RFR: 8294715: Add IR checks to the reduction vectorization tests [v3]

[Fei Gao]

On Tue, 28 Feb 2023 08:03:20 GMT, Daniel Skantz <duke at openjdk.org> wrote:

>> We are lifting some loopopts/superword tests to use the IR framework, and add IR annotations to check that vector reductions take place on x86_64. This can be useful to prevent issues such as JDK-8300865.
>> 
>> Approach: lift the more general tests in loopopts/superword, mainly using matching rules in cpu/x86/x86.ad, but leave tests mostly unchanged otherwise. Some reductions are considered non-profitable (superword.cpp), so we might need to raise sse/avx value pre-conditions from what would be a strict reading of x86.ad (as noted by @eme64).
>> 
>> Testing: Local testing (x86_64) using UseSSE={2,3,4}, UseAVX={0,1,2,3}. Tested running all jtreg compiler tests. Tier1-tier5 runs to my knowledge never showed any compiler-related regression in other tests as a result from this work. GHA. Validation: all tests fail if we put unreasonable counts for the respective reduction node, such as counts = {IRNode.ADD_REDUCTION_VI, ">= 10000000"}).
>> 
>> Thanks @robcasloz  and @eme64 for advice.
>> 
>> Notes: ProdRed_Double does not vectorize (JDK-8300865). SumRed_Long does not vectorize on 32-bit, according to my reading of source, test on GHA and cross-compiled JDK on 32-bit Linux, so removed these platforms from @requires. Lifted the AbsNeg tests too but added no checks, as these are currently not run on x86_64.
>
> Daniel Skantz has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Address further review comments (edits)

Changes requested by fgao (Committer).

test/hotspot/jtreg/compiler/loopopts/superword/RedTest_long.java line 150:

> 148:     @IR(applyIfOr = {"SuperWordReductions", "false", "LoopMaxUnroll", "< 8"},
> 149:         failOn = {IRNode.ADD_REDUCTION_VL})
> 150:     @IR(applyIfCPUFeature = {"avx2", "false"},

I'm afraid the restriction here is not safe because all non-avx2 machines would meet the condition, like aarch64 and riscv64. SVE supports `AndReductionV`, `OrReductionV `, `XorReductionV`, `AddReductionVL`, `AddReductionVD`, `SqrtVD`, and so on.

test/hotspot/jtreg/compiler/loopopts/superword/RedTest_long.java line 172:

> 170:     @IR(applyIfOr = {"SuperWordReductions", "false", "LoopMaxUnroll", "< 8"},
> 171:         failOn = {IRNode.OR_REDUCTION_V})
> 172:     @IR(applyIfCPUFeature = {"avx2", "false"},

ditto

test/hotspot/jtreg/compiler/loopopts/superword/RedTest_long.java line 194:

> 192:     @IR(applyIfOr = {"SuperWordReductions", "false", "LoopMaxUnroll", "< 8"},
> 193:         failOn = {IRNode.AND_REDUCTION_V})
> 194:     @IR(applyIfCPUFeature = {"avx2", "false"},

ditto

test/hotspot/jtreg/compiler/loopopts/superword/RedTest_long.java line 216:

> 214:     @IR(applyIfOr = {"SuperWordReductions", "false", "LoopMaxUnroll", "< 8"},
> 215:         failOn = {IRNode.XOR_REDUCTION_V})
> 216:     @IR(applyIfCPUFeature = {"avx2", "false"},

ditto

test/hotspot/jtreg/compiler/loopopts/superword/SumRedSqrt_Double.java line 95:

> 93:     @IR(applyIfOr = {"SuperWordReductions", "false", "LoopMaxUnroll", "< 8"},
> 94:         failOn = {IRNode.ADD_REDUCTION_VD, IRNode.SQRT_V})
> 95:     @IR(applyIfCPUFeature = {"avx", "false"},

ditto

test/hotspot/jtreg/compiler/loopopts/superword/SumRed_Long.java line 98:

> 96:     @IR(applyIfOr = {"SuperWordReductions", "false", "LoopMaxUnroll", "< 8"},
> 97:         failOn = {IRNode.ADD_REDUCTION_VL})
> 98:     @IR(applyIfCPUFeature = {"avx2", "false"},

ditto

-------------

PR: https://git.openjdk.org/jdk/pull/12683