search for: klochkov

...c(0L, 0L, 0L, 0L, 0L, 0L ), Tommaso.Teofili = c(0L, 0L, 0L, 0L, 0L, 0L), Jeff.Hammerbacher. = c(0L, 0L, 0L, 0L, 0L, 0L), Andrew.Purtell. = c(0L, 0L, 0L, 0L, 0L, 0L), Johan.Oskarsson = c(0L, 0L, 0L, 0L, 0L, 0L), Erik.Frey = c(0L, 0L, 0L, 0L, 0L, 0L), Sebb = c(0L, 0L, 0L, 0L, 0L, 0L), Andrew.klochkov = c(0L, 0L, 0L, 0L, 0L, 0L), Paul.Egan = c(0L, 0L, 0L, 0L, 0L, 0L), Hudson = c(0L, 0L, 0L, 0L, 0L, 0L), Hadoop = c(0L, 0L, 0L, 0L, 0L, 0L), Aaron = c(0L, 0L, 0L, 0L, 0L, 0L), henry = c(0L, 0L, 0L, 0L, 0L, 0L), steve = c(0L, 0L, 0L, 0L, 0L, 0L), russel = c(0L, 0L, 0L, 0L, 0L, 0L)), .Name...

I noticed that the operand commuting code in X86InstrInfo.cpp treats scalar FMA intrinsics specially. It prevents operand commuting on these scalar instructions because the scalar FMA instructions preserve the upper bits of the vector. Presumably, the restrictions are there because commuting operands potentially changes the result upper bits. However, AFAIK the Intel and GNU FMA intrinsics