NagyDonat wrote: > RT is not a concern to me in this case.
The reason why I decided to check the runtime is because you said that "It not only simplifies it but also makes if faster. " in an earlier commit. My first runtime measurement finished and unfortunately it seems that this PR somehow _increases_ the runtime :confused: on average: | Project | before this PR| this PR | Ratio | | --- | --- | --- | --- | | memcached | 23.78 | 26.44 | 111.2% | | tmux | 74.92 | 73.93 | 98.7% | | curl | 231.98 | 238.36 | 102.8% | | twin | 57.52 | 54.77 | 95.2% | | vim | 179.32 | 164.70 | 91.8% | | openssl | 407.68 | 407.90 | 100.1% | | sqlite | 878.11 | 835.58 | 95.2% | | ffmpeg | 1306.35 | 1288.38 | 98.6% | | postgres | 401.11 | 397.88 | 99.2% | | tinyxml2 | 28.74 | 28.44 | 98.9% | | libwebm | 57.17 | 60.29 | 105.5% | | xerces | 204.98 | 204.06 | 99.6% | | bitcoin | 160.97 | 158.15 | 98.2% | | protobuf | 2332.42 | 2435.38 | 104.4% | | qtbase | 3226.83 | 3851.08 | 119.3% | | openrct2 | 1154.17 | 1172.49 | 101.6% | | **total** | 10726.05 | 11397.83 | 106.3% | I restarted this measurement to see whether this is a real tendency or just noise... but if this is a real slowdown, then I'll probably abandon this direction. https://github.com/llvm/llvm-project/pull/139939 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
