On Wed, May 6, 2020 at 6:26 AM Richard Biener <rguent...@suse.de> wrote: > > On Tue, 5 May 2020, Richard Biener wrote: > > > > > This rewrites store-motion to process candidates where we can > > ensure order preserving separately and with no need to disambiguate > > against all stores. Those candidates we cannot handle this way > > are validated to be independent on all stores (w/o TBAA) and then > > processed as "unordered" (all conditionally executed stores are so > > as well). > > > > This will necessary cause > > FAIL: gcc.dg/graphite/pr80906.c scan-tree-dump graphite "isl AST to > > Gimple succeeded" > > because the SM previously performed is not valid for exactly the PR57359 > > reason, we still perform SM of qc for the innermost loop but that's not > > enough. > > > > There is still room for improvements because we still check some constraints > > for the order preserving cases that are only necessary in the current > > strict way for the unordered ones. Leaving that for the furture. > > > > Bootstrapped and tested on x86_64-unknown-linux-gnu, a final > > SPEC 2006 evaluation is running. > > A first complete 3-run looked too good so I repeated the off-noise > ones which eliminated most of them as noise ... there's two > consistent parts namely a ~1% regression on 401.bzip2 and a > ~1% progression on 464.h264ref. Given earlier much larger (~5-10%) > "consistent" noise on other tests I declare it a wash ... > > Means, I'll push the patch to master once 10.1 is released (unless > there are comments on the patch itself).
This caused: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94988 -- H.J.
