AVX-512 vector shuffle lowering

Demikhovsky, Elena elena.demikhovsky at intel.com
Mon Jun 29 00:42:09 PDT 2015


Ping*

-  Elena


-----Original Message-----
From: Elena Demikhovsky [mailto:elena.demikhovsky at intel.com] 
Sent: Wednesday, June 24, 2015 11:58
To: Demikhovsky, Elena; chandlerc at gmail.com; qcolombet at apple.com; Andrea_DiBiagio at sn.scee.net
Cc: llvm-commits at cs.uiuc.edu
Subject: [PATCH] AVX-512 vector shuffle lowering

Hi chandlerc, qcolombet, andreadb,

Implemented lowering for 512-bit vector shuffles. 
Vector types: <8 x 64>, <16 x 32>, <32 x 16> float and integer. 

AVX-512 provides vector shuffle instructions with variable mask (mask in register) for one and two sources (VPERM and VPERMT2).
Use them instead of splitting vectors.

All new shuffle instructions are for integer and FP data types.

More optimizations in the next patch.

REPOSITORY
  rL LLVM

http://reviews.llvm.org/D10683

Files:
  lib/Target/X86/X86ISelLowering.cpp
  test/CodeGen/X86/vector-shuffle-512-v16.ll
  test/CodeGen/X86/vector-shuffle-512-v32.ll
  test/CodeGen/X86/vector-shuffle-512-v8.ll

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/
---------------------------------------------------------------------
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