projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 0d7c42c) | patch
Integer legalization: fix MUL expansion
authorPawel Bylica <chfast@gmail.com>
Tue, 15 Nov 2016 18:29:24 +0000 (18:29 +0000)
committerPawel Bylica <chfast@gmail.com>
Tue, 15 Nov 2016 18:29:24 +0000 (18:29 +0000)
commitc3f6c97f713933554f1555397887f22e64f555b7
tree49b6ddee17be05e7dda124aa1cb6037442dbe0d1tree | snapshot
parent0d7c42c7abb3a822ce5b3c5960923dee07cc496fcommit | diff
Integer legalization: fix MUL expansion

Summary:
This fixes the runtime results produces by the fallback multiplication expansion introduced in r270720.

For tests I created a fuzz tester that compares the results with Boost.Multiprecision.

Reviewers: hfinkel

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D26628

llvm-svn: 286998
llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp diff | blob | history
llvm/test/CodeGen/X86/mul-i1024.ll diff | blob | history
llvm/test/CodeGen/X86/mul-i256.ll diff | blob | history
llvm/test/CodeGen/X86/mul-i512.ll diff | blob | history
Domain: System / Toolchain;