1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 03:33:20 +01:00
llvm-mirror/unittests
Lang Hames 2082c9d610 [ADT] Fix PR20728 - Incorrect APFloat::fusedMultiplyAdd results for x86_fp80.
As detailed at http://llvm.org/PR20728, due to an internal overflow in
APFloat::multiplySignificand the APFloat::fusedMultiplyAdd method can return
incorrect results for x87DoubleExtended (x86_fp80) values. This commonly
manifests as incorrect constant folding of libm fmal calls on x86. E.g.

fmal(1.0L, 1.0L, 3.0L) == 0.0L      (should be 4.0L)

This patch fixes PR20728 by adding an extra bit to the significand for
intermediate results of APFloat::multiplySignificand, avoiding the overflow.

llvm-svn: 222374
2014-11-19 19:15:41 +00:00
..
ADT [ADT] Fix PR20728 - Incorrect APFloat::fusedMultiplyAdd results for x86_fp80. 2014-11-19 19:15:41 +00:00
Analysis CallGraphTest.cpp: Remove invalid tests. ++S might step over F if S == F. 2014-11-18 12:23:19 +00:00
Bitcode [Bitcode] AtEndOfStream should only check against the size if it's known. 2014-11-13 00:08:41 +00:00
CodeGen
DebugInfo
ExecutionEngine [MCJIT] Make repeat calls to MCJIT::getPointerToFunction for declarations safe. 2014-10-22 23:18:42 +00:00
IR Add unittest for extreme alignments. 2014-10-23 14:45:19 +00:00
LineEditor [CMake] LineEditorTests: Add Support to link_components. 2014-07-24 11:44:33 +00:00
Linker Fix the autoconf build. 2014-11-17 21:06:38 +00:00
MC Dispose disassembler after use in unit test. 2014-11-15 10:53:12 +00:00
Option
Support Fixing some sign comparison warnings from MSVC; NFC. 2014-11-13 13:39:49 +00:00
Transforms Move the complex address expression out of DIVariable and into an extra 2014-10-01 18:55:02 +00:00
CMakeLists.txt
Makefile
Makefile.unittest