NVIDIA CMP 90HX vs NVIDIA Tesla P40
Grafikprozessor
| GPU Name | GA102 | GP102 |
|---|---|---|
| GPU Variante | GA102-100-A1 | — |
| Architektur | Ampere | Pascal |
| Hersteller | Samsung | TSMC |
| Prozessgröße | 8 nm | 16 nm |
| Transistoren | 28,300 million | 11,800 million |
| Chipgröße | 628 mm² | 471 mm² |
Grafikkarte
| Veröffentlichungsdatum | Jul 28th, 2021 | Sep 13th, 2016 |
|---|---|---|
| Generation | Mining GPUs | Tesla |
| Produktion | Active | End-of-life |
| Businterface | PCIe 4.0 x16 | PCIe 3.0 x16 |
| Originalpreis | — | 5,699 USD |
Takt-Geschwindigkeiten
| Basis-Taktfrequenz | 1500 MHz | 1303 MHz |
|---|---|---|
| Boost-Takt | 1710 MHz | 1531 MHz |
| Speicher-Taktung | 1188 MHz 19 Gbps effective | 1808 MHz 14.5 Gbps effective |
Speicher
| Speichergröße | 10 GB | 24 GB |
|---|---|---|
| Speichertyp | GDDR6X | GDDR5X |
| Speicherbus | 320 bit | 384 bit |
| Bandbreite | 760.3 GB/s | 694.3 GB/s |
Render-Konfiguration
| Shading-Einheiten | 6400 | 3840 |
|---|---|---|
| TMUs | 200 | 240 |
| ROPs | 80 | 96 |
| Anzahl SM | 50 | 30 |
| Tensor Cores | 200 | — |
| RT-Kerne | 50 | — |
| Cache L1 | 128 KB (per SM) | 48 KB (per SM) |
| Cache L2 | 5 MB | 3 MB |
Theoretische Leistung
| Pixelrate | 136.8 GPixel/s | 147.0 GPixel/s |
|---|---|---|
| Texturrate | 342.0 GTexel/s | 367.4 GTexel/s |
| FP16 (half) Leistung | 21.89 TFLOPS (1:1) | 183.7 GFLOPS (1:64) |
| FP32 (float) Leistung | 21.89 TFLOPS | 11.76 TFLOPS |
| FP64 (double) Leistung | 342.0 GFLOPS (1:64) | 367.4 GFLOPS (1:32) |
Boarddesign
| Steckplatzbreite | Dual-slot | Dual-slot |
|---|---|---|
| Länge | 285 mm 11.2 inches | 267 mm 10.5 inches |
| Breite | 112 mm 4.4 inches | 111 mm 4.4 inches |
| TDP | 320 W | 250 W |
| Vorgeschlagenes Netzteil | 700 W | 600 W |
| Ausgabe | No outputs | No outputs |
| Stromverbindung | 2x 8-pin | 1x 6-pin + 1x 8-pin |
| Board-Nummer | PG132 SKU 100 | PG610 SKU 200 |
Grafik-Features
| DirectX | 12 Ultimate (12_2) | 12 (12_1) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 3.0 | 3.0 |
| Vulkan | 1.2 | 1.2 |
| CUDA | 8.6 | 6.1 |
| Shader-Modell | 6.6 | 6.4 |
