1. 09 Jan, 2016 1 commit
  2. 06 Jan, 2016 1 commit
  3. 02 Dec, 2015 1 commit
  4. 01 Oct, 2015 1 commit
  5. 27 May, 2015 1 commit
  6. 28 Mar, 2015 1 commit
  7. 11 Nov, 2014 2 commits
  8. 19 Aug, 2014 2 commits
  9. 15 Aug, 2014 1 commit
  10. 14 Jul, 2014 10 commits
  11. 21 Feb, 2014 1 commit
  12. 14 Feb, 2014 1 commit
  13. 20 Nov, 2013 1 commit
  14. 15 Nov, 2013 1 commit
  15. 01 Nov, 2013 1 commit
    • Eric Anholt's avatar
      Skip counting the second pair of an instruction in 16-wide mode. · 703de8e8
      Eric Anholt authored
      Generally we want our count to approximate time, and the two SIMD8
      instructions are same time as a SIMD16 instruction, except for the
      cost to instruction cache pressure.  This pattern happens for MAD and
      LRP a lot, and there are other cases, too.  Double-counting these
      meant that we basically couldn't use shader-db to validate
      optimizations involving those instructions.
      703de8e8
  16. 08 Apr, 2013 4 commits
  17. 25 Mar, 2013 4 commits
  18. 23 Mar, 2013 2 commits
  19. 22 Mar, 2013 4 commits