AMD Radeon Instinct MI8 vs NVIDIA GeForce GTX 980 Ti
Graphics Processor
| GPU model | Fiji | GM200 |
|---|---|---|
| GPU variant | Fiji XT | GM200-310-A1 |
| Architecture | GCN 3.0 | Maxwell 2.0 |
| Foundry | TSMC | TSMC |
| Manufacturing process | 28 nm | 28 nm |
| Number of transistors | 8,900 million | 8,000 million |
| Die size | 596 mm² | 601 mm² |
Graphics Card
| Release date | Dec 12th, 2016 | Jun 2nd, 2015 |
|---|---|---|
| Generation | Radeon Instinct | GeForce 900 |
| Production | End-of-life | End-of-life |
| Interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
| Launch price | — | 649 USD |
| Reviews | — | 169 in our database |
| Predecessor | — | GeForce 700 |
| Successor | — | GeForce 10 |
Clocks
| GPU clock | 1000 MHz | — |
|---|---|---|
| Memory clock | 500 MHz 1000 Mbps effective | 1753 MHz 7 Gbps effective |
| Base clock | — | 1000 MHz |
| Boost clock | — | 1076 MHz |
Memory Configuration
| Memory size | 4 GB | 6 GB |
|---|---|---|
| Memory type | HBM | GDDR5 |
| Memory bus width | 4096 bit | 384 bit |
| Bandwidth | 512.0 GB/s | 336.6 GB/s |
Render Configuration
| Shading units | 4096 | 2816 |
|---|---|---|
| TMUs | 256 | 176 |
| ROPs | 64 | 96 |
| Compute units | 64 | — |
| Cache L1 | 16 KB (per CU) | 48 KB (per SMM) |
| Cache L2 | 2 MB | 3 MB |
| SMM count | — | 22 |
Performance
| Pixel rate | 64.00 GPixel/s | 103.3 GPixel/s |
|---|---|---|
| Texture rate | 256.0 GTexel/s | 189.4 GTexel/s |
| FP16 (half) performance | 8.192 TFLOPS (1:1) | — |
| FP32 (float) performance | 8.192 TFLOPS | 6.060 TFLOPS |
| FP64 (double) performance | 512.0 GFLOPS (1:16) | 189.4 GFLOPS (1:32) |
Dimensions & Outputs
| Slot width | Dual-slot | Dual-slot |
|---|---|---|
| Length | 152 mm 6 inches | 267 mm 10.5 inches |
| Width | 111 mm 4.4 inches | 111 mm 4.4 inches |
| TDP | 175 W | 250 W |
| Suggested PSU | 450 W | 600 W |
| Outputs | No outputs | 1x DVI1x HDMI3x DisplayPort |
| Power connectors | 1x 8-pin | 1x 6-pin + 1x 8-pin |
| Height | — | 40 mm 1.6 inches |
| Board number | — | PG600 SKU 30 |
API Support & Features
| DirectX | 12 (12_0) | 12 (12_1) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 2.0 | 3.0 |
| Vulkan | 1.2 | 1.1 |
| Shader model | 6.3 | 6.4 |
| CUDA | — | 5.2 |
