Commits

Breadcrumbs

History for

onf0817a3

User selector

All users

All time

Commit History

Commits on Feb 4, 2014

• do not apply de-morgan's laws on floats, nans break them

  kripken

  committed
  71c8f19

  Copy full SHA for 71c8f19

• handle fround in optimizing out of tempDoublePtr

  kripken

  committed
  b6782ad

  Copy full SHA for b6782ad

Commits on Dec 19, 2013

• Optimize away masks on stores to HEAP8 and HEAP16.

  Show description for 47e803a
  

  Dan Gohman

  committed
  47e803a

  Copy full SHA for 47e803a

Commits on Oct 3, 2013

• update js optimizer tests

  kripken

  committed
  8b76d21

  Copy full SHA for 8b76d21

Commits on Jul 18, 2013

• consolidate simplifyExpressions to one pass that happens before closure, where simplifyNotComps is guaranteed to be valid

  kripken

  committed
  646e5af

  Copy full SHA for 646e5af

Commits on Jul 14, 2013

• do not eliminate away tempDoublePtr uses that are alignment fixes; fixes #1375

  kripken

  committed
  d98e6e3

  Copy full SHA for d98e6e3

Commits on Jul 13, 2013

• avoid creating ~~~ (from ~~ ^ -1) which is confusing for asm given the role of ~~

  kripken

  committed
  aff455b

  Copy full SHA for aff455b

Commits on Jul 8, 2013

• Add '!==' as a comparison operator.

  

  Dan Gohman

  committed
  30d6e3c

  Copy full SHA for 30d6e3c

• Xor optimizations.

  Show description for b090967
  

  Dan Gohman

  committed
  b090967

  Copy full SHA for b090967

Commits on Jul 2, 2013

• Optimize (x<y)&1

  Show description for 0a2260a
  

  Dan Gohman

  committed
  0a2260a

  Copy full SHA for 0a2260a

Commits on Jun 27, 2013

• do not collapse & inside a FUNCTION_TABLE call, we need the mask

  kripken

  committed
  91920dc

  Copy full SHA for 91920dc

Commits on Jun 21, 2013

• update tests

  kripken

  committed
  df42061

  Copy full SHA for df42061

Commits on Jun 19, 2013

• fix bug with reducing f()|0|const

  kripken

  committed
  8215a28

  Copy full SHA for 8215a28

• keep |0 on function calls, allow other bitwise ops on heap accesses etc.

  kripken

  committed
  ea9efe3

  Copy full SHA for ea9efe3

• keep coercions on heap accesses and function calls, but fully optimize them otherwise

  kripken

  committed
  e4f8c81

  Copy full SHA for e4f8c81

Commits on Jun 8, 2013

• only remove seq|0 when seq is an assign value

  kripken

  committed
  8cc0453

  Copy full SHA for 8cc0453

• remove unneeded outside |0 when a sequence ends in a safe bitop

  kripken

  committed
  9ea131b

  Copy full SHA for 9ea131b

• optimize away bitcasts to variables that will only be used in an assign to the parallel heap type

  kripken

  committed
  430cfef

  Copy full SHA for 430cfef

• improve tempDoublePtr elimination and only run analysis when necessary

  kripken

  committed
  0f38235

  Copy full SHA for 0f38235

• eliminate bitcasts of immediately loaded values

  kripken

  committed
  8c3c887

  Copy full SHA for 8c3c887

• optimize some HEAP32 expressions, including some unnecessary bitcasts through tempDoublePtr

  kripken

  committed
  9e7ce49

  Copy full SHA for 9e7ce49

Commits on May 22, 2013

• Optimize (x&A)<<B>>B.

  Show description for c571228
  

  Dan Gohman

  committed
  c571228

  Copy full SHA for c571228

Commits on Apr 20, 2013

• optimize HEAPU?8[..] << 24 >> 24 and similar

  kripken

  committed
  6f8af49

  Copy full SHA for 6f8af49

Commits on Mar 10, 2013

• move asm return adding test to right place

  kripken

  committed
  683c5b4

  Copy full SHA for 683c5b4

• remove coercions on asm function calls that have no capturing variables

  kripken

  committed
  b5c3e1b

  Copy full SHA for b5c3e1b

Commits on Feb 4, 2013

• optimize num >> num in js optimizer, necessary for asm now that we do not do shift optimization there

  kripken

  committed
  87c84e8

  Copy full SHA for 87c84e8

Commits on Jan 19, 2013

• fix simplyExpressions bug with removing too many |0s

  kripken

  committed
  425fcf6

  Copy full SHA for 425fcf6

Commits on Jan 9, 2013

• optimize |,& on constants

  kripken

  committed
  d7c3e10

  Copy full SHA for d7c3e10

• improve |0 removal in asm mode a little

  kripken

  committed
  c31a515

  Copy full SHA for c31a515

• keep a coercion right on top of heap accesses in asm mode

  kripken

  committed
  78dbafb

  Copy full SHA for 78dbafb

Loading