3 ms·
While it is not mentioned in the article, I am aware of /arch:IA32 (don't use SSE instructions) and I have tried it - no luck. On the other hand, using /fp:fas
by CookiePLMonster 8y ago
While it is not mentioned in the article, I am aware of /arch:IA32 (don't use SSE instructions) and I have tried it - no luck.
On the other hand, using /fp:fast with VS2017 did give me similar results to what VS2003/VS2010 produces - so if I were to guess, VS2003 generates x87 math in a manner similar to fast math, while precise math breaks it?
- moogly 8y agoIt appears VS2003 didn't have a defined (default) floating point model. It was introduced in VS2005 (default: precise): https://social.msdn.microsoft.com/Forums/sqlserver/en-US/44ed81d5-54a7-4c01-9055-2d235fc0a2cc/floatingpoint-models-in-vs2005-vs-vs2003?forum=vclanguage https://social.msdn.microsoft.com/Forums/sqlserver/en-US/44e...
- CookiePLMonster 8y agoThen again, it may not be strictly due to /fp:precise - I tried compiling with VS2010 too, and even though it was /fp:precise the bug did not manifest itself. Perhaps comparing assembly code would have given a definite answer, but since it is so obvious from source I don't think it is worth the effort.