jdk
8307513: C2: intrinsify Math.max(long,long) and Math.min(long,long)
#20098
Open
Overview
Commits
36
Changes
View On GitHub
Commits
  • 8307513: C2: intrinsify Math.max(long,long) and Math.min(long,long)
    galderz committed 169 days ago
  • Add IR test
    galderz committed 160 days ago
  • Refactor inline methods to unify their implementations
    galderz committed 159 days ago
  • Add math vectorized JMH benchmark
    galderz committed 155 days ago
  • Rename benchmark class to MathLoopBench
    galderz committed 120 days ago
  • Fix multi long tests to use long arrays
    galderz committed 120 days ago
  • Implement cmovL as a jump+mov branch
    galderz committed 107 days ago
  • Switch movl to movq
    galderz committed 105 days ago
  • Fix format of assembly for the movl to movq switch
    galderz committed 105 days ago
  • Distribute values targetting a branch percentage
    galderz committed 104 days ago
  • Fix min case to distribute numbers as per probability
    galderz committed 104 days ago
  • Fix compilation error
    galderz committed 104 days ago
  • Add an intermediate % that is more representative of real life
    galderz committed 104 days ago
  • Skip single array benchmarks
    galderz committed 100 days ago
  • Add min/max benchmark that includes loops and reductions
    galderz committed 92 days ago
  • Renamed benchmark methods
    galderz committed 92 days ago
  • Multiply array value in reduction for vectorization to kick in
    galderz committed 91 days ago
  • Remove previous benchmark effort
    galderz committed 89 days ago
  • Revert "Fix format of assembly for the movl to movq switch"
    galderz committed 89 days ago
  • Revert "Switch movl to movq"
    galderz committed 89 days ago
  • Revert "Implement cmovL as a jump+mov branch"
    galderz committed 89 days ago
  • Merge branch 'master' into topic.intrinsify-max-min-long
    galderz committed 69 days ago
  • Avoid creating result array in benchmark method
    galderz committed 69 days ago
  • Encapsulate benchmark state within an inner class
    galderz committed 69 days ago
  • Add clipping range benchmark that uses min/max
    galderz committed 69 days ago
  • Restore previous benchmark iterations and default param size
    galderz committed 69 days ago
  • Make state class non-final
    galderz committed 69 days ago
  • Double/Float tests only when avx enabled
    galderz committed 69 days ago
  • Renamed benchmark class
    galderz committed 69 days ago
  • Use same default size as in other vector reduction benchmarks
    galderz committed 69 days ago
  • Merge branch 'master' into topic.intrinsify-max-min-long
    galderz committed 13 days ago
  • Add empty line
    galderz committed 8 days ago
  • Add max reduction test
    galderz committed 8 days ago
  • Fix style
    galderz committed 8 days ago
  • Adjust min/max identity IR test expectations after changes
    galderz committed 7 days ago
  • Added comment around the assertions
    galderz committed 7 days ago
FAQ Terms Privacy Refunds Impressum
Loading