Product Compare
NVIDIA RTX PRO 4000 Blackwell vs AMD Radeon Instinct MI300A
| GPU | GPU | |
|---|---|---|
| Graphics Card | ||
| Release Date | Mar 2025 | Dec 2023 |
| Generation | Blackwell PRO W | Radeon Instinct |
| Type | Desktop | Desktop |
| Bus Interface | PCIe 5.0 x16 | PCIe 5.0 x16 |
| Clock Speeds | ||
| Base Clock | 1230 MHz | 1000 MHz |
| Boost Clock | 2055 MHz | 2100 MHz |
| Memory Clock | 1750 MHz | 2525 MHz |
| Memory | ||
| Memory Size | 24GB | 192GB |
| Memory Type | GDDR7 | HBM3 |
| Memory Bus | 192bit | 8192bit |
| Bandwidth | 672.0GB/s | 10.3TB/s |
| Render Config | ||
| SM Count | 70 | |
| Shading Units | 8960 | 19456 |
| TMUs | 280 | 1216 |
| ROPs | 96 | 0 |
| Tensor Cores | 280 | 1216 |
| RT Cores | 70 | |
| L1 Cache | 128 KB (per SM) | 16 KB (per CU) |
| L2 Cache | 48 MB | 16 MB |
| Compute Units | 304 | |
| L3 Cache | 256 MB | |
| Theoretical Performance | ||
| Pixel Rate | 197.3 GPixel/s | 0 MPixel/s |
| Texture Rate | 575.4 GTexel/s | 2554 GTexel/s |
| FP16 (half) | 36.83 TFLOPS | 653.7 TFLOPS |
| FP32 (float) | 36.83 TFLOPS | 81.72 TFLOPS |
| FP64 (double) | 575.4 GFLOPS | 81.72 TFLOPS |
| Graphics Processor | ||
| GPU Name | GB203 | Aqua Vanjaram |
| Architecture | Blackwell 2.0 | CDNA 3.0 |
| Foundry | TSMC | TSMC |
| Process Size | 5 nm | 5 nm |
| Transistors | 45.6 billion | 153 billion |
| Die Size | 378 mm² | 1017 mm² |
| Board Design | ||
| TDP | 140W | 750W |
| Suggested PSU | 300 W | 1150 W |
| Outputs | 4x DisplayPort 2.1b | No outputs |
| Power Connectors | 1x 16-pin | None |
| Graphics Features | ||
| DirectX | 12 Ultimate (12_2) | N/A |
| OpenGL | 4.6 | N/A |
| OpenCL | 3.0 | 3.0 |
| Vulkan | 1.4 | N/A |
| CUDA | 12.0 | |
| Shader Model | 6.8 | N/A |