[LLVMbugs] [Bug 10496] [AVX] missing intrinsics for integer min/max instructions

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=10496

Bruno Cardoso Lopes <bruno.cardoso at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |INVALID

--- Comment #6 from Bruno Cardoso Lopes <bruno.cardoso at gmail.com> 2011-08-03 18:38:41 CDT ---
Hi,

I think you're still making confusion! :)

There's no pminsd instruction for 256-bit vectors, only re-encoded ones for
128-bit. So, if you want to have min/max for 256-bit, you should issue two
@llvm.x86.sse41.pminsd and join the result vectors. This happens because llvm
doesn't have native "max" or "min", so if you're using intrinsics, you should
use the right ones, because the target independent legalizer won't touch it. 

AVX2 probably has a max/min 256-bit instruction, but this is future-future
work!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.