Skip to content

V
V8
Switch branch/tag
  1. 06 Nov, 2017 1 commit
    • Clemens Hammacher's avatar
      Reland "[bits] Consolidate Count{Leading,Trailing}Zeros" · 27ffc624
      Clemens Hammacher authored 
      This is a reland of 7d231e57, fixed to
avoid instantiating CountLeadingZeros for bits==0.

Original change's description:
> [bits] Consolidate Count{Leading,Trailing}Zeros
> 
> Instead of having one method for 32 bit integers and one for 64 bit,
> plus a templatized version to choose from those two, just implement one
> version which handles unsigned integers of any size. Also, make them
> constexpr.
> The Count{Leading,Trailing}Zeros{32,64} methods are kept for now in
> order to keep the amount of code changes small. Also, sometimes it
> improves readability by stating exactly the size of the argument,
> especially for leading zeros (where zero-extending would add more
> leading zeros).
> 
> CountLeadingZeros now uses a binary search inspired implementation
> as proposed in Hacker's Delight. It's more than 20% faster on x64 if
> the builtins are disabled.
> CountTrailingZeros falls back to CountPopulation instead of counting in
> a naive loop. This is ~50% faster.
> 
> R=mstarzinger@chromium.org
> 
> Change-Id: I1d8bf1d7295b930724163248150444bd17fbb34e
> Reviewed-on: https://chromium-review.googlesource.com/741231
> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
> Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#49106}

Change-Id: Icdff2510ec66d1c96a1912cef29d77d8550994ee
Reviewed-on: https://chromium-review.googlesource.com/753903Reviewed-by: 's avatarMichael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#49138}
      27ffc624
  3. 04 Nov, 2017 1 commit
    • Michael Achenbach's avatar
      Revert "[bits] Consolidate Count{Leading,Trailing}Zeros" · 1a1968fe
      Michael Achenbach authored 
      This reverts commit 7d231e57.

Reason for revert: Breaks revert for win-clang:
https://build.chromium.org/p/tryserver.chromium.win/builders/win_clang/builds/342755

Original change's description:
> [bits] Consolidate Count{Leading,Trailing}Zeros
> 
> Instead of having one method for 32 bit integers and one for 64 bit,
> plus a templatized version to choose from those two, just implement one
> version which handles unsigned integers of any size. Also, make them
> constexpr.
> The Count{Leading,Trailing}Zeros{32,64} methods are kept for now in
> order to keep the amount of code changes small. Also, sometimes it
> improves readability by stating exactly the size of the argument,
> especially for leading zeros (where zero-extending would add more
> leading zeros).
> 
> CountLeadingZeros now uses a binary search inspired implementation
> as proposed in Hacker's Delight. It's more than 20% faster on x64 if
> the builtins are disabled.
> CountTrailingZeros falls back to CountPopulation instead of counting in
> a naive loop. This is ~50% faster.
> 
> R=​mstarzinger@chromium.org
> 
> Change-Id: I1d8bf1d7295b930724163248150444bd17fbb34e
> Reviewed-on: https://chromium-review.googlesource.com/741231
> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
> Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#49106}

TBR=mstarzinger@chromium.org,clemensh@chromium.org

Change-Id: Iceeb35bf9c7539a1013c9bdbc47118008611bef2
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/753463Reviewed-by: 's avatarMichael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#49123}
      1a1968fe
  5. 03 Nov, 2017 1 commit
    • Clemens Hammacher's avatar
      [bits] Consolidate Count{Leading,Trailing}Zeros · 7d231e57
      Clemens Hammacher authored 
      Instead of having one method for 32 bit integers and one for 64 bit,
plus a templatized version to choose from those two, just implement one
version which handles unsigned integers of any size. Also, make them
constexpr.
The Count{Leading,Trailing}Zeros{32,64} methods are kept for now in
order to keep the amount of code changes small. Also, sometimes it
improves readability by stating exactly the size of the argument,
especially for leading zeros (where zero-extending would add more
leading zeros).

CountLeadingZeros now uses a binary search inspired implementation
as proposed in Hacker's Delight. It's more than 20% faster on x64 if
the builtins are disabled.
CountTrailingZeros falls back to CountPopulation instead of counting in
a naive loop. This is ~50% faster.

R=mstarzinger@chromium.org

Change-Id: I1d8bf1d7295b930724163248150444bd17fbb34e
Reviewed-on: https://chromium-review.googlesource.com/741231Reviewed-by: 's avatarMichael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#49106}
      7d231e57
  7. 20 Oct, 2017 1 commit
  9. 19 Oct, 2017 1 commit
  11. 13 Sep, 2017 1 commit
  13. 07 Sep, 2017 1 commit
  15. 14 Jul, 2017 1 commit
  17. 13 Jul, 2017 1 commit
  19. 23 Feb, 2017 1 commit
  21. 16 Feb, 2017 1 commit
    • Andreas Haas's avatar
      [turbofan] For Word32Shl optimizations only consider the last 5 bits of the shift · 5f1661aa
      Andreas Haas authored 
      One optimization in the machine-operator-reducer did not consider that
that word32 shift left instructions only consider the last 5 bits of
the shift input.

The issue only occurs for WebAssembly because in JavaScript we always
add a "& 0xf" on the shift value to the TurboFan graph.

For additional background: The JavaScript and WebAssembly spec both
say that only the last 5 bits of the shift value are used in the
word32-shift-left operation. This means that an "x << 0x29", in the
code is actually executed as "x << 0x09". Therefore the changes in
this CL are okay because they mask the last 5 bit of the shift value.

