AMD Radeon Instinct MI25 vs NVIDIA Tesla T4
Grafikprozessor
| GPU Name | Vega 10 | TU104 |
|---|---|---|
| GPU Variante | Vega 10 XT GL (215-0894124) | TU104-895-A1 |
| Architektur | GCN 5.0 | Turing |
| Hersteller | GlobalFoundries | TSMC |
| Prozessgröße | 14 nm | 12 nm |
| Transistoren | 12,500 million | 13,600 million |
| Chipgröße | 495 mm² | 545 mm² |
Grafikkarte
| Veröffentlichungsdatum | Jun 27th, 2017 | Sep 13th, 2018 |
|---|---|---|
| Generation | Radeon Instinct | Tesla |
| Produktion | Active | Active |
| Businterface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Takt-Geschwindigkeiten
| Basis-Taktfrequenz | 1400 MHz | 585 MHz |
|---|---|---|
| Boost-Takt | 1500 MHz | 1590 MHz |
| Speicher-Taktung | 852 MHz 1704 Mbps effective | 1250 MHz 10 Gbps effective |
Speicher
| Speichergröße | 16 GB | 16 GB |
|---|---|---|
| Speichertyp | HBM2 | GDDR6 |
| Speicherbus | 2048 bit | 256 bit |
| Bandbreite | 436.2 GB/s | 320.0 GB/s |
Render-Konfiguration
| Shading-Einheiten | 4096 | 2560 |
|---|---|---|
| TMUs | 256 | 160 |
| ROPs | 64 | 64 |
| Compute Units | 64 | — |
| Cache L1 | 16 KB (per CU) | 64 KB (per SM) |
| Cache L2 | 4 MB | 4 MB |
| Anzahl SM | — | 40 |
| Tensor Cores | — | 320 |
| RT-Kerne | — | 40 |
Theoretische Leistung
| Pixelrate | 96.00 GPixel/s | 101.8 GPixel/s |
|---|---|---|
| Texturrate | 384.0 GTexel/s | 254.4 GTexel/s |
| FP16 (half) Leistung | 24.58 TFLOPS (2:1) | 65.13 TFLOPS (8:1) |
| FP32 (float) Leistung | 12.29 TFLOPS | 8.141 TFLOPS |
| FP64 (double) Leistung | 768.0 GFLOPS (1:16) | 254.4 GFLOPS (1:32) |
Boarddesign
| Steckplatzbreite | Dual-slot | Single-slot |
|---|---|---|
| Länge | 267 mm 10.5 inches | 168 mm 6.6 inches |
| Breite | 111 mm 4.4 inches | — |
| TDP | 300 W | 70 W |
| Vorgeschlagenes Netzteil | 700 W | 250 W |
| Ausgabe | No outputs | No outputs |
| Stromverbindung | 2x 8-pin | None |
| Board-Nummer | D051 | PG183 SKU 200 |
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 |
