NVIDIA Quadro P600 vs NVIDIA Tesla P4
Graphics Processor
| GPU model | GP107 | GP104 |
|---|---|---|
| Architecture | Pascal | Pascal |
| Foundry | Samsung | TSMC |
| Manufacturing process | 14 nm | 16 nm |
| Number of transistors | 3,300 million | 7,200 million |
| Die size | 132 mm² | 314 mm² |
| GPU variant | — | GP104-895-A1 |
Graphics Card
| Release date | Feb 7th, 2017 | Sep 13th, 2016 |
|---|---|---|
| Generation | Quadro | Tesla |
| Production | End-of-life | End-of-life |
| Interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Clocks
| Base clock | 1329 MHz | 886 MHz |
|---|---|---|
| Boost clock | 1557 MHz | 1114 MHz |
| Memory clock | 1002 MHz 4 Gbps effective | 1502 MHz 6 Gbps effective |
Memory Configuration
| Memory size | 2 GB | 8 GB |
|---|---|---|
| Memory type | GDDR5 | GDDR5 |
| Memory bus width | 128 bit | 256 bit |
| Bandwidth | 64.13 GB/s | 192.3 GB/s |
Render Configuration
| Shading units | 384 | 2560 |
|---|---|---|
| TMUs | 24 | 160 |
| ROPs | 16 | 64 |
| SM count | 3 | 20 |
| Cache L1 | 48 KB (per SM) | 48 KB (per SM) |
| Cache L2 | 1024 KB | 2 MB |
Performance
| Pixel rate | 24.91 GPixel/s | 71.30 GPixel/s |
|---|---|---|
| Texture rate | 37.37 GTexel/s | 178.2 GTexel/s |
| FP16 (half) performance | 18.68 GFLOPS (1:64) | 89.12 GFLOPS (1:64) |
| FP32 (float) performance | 1,196 GFLOPS | 5.704 TFLOPS |
| FP64 (double) performance | 37.37 GFLOPS (1:32) | 178.2 GFLOPS (1:32) |
Dimensions & Outputs
| Slot width | Single-slot | Single-slot |
|---|---|---|
| Length | 150 mm 5.9 inches | 168 mm 6.6 inches |
| Width | 69 mm 2.7 inches | — |
| TDP | 40 W | 75 W |
| Suggested PSU | 200 W | 250 W |
| Outputs | 4x mini-DisplayPort | No outputs |
| Power connectors | None | None |
| Board number | PG212 SKU 501 | — |
API Support & Features
| DirectX | 12 (12_1) | 12 (12_1) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 3.0 | 3.0 |
| Vulkan | 1.2 | 1.2 |
| CUDA | 6.1 | 6.1 |
| Shader model | 6.4 | 6.4 |