BUG=chromium:689450

Change-Id: Id92f298ed6d7f1714b109b3f4fbcecd5ac6d30f7
Reviewed-on: https://chromium-review.googlesource.com/439312Reviewed-by: 's avatarBen Titzer <titzer@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43245}
      5f1661aa
  23. 03 Feb, 2017 1 commit
  25. 02 Feb, 2017 1 commit
  27. 01 Feb, 2017 1 commit
  29. 26 Jan, 2017 1 commit
    • ahaas's avatar
      [wasm] Fix constant folding with signalling NaN. · 55aed782
      ahaas authored 
      According to the WebAssembly spec no arithmetic operation should ever
return a signalling NaN. With the constant folding in V8, however, it
was possible that some arithmetic operations were elided, and if the
input of the arithmetic operation was a signalling NaN, then also the
result was the same signalling NaN. This CL removes some constant
folding optimizations and adjusts others so that even with constant
folding the result of an arithmetic operation is never a signalling NaN.

R=titzer@chromium.org, rossberg@chromium.org, bmeurer@chromium.org

Review-Url: https://codereview.chromium.org/2647353007
Cr-Commit-Position: refs/heads/master@{#42694}
      55aed782
  31. 04 Jan, 2017 1 commit
  33. 30 Nov, 2016 1 commit
  35. 27 Sep, 2016 1 commit
  37. 16 Sep, 2016 1 commit
  39. 13 Sep, 2016 1 commit
  41. 10 Aug, 2016 1 commit
  43. 08 Aug, 2016 1 commit
  45. 04 Aug, 2016 1 commit
  47. 03 Aug, 2016 1 commit
  49. 25 Jul, 2016 1 commit
  51. 22 Jul, 2016 1 commit
  53. 14 Jul, 2016 1 commit
  55. 01 Jul, 2016 1 commit
    • bmeurer's avatar
      [builtins] Unify most of the remaining Math builtins. · 0a0fe8fb
      bmeurer authored 
      Import fdlibm versions of acos, acosh, asin and asinh, which are more
precise and produce the same result across platforms (we were using
libm versions for asin and acos so far, where both speed and precision
depended on the operating system so far). Introduce appropriate TurboFan
operators for these functions and use them both for inlining and for the
generic builtin.

Also migrate the Math.imul and Math.fround builtins to TurboFan builtins
to ensure that their behavior is always exactly the same as the inlined
TurboFan version (i.e. C++ truncation semantics for double to float
don't necessarily meet the JavaScript semantics).

For completeness, also migrate Math.sign, which can even get some nice
love in TurboFan.

Drive-by-fix: Some alpha-sorting on the Math related functions, and
cleanup the list of Math intrinsics that we have to export via the
native context currently.

BUG=v8:3266,v8:3496,v8:3509,v8:3952,v8:5169,v8:5170,v8:5171,v8:5172
TBR=rossberg@chromium.org
R=franzih@chromium.org

Review-Url: https://codereview.chromium.org/2116753002
Cr-Commit-Position: refs/heads/master@{#37476}
      0a0fe8fb
  57. 30 Jun, 2016 1 commit
  59. 28 Jun, 2016 1 commit
    • bmeurer's avatar
      [turbofan] Introduce Float64Pow and NumberPow operators. · e607e12e
      bmeurer authored 
      Introduce a new machine operator Float64Pow that for now is backed by
the existing MathPowStub to start the unification of Math.pow, and at
the same time address the main performance issue that TurboFan still has
with the imaging-darkroom benchmark in Kraken.

Also migrate the Math.pow builtin itself to a TurboFan builtin and
remove a few hundred lines of hand-written platform code for special
handling of the fullcodegen Math.pow version.

BUG=v8:3599,v8:5086,v8:5157

Review-Url: https://codereview.chromium.org/2103733003
Cr-Commit-Position: refs/heads/master@{#37323}
      e607e12e
  61. 22 Jun, 2016 1 commit
  63. 20 Jun, 2016 1 commit
    • bmeurer's avatar
      [builtins] Introduce proper Float64Tan operator. · c87168bc
      bmeurer authored 
      Import base::ieee754::tan() from fdlibm and introduce Float64Tan TurboFan
operator based on that, similar to what we do for Float64Cos and Float64Sin.
Rewrite Math.tan() as TurboFan builtin and use those operators to also
inline Math.tan() into optimized TurboFan functions.

Drive-by-fix: Kill the %_ConstructDouble intrinsics, and provide only
the %ConstructDouble runtime entry for writing tests.

BUG=v8:5086,v8:5126
R=yangguo@chromium.org

Review-Url: https://codereview.chromium.org/2083453002
Cr-Commit-Position: refs/heads/master@{#37087}
      c87168bc
  65. 17 Jun, 2016 4 commits
  67. 16 Jun, 2016 3 commits
  69. 13 Jun, 2016 1 commit
    • bmeurer's avatar
      [builtins] Introduce proper Float64Atan and Float64Atan2 operators. · 89d8c57b
      bmeurer authored 
      Import base::ieee754::atan() and base::ieee754::atan2() from fdlibm and
introduce Float64Atan and Float64Atan2 TurboFan operators based on those,
similar to what we already did for Float64Log and Float64Log1p. Rewrite
Math.atan() and Math.atan2() as TurboFan builtin and use the operators
to also inline Math.atan() and Math.atan2() into optimized TurboFan functions.

R=yangguo@chromium.org
BUG=v8:5086,v8:5095

Review-Url: https://codereview.chromium.org/2065503002
Cr-Commit-Position: refs/heads/master@{#36916}
      89d8c57b