Commits

Branch selector

 master

User selector

chandlerc

Datepicker

All time

Commit History

Commits on Sep 25, 2014

• [x86] Teach the new vector shuffle lowering to use AVX2 instructions for

  Show description for 4b667ee
  chandlerc

  committed
  4b667ee

  Copy full SHA for 4b667ee

• [x86] Teach the new vector shuffle lowering a fancier way to lower

  Show description for 05901d8
  chandlerc

  committed
  05901d8

  Copy full SHA for 05901d8

• [x86] Fix an oversight in the v8i32 path of the new vector shuffle

  Show description for 2e8d2c7
  chandlerc

  committed
  2e8d2c7

  Copy full SHA for 2e8d2c7

• [x86] Rearrange the code for v16i16 lowering a bit for clarity and to

  Show description for e5fb4ad
  chandlerc

  committed
  e5fb4ad

  Copy full SHA for e5fb4ad

• [x86] Implement AVX2 support for v32i8 in the new vector shuffle

  Show description for e3bb4bb
  chandlerc

  committed
  e3bb4bb

  Copy full SHA for e3bb4bb

• [x86] More tweaks to the v32i8 test cases.

  Show description for 1d63231
  chandlerc

  committed
  1d63231

  Copy full SHA for 1d63231

• [x86] Re-work a bunch of the v32i8 test cases to actually involve byte

  Show description for a87d04a
  chandlerc

  committed
  a87d04a

  Copy full SHA for a87d04a

• [x86] Remove the defunct X86ISD::BLENDV entry -- we use vector selects

  Show description for 6a289bb
  chandlerc

  committed
  6a289bb

  Copy full SHA for 6a289bb

• [x86] Fix the v16i16 blend logic I added in the prior commit and add the

  Show description for ef673b3
  chandlerc

  committed
  ef673b3

  Copy full SHA for ef673b3

• [x86] Implement v16i16 support with AVX2 in the new vector shuffle

  Show description for bdecfeb
  chandlerc

  committed
  bdecfeb

  Copy full SHA for bdecfeb

Commits on Sep 24, 2014

• [x86] Factor out the logic to generically decombose a vector shuffle

  Show description for c88ae96
  chandlerc

  committed
  c88ae96

  Copy full SHA for c88ae96

• [x86] Teach the instruction lowering to add comments describing constant

  Show description for 10cd809
  chandlerc

  committed
  10cd809

  Copy full SHA for 10cd809

• [x86] More refactoring of the shuffle comment emission. The previous

  Show description for f00b50b
  chandlerc

  committed
  f00b50b

  Copy full SHA for f00b50b

• [x86] Bypass the shuffle mask comment generation when not using verbose

  Show description for db5e1da
  chandlerc

  committed
  db5e1da

  Copy full SHA for db5e1da

• [x86] Hoist the logic for extracting the relevant bits of information

  Show description for 7e0f903
  chandlerc

  committed
  7e0f903

  Copy full SHA for 7e0f903

• [x86] Start refactoring the comment printing logic in the MC lowering of

  Show description for 5f671ba
  chandlerc

  committed
  5f671ba

  Copy full SHA for 5f671ba

• [x86] Teach the new vector shuffle lowering to lower v8i32 shuffles with

  Show description for 6717f9d
  chandlerc

  committed
  6717f9d

  Copy full SHA for 6717f9d

• [x86] Fix a really terrible bug in the repeated 128-bin-lane shuffle

  Show description for 8415f84
  chandlerc

  committed
  8415f84

  Copy full SHA for 8415f84

Commits on Sep 23, 2014

• [x86] Teach the new vector shuffle lowering to lower v4i64 vector

  Show description for 30ce74b
  chandlerc

  committed
  30ce74b

  Copy full SHA for 30ce74b

• [x86] Teach the rest of the 'target shuffle' machinery about blends and

  Show description for 798f284
  chandlerc

  committed
  798f284

  Copy full SHA for 798f284

• [x86] Teach the new shuffle lowering's blend functionality to use AVX2's

  Show description for 7024c7e
  chandlerc

  committed
  7024c7e

  Copy full SHA for 7024c7e

• [x86] Teach the vector comment parsing and printing to correctly handle

  Show description for 4850be4
  chandlerc

  committed
  4850be4

  Copy full SHA for 4850be4

• [x86] Teach the AVX1 path of the new vector shuffle lowering one more

  Show description for 8f63778
  chandlerc

  committed
  8f63778

  Copy full SHA for 8f63778

Commits on Sep 22, 2014

• [x86] Rename X86ISD::VPERMILP to X86ISD::VPERMILPI (and the same for the

  Show description for 4b36515
  chandlerc

  committed
  4b36515

  Copy full SHA for 4b36515

• [x86] Stub out the integer lowering of 256-bit vectors with AVX2

  Show description for 8571ae3
  chandlerc

  committed
  8571ae3

  Copy full SHA for 8571ae3

• [x86] Introduce tests covering the gamut of 256-bit vector shuffling.

  Show description for 56c7cfe
  chandlerc

  committed
  56c7cfe

  Copy full SHA for 56c7cfe

• [x86] Move the AVX v4i64 test cases down to group them together.

  Show description for ec35919
  chandlerc

  committed
  ec35919

  Copy full SHA for ec35919

• [x86] Back out a bad choice about lowering v4i64 and pave the way for

  Show description for de95c38
  chandlerc

  committed
  de95c38

  Copy full SHA for de95c38

Commits on Sep 21, 2014

• [x86] Teach the new vector shuffle lowering how to cleverly lower single

  Show description for 37bb4b0
  chandlerc

  committed
  37bb4b0

  Copy full SHA for 37bb4b0

• [x86] Add a bunch of test cases where we have different shuffle patterns

  Show description for 7da57cf
  chandlerc

  committed
  7da57cf

  Copy full SHA for 7da57cf

• [x86] With the stronger canonicalization of shuffles added in r218216,

  Show description for 974e872
  chandlerc

  committed
  974e872

  Copy full SHA for 974e872

• [x86] Teach the new vector shuffle lowering to re-use the SHUFPS

  Show description for 974542d
  chandlerc

  committed
  974542d

  Copy full SHA for 974542d

• [x86] Refactor the logic to form SHUFPS instruction patterns to lower

  Show description for 38e1816
  chandlerc

  committed
  38e1816

  Copy full SHA for 38e1816

• [x86] Teach the new vector shuffle lowering the basics about insertion

  Show description for 1a5f7f5
  chandlerc

  committed
  1a5f7f5

  Copy full SHA for 1a5f7f5

• [x86] Teach the new vector shuffle lowering how to lower to UNPCKLPS and

  Show description for 6ef31b0
  chandlerc

  committed
  6ef31b0

  Copy full SHA for 6ef31b0

Pagination

Previous

Next