[PATCH][AVX512] Add mem patterns for 512b FP rounds

Demikhovsky, Elena elena.demikhovsky at intel.com
Thu Feb 19 02:53:13 PST 2015


Hi Cameron,

I implemented (229837) all forms VRNDSCALE scalar with masks, sae, memory and so on.
If you want to complete the vector form, it should be done in the similar way.

Thanks.

-  Elena


-----Original Message-----
From: Demikhovsky, Elena 
Sent: Wednesday, February 18, 2015 23:29
To: 'cameron.mcinally at nyu.edu'
Cc: llvm-commits at cs.uiuc.edu
Subject: RE: [PATCH][AVX512] Add mem patterns for 512b FP rounds

Hi Cameron,

Please wait, these instructions should be refactored with masks, broadcast, sae and other forms.
I plan to complete the full form of the scalar version tomorrow.

-  Elena


-----Original Message-----
From: Cameron McInally [mailto:cameron.mcinally at nyu.edu] 
Sent: Tuesday, February 17, 2015 18:23
To: llvm-commits at cs.uiuc.edu
Cc: Demikhovsky, Elena; Robert Khasanov
Subject: [PATCH][AVX512] Add mem patterns for 512b FP rounds

Hey guys,

Attached is a patch to add patterns for 512b FP rounds that fold a load.

I'm somewhat unhappy with the verbosity of this patch, but couldn't find a more compact representation. Any suggestions?

Thanks,
Cameron
---------------------------------------------------------------------
Intel Israel (74) Limited

This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.




More information about the llvm-commits mailing list