projects / platform / upstream / dotnet / runtime.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: fa9409c) | patch
Implement long to float cast for x86
authorMike Danes <onemihaid@hotmail.com>
Sun, 11 Sep 2016 12:49:12 +0000 (15:49 +0300)
committerMike Danes <onemihaid@hotmail.com>
Mon, 12 Sep 2016 04:35:43 +0000 (07:35 +0300)
commite01ed457ab3500d18749d1f44e812e859215cfcc
treea6da5e2953e248c49a9e5a7c4b8812082f66848atree | snapshot
parentfa9409c010b084a6a6a0e0ed04599e03ab966a8bcommit | diff
Implement long to float cast for x86

Convert long to float/double casts to helper calls. Despite the call this version is 2x faster than JIT32's FILD implementation which suffers a significant store forwarding stall penalty.

Commit migrated from https://github.com/dotnet/coreclr/commit/398317d18b0845681671ba449b8d8d9ca2cd505c
src/coreclr/src/jit/codegenxarch.cpp diff | blob | history
src/coreclr/src/jit/morph.cpp diff | blob | history
src/coreclr/src/vm/jithelpers.cpp diff | blob | history
Domain: Dotnet / Core;