CUDA Occupancy Calculator
Analyze and optimize your GPU kernel's performance
โ๏ธ Kernel Configuration
๐๏ธ GPU Architecture Specifications
Architecture | Max Threads/SM | Max Blocks/SM | Max Warps/SM | Shared Memory/SM | Registers/SM | Warp Size |
---|---|---|---|---|---|---|
Compute 1.x (Tesla) | 768 | 8 | 24 | 16 KB | 8,192 | 32 |
Compute 2.x (Fermi) | 1,536 | 8 | 48 | 48 KB | 32,768 | 32 |
Compute 3.x (Kepler) | 2,048 | 16 | 64 | 48 KB | 65,536 | 32 |
Compute 5.x (Maxwell) | 2,048 | 32 | 64 | 96 KB | 65,536 | 32 |
Compute 6.x (Pascal) | 2,048 | 32 | 64 | 96 KB | 65,536 | 32 |
Compute 7.x (Volta/Turing) | 2,048 | 32 | 64 | 96 KB | 65,536 | 32 |
Compute 8.x (Ampere) | 2,048 | 32 | 64 | 164 KB | 65,536 | 32 |
Compute 9.0 (Hopper) | 2,048 | 32 | 64 | 228 KB | 65,536 | 32 |
Compute 10.0 (Blackwell) | 2,048 | 32 | 64 | 256 KB | 65,536 | 32 |
Read Also: