[PATCH] [X86][SSE] Sign extension for target vector sizes less than 128 bits.

[Simon Pilgrim]

Hi qcolombet, delena, andreadb, spatel,

To add support for v2i8/v2i16 to v2f64, this patch adds support for sign extension for sub 128-bit vectors, such as to v2i32 (which can then be converted to v2f64). It concatenates with UNDEF subvectors up to 128-bits, performs the sign extension (i.e. as v4i32) and then extracts the target subvector.

This is really 2 patches (the sign extension of subvectors + conversion of v2i8/v2i16 to v2f64) so I will commit these separately after combined review.

REPOSITORY
  rL LLVM

http://reviews.llvm.org/D10589

Files:
  lib/Target/X86/X86ISelLowering.cpp
  test/CodeGen/X86/vec_int_to_fp.ll
  test/CodeGen/X86/vector-sext.ll
  test/CodeGen/X86/widen_conv-3.ll

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D10589.28080.patch
Type: text/x-patch
Size: 8826 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150620/14edb858/attachment.bin>