[llvm] [TableGen] Optimize intrinsic info type signature encoding (PR #106809)
Rahul Joshi via llvm-commits
llvm-commits at lists.llvm.org
Mon Sep 2 06:46:39 PDT 2024
jurahul wrote:
Actual benchmark runs:
```
Old:
~/upstream_clean$ ./build/benchmarks/GetIntrinsicInfoTableEntriesBM --benchmark_repetitions=40 --benchmark_display_aggregates_only
2024-09-01T08:16:40-07:00
Running ./build/benchmarks/GetIntrinsicInfoTableEntriesBM
Run on (20 X 2918.4 MHz CPU s)
CPU Caches:
L1 Data 48 KiB (x10)
L1 Instruction 32 KiB (x10)
L2 Unified 1280 KiB (x10)
L3 Unified 24576 KiB (x1)
Load Average: 0.43, 0.29, 0.40
---------------------------------------------------------------------------------
Benchmark Time CPU Iterations
---------------------------------------------------------------------------------
BM_GetIntrinsicInfoTableEntries_mean 511908 ns 511877 ns 40
BM_GetIntrinsicInfoTableEntries_median 510461 ns 510461 ns 40
BM_GetIntrinsicInfoTableEntries_stddev 9411 ns 9428 ns 40
BM_GetIntrinsicInfoTableEntries_cv 1.84 % 1.84 % 40
new:
~/upstream_llvm/llvm-project$ ./build/benchmarks/GetIntrinsicInfoTableEntriesBM --benchmark_repetitions=40 --benchmark_display_aggregates_only
2024-09-01T08:15:59-07:00
Running ./build/benchmarks/GetIntrinsicInfoTableEntriesBM
Run on (20 X 2918.4 MHz CPU s)
CPU Caches:
L1 Data 48 KiB (x10)
L1 Instruction 32 KiB (x10)
L2 Unified 1280 KiB (x10)
L3 Unified 24576 KiB (x1)
Load Average: 0.19, 0.23, 0.39
---------------------------------------------------------------------------------
Benchmark Time CPU Iterations
---------------------------------------------------------------------------------
BM_GetIntrinsicInfoTableEntries_mean 474298 ns 474293 ns 40
BM_GetIntrinsicInfoTableEntries_median 474271 ns 474270 ns 40
BM_GetIntrinsicInfoTableEntries_stddev 5285 ns 5287 ns 40
BM_GetIntrinsicInfoTableEntries_cv 1.11 % 1.11 % 40
```
Speedup = 511877 -> 474293 = 7.3%
https://github.com/llvm/llvm-project/pull/106809
More information about the llvm-commits
mailing list