Intel Haswell Scheduler Updated In LLVM
Last month in LLVM there was new Sandy Bridge scheduler information to improve the instruction scheduling and other hardware detail changes so LLVM can generate more efficient code for those older CPUs. At that time we learned Intel developers were also planning improvements too for LLVM with newer Haswell / Broadwell / Skylake / Skylake-X CPUs. Improvements have now landed for Haswell.
As of this morning in LLVM Git is updated scheduling information for Haswell CPUs. This updates the Haswell architecture target with new instruction scheduling data, including latency, number of micro-ops, etc. This is likely to result in some code performance fluctuations.
The updated Haswell scheduler can be found via LLVM in their 6.0 development code. I'll be working on some fresh LLVM/Clang Haswell/SandyBridge benchmarks shortly at Phoronix.
As of this morning in LLVM Git is updated scheduling information for Haswell CPUs. This updates the Haswell architecture target with new instruction scheduling data, including latency, number of micro-ops, etc. This is likely to result in some code performance fluctuations.
The updated Haswell scheduler can be found via LLVM in their 6.0 development code. I'll be working on some fresh LLVM/Clang Haswell/SandyBridge benchmarks shortly at Phoronix.
11 Comments