AMD Radeon Instinct MI50 vs NVIDIA CMP 40HX
Grafikprozessor
| GPU Name | Vega 20 | TU106 |
|---|---|---|
| GPU Variante | Vega 20 GLXT | TU106-100-A1 |
| Architektur | GCN 5.1 | Turing |
| Hersteller | TSMC | TSMC |
| Prozessgröße | 7 nm | 12 nm |
| Transistoren | 13,230 million | 10,800 million |
| Chipgröße | 331 mm² | 445 mm² |
Grafikkarte
| Veröffentlichungsdatum | Nov 18th, 2018 | Feb 25th, 2021 |
|---|---|---|
| Generation | Radeon Instinct | Mining GPUs |
| Produktion | Active | Active |
| Businterface | PCIe 4.0 x16 | PCIe 3.0 x16 |
| Originalpreis | — | 699 USD |
Takt-Geschwindigkeiten
| Basis-Taktfrequenz | 1200 MHz | 1470 MHz |
|---|---|---|
| Boost-Takt | 1746 MHz | 1650 MHz |
| Speicher-Taktung | 1000 MHz 2 Gbps effective | 1750 MHz 14 Gbps effective |
Speicher
| Speichergröße | 16 GB | 8 GB |
|---|---|---|
| Speichertyp | HBM2 | GDDR6 |
| Speicherbus | 4096 bit | 256 bit |
| Bandbreite | 1,024 GB/s | 448.0 GB/s |
Render-Konfiguration
| Shading-Einheiten | 3840 | 2304 |
|---|---|---|
| TMUs | 240 | 144 |
| ROPs | 64 | 64 |
| Compute Units | 60 | — |
| Cache L1 | 16 KB (per CU) | 64 KB (per SM) |
| Cache L2 | 4 MB | 4 MB |
| Anzahl SM | — | 36 |
| Tensor Cores | — | 288 |
| RT-Kerne | — | 36 |
Theoretische Leistung
| Pixelrate | 111.7 GPixel/s | 105.6 GPixel/s |
|---|---|---|
| Texturrate | 419.0 GTexel/s | 237.6 GTexel/s |
| FP16 (half) Leistung | 26.82 TFLOPS (2:1) | 15.21 TFLOPS (2:1) |
| FP32 (float) Leistung | 13.41 TFLOPS | 7.603 TFLOPS |
| FP64 (double) Leistung | 6.705 TFLOPS (1:2) | 237.6 GFLOPS (1:32) |
Boarddesign
| Steckplatzbreite | Dual-slot | Dual-slot |
|---|---|---|
| Länge | 267 mm 10.5 inches | 229 mm 9 inches |
| TDP | 300 W | 185 W |
| Vorgeschlagenes Netzteil | 700 W | 450 W |
| Ausgabe | 1x mini-DisplayPort | No outputs |
| Stromverbindung | 1x 6-pin + 1x 8-pin | 1x 8-pin |
| Breite | — | 111 mm 4.4 inches |
| Höhe | — | 35 mm 1.4 inches |
| Board-Nummer | — | PG161 SKU 100 |
Grafik-Features
| DirectX | 12 (12_1) | 12 Ultimate (12_2) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 2.1 | 3.0 |
| Vulkan | 1.2 | 1.2 |
| Shader-Modell | 6.4 | 6.6 |
| CUDA | — | 7.5 |
