On Fri, 25 Apr 2025 09:17:02 GMT, Jatin Bhateja wrote:
>> Thanks for telling me this information. Another more important reason to
>> check outcnt here is to prevent this optimization when the uses of
>> VectorMaskCmp is greater than 1, because this optimization may not be
>> worthwhile. For e
On Mon, 28 Apr 2025 14:12:02 GMT, Emanuel Peter wrote:
> I suppose in that case you can assert that you NEVER get those nodes, because
> if you have vectors not supported, they will not show up because of that, and
> if you do support vectors, they should be optimized away.
This is expected.
On Mon, 28 Apr 2025 07:46:14 GMT, erifan wrote:
>> I would also prefer if you added the IR restrictions rather than the JTREG
>> requires.
>> The benefit is that we can still run the tests on all platforms, at least
>> for result verification.
>>
>> Imagine someone adds optimizations to a new
On Mon, 28 Apr 2025 06:45:58 GMT, Emanuel Peter wrote:
>> This is not specifically required on x86, but because this test fails on x86
>> when `-XX:UseAVX=0` is specified. When `-XX:UseAVX=0` is specified, the
>> sub-graph is like this:
>> `(XorV (VectorMaskCmp (LoadVector ...)) (Replicate -1))
On Sun, 27 Apr 2025 10:09:48 GMT, erifan wrote:
>> I don't see XorVMask implemented on all non-x86 target, like PPC etc..
>
> This is not specifically required on x86, but because this test fails on x86
> when `-XX:UseAVX=0` is specified. When `-XX:UseAVX=0` is specified, the
> sub-graph is lik
On Fri, 25 Apr 2025 09:48:59 GMT, Jatin Bhateja wrote:
>> Since this is a platform independent optimization, I tend to use this
>> `@requires` because it's simpler. If we use `applyIfCPUFeatureOr`, we need
>> to add the same restriction before each test. In addition, if a new
>> architecture
On Thu, 24 Apr 2025 08:57:14 GMT, erifan wrote:
>> erifan has updated the pull request with a new target base due to a merge or
>> a rebase. The incremental webrev excludes the unrelated changes brought in
>> by the merge/rebase. The pull request contains two additional commits since
>> the la
On Thu, 24 Apr 2025 09:46:24 GMT, erifan wrote:
>> test/hotspot/jtreg/compiler/vectorapi/VectorMaskCompareNotTest.java line 38:
>>
>>> 36: * @summary test combining vector not operation with compare
>>> 37: * @modules jdk.incubator.vector
>>> 38: * @requires ((os.arch!="x86" & os.arch!="i386"
On Thu, 24 Apr 2025 09:37:07 GMT, erifan wrote:
>> src/hotspot/share/opto/vectornode.cpp line 2243:
>>
>>> 2241: in1 = in1->in(1);
>>> 2242: }
>>> 2243: if (in1->Opcode() != Op_VectorMaskCmp || in1->outcnt() > 1 ||
>>
>> Checks on outcnt on line 2243 and 2238 can be removed. Idealizatio
On Thu, 24 Apr 2025 10:28:15 GMT, Andrew Haley wrote:
>> erifan has updated the pull request with a new target base due to a merge or
>> a rebase. The incremental webrev excludes the unrelated changes brought in
>> by the merge/rebase. The pull request contains two additional commits since
>>
On Thu, 24 Apr 2025 09:39:58 GMT, erifan wrote:
>> src/hotspot/share/opto/vectornode.cpp line 2265:
>>
>>> 2263: vmcmp = new VectorMaskCastNode(phase->transform(vmcmp),
>>> vmcast_vt);
>>> 2264: }
>>> 2265: return vmcmp;
>>
>> It would be preferable if you could kindly re-factor the co
On Fri, 18 Apr 2025 01:36:10 GMT, erifan wrote:
>> This patch optimizes the following patterns:
>> For integer types:
>>
>> (XorV (VectorMaskCmp src1 src2 cond) (Replicate -1))
>> => (VectorMaskCmp src1 src2 ncond)
>> (XorVMask (VectorMaskCmp src1 src2 cond) (MaskAll m1))
>> => (VectorMa
On Wed, 23 Apr 2025 12:09:51 GMT, Jatin Bhateja wrote:
>> erifan has updated the pull request with a new target base due to a merge or
>> a rebase. The incremental webrev excludes the unrelated changes brought in
>> by the merge/rebase. The pull request contains two additional commits since
>>
On Fri, 18 Apr 2025 01:36:10 GMT, erifan wrote:
>> This patch optimizes the following patterns:
>> For integer types:
>>
>> (XorV (VectorMaskCmp src1 src2 cond) (Replicate -1))
>> => (VectorMaskCmp src1 src2 ncond)
>> (XorVMask (VectorMaskCmp src1 src2 cond) (MaskAll m1))
>> => (VectorMa
On Fri, 18 Apr 2025 01:36:10 GMT, erifan wrote:
>> This patch optimizes the following patterns:
>> For integer types:
>>
>> (XorV (VectorMaskCmp src1 src2 cond) (Replicate -1))
>> => (VectorMaskCmp src1 src2 ncond)
>> (XorVMask (VectorMaskCmp src1 src2 cond) (MaskAll m1))
>> => (VectorMa
On Fri, 18 Apr 2025 01:36:10 GMT, erifan wrote:
>> This patch optimizes the following patterns:
>> For integer types:
>>
>> (XorV (VectorMaskCmp src1 src2 cond) (Replicate -1))
>> => (VectorMaskCmp src1 src2 ncond)
>> (XorVMask (VectorMaskCmp src1 src2 cond) (MaskAll m1))
>> => (VectorMa
> This patch optimizes the following patterns:
> For integer types:
>
> (XorV (VectorMaskCmp src1 src2 cond) (Replicate -1))
> => (VectorMaskCmp src1 src2 ncond)
> (XorVMask (VectorMaskCmp src1 src2 cond) (MaskAll m1))
> => (VectorMaskCmp src1 src2 ncond)
>
> cond can be eq, ne, le, ge, l
17 matches
Mail list logo
