forked from OSchip/llvm-project
be8dc6499a
Vectors were being manually scalarized by the backend. Instead, let the target-independent code do all of the work. The manual scalarization was from a time before good target-independent support for scalarization in LLVM. However, this forces us to specially-handle vector loads and stores, which we can turn into PTX instructions that produce/consume multiple operands. llvm-svn: 174968 |
||
|---|---|---|
| .. | ||
| annotations.ll | ||
| arithmetic-fp-sm10.ll | ||
| arithmetic-fp-sm20.ll | ||
| arithmetic-int.ll | ||
| calling-conv.ll | ||
| compare-int.ll | ||
| convert-fp.ll | ||
| convert-int-sm10.ll | ||
| convert-int-sm20.ll | ||
| fma-disable.ll | ||
| fma.ll | ||
| global-ordering.ll | ||
| intrin-nocapture.ll | ||
| intrinsic-old.ll | ||
| intrinsics.ll | ||
| ld-addrspace.ll | ||
| ld-generic.ll | ||
| lit.local.cfg | ||
| param-align.ll | ||
| pr13291-i1-store.ll | ||
| ptx-version-30.ll | ||
| ptx-version-31.ll | ||
| simple-call.ll | ||
| sm-version-10.ll | ||
| sm-version-11.ll | ||
| sm-version-12.ll | ||
| sm-version-13.ll | ||
| sm-version-20.ll | ||
| sm-version-21.ll | ||
| sm-version-30.ll | ||
| sm-version-35.ll | ||
| st-addrspace.ll | ||
| st-generic.ll | ||
| tuple-literal.ll | ||
| vector-compare.ll | ||
| vector-loads.ll | ||
| vector-select.ll | ||