TABLE I.
GPU Organization | |
---|---|
TPCs (Thread Processing Cluster) | 10 total |
SMs (Streaming Multiprocessor) | 3 per TPC |
Shader Clock | 1.48 GHz |
Memory (DRAM) Clock | 1.24 GHz |
Memory (DRAM) Bus Width | 512-bit |
Memory (DRAM) Latency | 400 – 600 cycles |
| |
SM Resources (30 SMs total)
| |
SPs (Scalar Processor) | 8 per SM |
SFUs (Special Function Unit) | 2 per SM |
DPUs (Double Precision Unit) | 1 per SM |
Registers | 16,384 per SM |
Shared Memory | 16 KB per SM |
Constant Cache | 8 KB per SM |
Texture Cache | 6–8 KB per SM |
| |
Programming Model
| |
Warps | 32 threads |
Max number of threads per block | 512 threads |
Max sizes of each dimension of a block | 512 x 512 x 64 |
Max sizes of each dimension of a grid | 65,535 x 65,535 x 1 |
Global Memory | 1 GB total |
Constant Memory | 64 KB total |