixel/s |
---|---|---|
Texture rate | 256.0 GTexel/s | 199.0 GTexel/s |
FP16 (half) performance | 8.192 TFLOPS (1:1) | 12.74 TFLOPS (1:1) |
FP32 (float) performance | 8.192 TFLOPS | 12.74 TFLOPS |
FP64 (double) performance | 512.0 GFLOPS (1:16) | 199.0 GFLOPS (1:64) |
Slot width | Dual-slot | Dual-slot |
---|---|---|
Length | 154 mm 6.1 inches | 242 mm 9.5 inches |
Width | 112 mm 4.4 inches | 112 mm 4.4 inches |
Height | 40 mm 1.6 inches | — |
TDP | 175 W | 170 W |
Suggested PSU | 450 W | 450 W |
Outputs | 1x HDMI3x DisplayPort | 1x HDMI3x DisplayPort |
Power connectors | 1x 8-pin | 1x 12-pin |
Board number | C882-37 | PG190 SKU 60 |
DirectX | 12 (12_0) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 2.0 | 3.0 |
Vulkan | 1.2 | 1.2 |
Shader model | 6.3 | 6.6 |
CUDA | — | 8.6 |