DirectCompute Optimizations and Best Practices - Nvidia

More documents

Recommendations

Info

Parallel Reduction: Interleaved Addressing Values (shared memory) Step 1 Stride 1 Step 2 Stride 2 Step 3 Stride 4 Step 4 Stride 8 Thread IDs Values Thread IDs Values Thread IDs Values Thread IDs Values 10 1 8 -1 0 -2 3 5 -2 -3 2 7 0 11 0 2 0 2 4 6 8 10 12 14 11 1 7 -1 -2 -2 8 5 -5 -3 9 7 11 11 2 2 0 4 8 12 18 1 7 -1 6 -2 8 5 4 -3 9 7 13 11 2 2 0 8 24 1 7 -1 6 -2 8 5 17 -3 9 7 13 11 2 2 0 41 1 7 -1 6 -2 8 5 17 -3 9 7 13 11 2 2
Performance for 4M element reduction Shader 1: interleaved addressing with divergent branching Time (2 22 ints) B<strong>and</strong>width 8.054 ms 2.083 GB/s Note: Block Size = 128 threads for all tests
Page 1 and 2: DirectCompute Optimizations and Bes
Page 3 and 4: Why GPUs • GPUs are throughput or
Page 5 and 6: Advantages of DirectCompute • Dir
Page 7 and 8: Parallel Execution Model Thread Gro
Page 9 and 10: Memory Coalescing • A coordinated
Page 11 and 12: Uncoalesced Access: Reading floats
Page 13 and 14: Compute 1.2+ Coalesced Access: Read
Page 15 and 16: Thread Group Shared Memory (TGSM)
Page 17 and 18: Shared Memory Bank Addressing • N
Page 19 and 20: What is Occupancy GPUs typical
Page 21 and 22: Maximizing HW Occupancy • Registe
Page 23 and 24: DirectCompute Optimization Example
Page 25 and 26: Parallel Reduction • Tree-based a
Page 27 and 28: Solution: Shader Decomposition •
Page 29: Reduction #1: Interleaved Addressin
Page 33 and 34: What is Thread Divergence • Diver
Page 35 and 36: Parallel Reduction: Interleaved Add
Page 37 and 38: Thread Group Shared Memory Bank Con
Page 39 and 40: Reduction #3: Sequential Addressing
Page 41 and 42: Idle Threads Problem: for (unsigned
Page 43 and 44: Performance for 4M element reductio
Page 45 and 46: Unrolling the Last Warp • As redu
Page 49 and 50: Reduction #6: Completely Unrolled #
Page 51 and 52: Reduction #7: Multiple Adds / Threa
Page 55 and 56: Time (ms) Performance Comparison 10
Page 58 and 59: Extra Slides
Page 60 and 61: Memory Coalescing (Matrix Multiply)
Page 62: Matrix Multiplication (cont.) Optim

DirectCompute Optimizations and Best Practices - Nvidia

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?