* Jakub Jelinek: > On Mon, Sep 19, 2022 at 11:25:13AM +0200, Florian Weimer wrote: >> * Jakub Jelinek: >> >> > The disadvantage of the patch is that touching reg[x].loc and how[x] >> > now means 2 cachelines rather than one as before, and I admit beyond >> > bootstrap/regtest I haven't benchmarked it in any way. Florian, could >> > you retry whatever you measured to get at the 40% of time spent on the >> > stack clearing to see how the numbers change? >> >> A benchmark that unwinds through 100 frames containing a std::string >> variable goes from (0b5b8ac5cb7fe92dd17ae8bd7de84640daa59e84): >> >> min: 24418 ns >> 25%: 24740 ns >> 50%: 24790 ns >> 75%: 24840 ns >> 95%: 24937 ns >> 99%: 26174 ns >> max: 42530 ns >> avg: 24826.1 ns >> >> to (0b5b8ac5cb7fe92dd17ae8bd7de84640daa59e84 with this patch): >> >> min: 22307 ns >> 25%: 22640 ns >> 50%: 22713 ns >> 75%: 22787 ns >> 95%: 22948 ns >> 99%: 24839 ns >> max: 52658 ns >> avg: 22863.4 ns >> >> So 227 ns per frame instead of 248 ns per frame, or ~9% less. > > Thanks for doing that.
So it turns out my test program had 100 frames, but not with std::string. With std::string objects, the numbers are: Before: min: 71236 ns 25%: 71637 ns 50%: 71724 ns 75%: 71857 ns 95%: 73148 ns 99%: 74023 ns max: 120735 ns avg: 71973.1 ns After: min: 69547 ns 25%: 69961 ns 50%: 70034 ns 75%: 70112 ns 95%: 71273 ns 99%: 71511 ns max: 82691 ns avg: 70121.3 ns So slightly less improvement per frame, but it's still there. >> Moving cfa_how after how in struct frame_state_reg_info as an 8-bit >> bitfield should avoid zeroing another 8 bytes. This shaves off another >> 3 ns per frame in my testing (on a Core i9-10900T, so with ERMS). > > Good idea. Won't help always, on some targets how could have size divisible > by pointer alignment, but when it is at the end it always increases the > size by alignment of pointer, while after how array it only does so if > how is multiple of pointer alignment. Okay, I'll send a separate patch once yours is in, along with some other simple changes. Thanks, Florian
