[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Zachary Ware added the comment: Committed! Thank you Steve for the suggestion and Stefan for the approval. -- ___ Python tracker ___

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Zachary Ware added the comment: There's not a 32-bit ICC buildbot, though I could force one. But since you say commit it, I'll commit it ;) -- ___ Python tracker

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Roundup Robot added the comment: New changeset 863407e80370 by Zachary Ware in branch '3.5': Issue #24974: Force fp-model precice in mpdecimal.c on Windows https://hg.python.org/cpython/rev/863407e80370 New changeset 88c28d29afe4 by Zachary Ware in branch 'default': Closes #24974: Merge with

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Stefan Krah added the comment: Hmm, I don't see a 32-bit ICC buildbot. This problem can only occur on 32-bit with -DPPRO defined. Please go ahead and commit the patch. The inline asm miscompilation problem I mentioned earlier was solved in 12.0 (I think), so let's just assume it's gone

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Stefan Krah added the comment: We can always blame any fallout on ICC. ;) -- ___ Python tracker ___ ___

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Zachary Ware added the comment: I tested both 32 and 64 bit, with MSVC 14 (VS 2015) and ICC 15.0 (backed by VS 2015). I can go ahead and commit this to a sandbox for buildbot testing if it would help. -- ___ Python tracker

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Stefan Krah added the comment: Are you sure that you always tested the 32-bit build, also with ICC 15.0? -- ___ Python tracker ___

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Zachary Ware added the comment: As far as I can tell, this patch fixes the issue and doesn't break anything. Independent verification of that assertion would be lovely :) For the record, I was able to reproduce the issue on one of the Windows Server 2012 R2 machines after installing ICC

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Stefan Krah added the comment: Also note that ICC <= 11.0 did not handle __GNUC__ inline asm correctly, see the comment: Modules/_decimal/libmpdec/umodarith.h:391 I've disabled asm for Linux/ICC in setup.py. I don't know how the situation is with MASM inline asm, but I'd recommend to test

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Stefan Krah added the comment: Steve: What do you mean by "global option"? The C99 standard says that FENV_ACCESS (if set), is active until the end of the translation unit (here: mpdecimal.c). -- ___ Python tracker

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Stefan Krah added the comment: On Linux ICC has something like "fast-math" by default. The situation is a bit annoying: On Unix ICC defines __GNUC__, but does not really have all the features. Here ICC defines _MSC_VER, but does not behave like cl.exe. [I know it's a hard problem to be fully

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Changes by Facundo Batista : -- nosy: -facundobatista ___ Python tracker ___ ___

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Zachary Ware added the comment: Steve Dower added the comment: > fenv_access is not available when compiling with /fp:fast, which is > apparently ICC's default. > > The proposed workaround here changes that default to /fp:strict, which is a > very different model, for all of CPython. I

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Steve Dower added the comment: fenv_access is not available when compiling with /fp:fast, which is apparently ICC's default. The proposed workaround here changes that default to /fp:strict, which is a very different model, for all of CPython. I proposed using #pragma float_control to force

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Stefan Krah added the comment: If "#pragma float_control(precise, push)" is exactly the MSVC default then I'm fine with putting it on top of FENV_ACCESS. There's nothing speed sensitive going on here: In the 32-bit build the x87 FPU is used for modular multiplication of integers and needs the

[issue24974] ICC on Windows 8.1: _decimal fails to compile with default fp model

Changes by Zachary Ware : -- nosy: +christopher.hogan title: ICC on Windows 8.1: _decimal fails to compile without /fp:strict -> ICC on Windows 8.1: _decimal fails to compile with default fp model ___ Python tracker