search for: origloop

...bination of optimization from *multiple* passes, neither pass would make the decision locally to version. Also, it’s probably beneficial to perform a single loop versioning even if multiple passes would like to check different accesses. E.g. rather than: Checks_1 / \ / \ OrigLoop Checks_2 \ / \ \ / \ \ NoAlias_1 NoAlias_2 \ | / \ | / \ | / Join But instead: Checks_1+Check_2 / \ / \ OrigLoop NoAlias_2 \ / \ / \ / Join This is e...

...n. > > > > > > Also, it’s probably beneficial to perform a single loop versioning > even if multiple passes would like to check different accesses. E.g. > rather than: > > > > > > Checks_1 > > > / \ > > > / \ > > > OrigLoop Checks_2 > > > \ / \ > > > \ / \ > > > \ NoAlias_1 NoAlias_2 > > > \ | / > > > \ | / > > > \ | / > > > Join > > > > > > But instead: > > > > > > Checks_1+Check_2 > > &g...

...n. > > > > > > Also, it’s probably beneficial to perform a single loop versioning > even if multiple passes would like to check different accesses. E.g. > rather than: > > > > > > Checks_1 > > > / \ > > > / \ > > > OrigLoop Checks_2 > > > \ / \ > > > \ / \ > > > \ NoAlias_1 NoAlias_2 > > > \ | / > > > \ | / > > > \ | / > > > Join > > > > > > But instead: > > > > > > Checks_1+Check_2 > > &g...

...to perform a single loop versioning >> even if multiple passes would like to check different accesses. E.g. >> rather than: >> >> >> >> >> >> Checks_1 >> >> >> / \ >> >> >> / \ >> >> >> OrigLoop Checks_2 >> >> >> \ / \ >> >> >> \ / \ >> >> >> \ NoAlias_1 NoAlias_2 >> >> >> \ | / >> >> >> \ | / >> >> >> \ | / >> >> >> Join >> >> >>...

...o down that path…… the following would be the basics steps. 1) First, look at the comments in createVectorizedLoopSkeleton(). “old scalar loop to handle remainder” is the one you want to vectorize again. Near the end of that function, LoopScalarBody is set. This is the loop body of interest. OrigLoop becomes the scalar Loop. First, try attaching a new metadata (with VF info) to it. 2) Create a RemainderVectorizePass, very similar to LoopVectorizePass, but let it work only on the loops that has the metadata. Add it to IPO/PassManagerBuilder right after LoopVectorize. You could also try add...

Hi Hameeza, Aside from Ashutosh's patch..... When the vector width is that large, we can't keep vectorizing remainder like below. It'll be a huge code size if nothing else ---- hitting ITLB miss because of this is very bad, for example. VF=2048 // main vector loop VF=1024 // vectorized remainder 1 VF=512 // vectorized remainder 2 ... Vectorize remainder until trip count is