Accelerate ArraysSupport.vectorizedMismatch in IL #16662

Spencer-Comin · 2023-02-02T21:51:24Z

This patch adds ArraysSupport.vectorizedMismatch as a recognized method. If the arraycmp IL opcode is supported, vectorizedMismatch call nodes are transformed to a functionally equivalent tree using arraycmp.

vectorizedMismatch is functionally equivalent to:

   lengthInBytes = length << log2ArrayIndexScale
   mask = (log2ArrayIndexScale<2) ? 3 : 7
   n = lengthInBytes & ~(mask)
   res = arrayCmpLen(a+aOffset, b+bOffset, n)
   if (res == n)  // no mismatch found
      return ~((lengthInBytes & mask) >> log2ArrayIndexScale)
   else           // mismatch found
      return res >> log2ArrayIndexScale

Therefore the call nodes can be translated like so:

n9n    icall  jdk/internal/util/ArraysSupport.vectorizedMismatch(Ljava/lang/Object;JLjava/lang/Object;JII)I
n3n      <a>
n4n      <aOffset>
n5n      <b>
n6n      <bOffset>
n7n      <length>
n8n      <log2ArrayIndexScale>

becomes

n9n         iselect ()
n31n          lcmpeq
n22n            arraycmp (arrayCmpLen )
n23n              aladd
n3n                 <a>
n4n                 <aOffset>
n24n              aladd
n5n                 <b>
n6n                 <bOffset>
n21n              land
n14n                lshl
n13n                  i2l
n7n                     <length>
n12n                  i2l
n8n                     <log2ArrayIndexScale>
n20n                lxor
n18n                  lor
n16n                    lshl
n12n                      ==>i2l
n15n                      lconst 1
n17n                    lconst 3
n19n                  lconst -1
n21n            ==>land
n29n          ixor
n27n            l2i
n26n              lshr
n25n                land
n14n                  ==>lshl
n18n                  ==>lor
n12n                ==>i2l
n28n            iconst -1
n30n          ishr
n22n            ==>arraycmp
n8n             <log2ArrayIndexScale>

ArraysSupport.vectorizedMismatch is used in the implementations of Arrays.mismatch, Arrays.compare, and Arrays.equals for primitive types.

Benchmarking Arrays.compare on x86, Power, and Z for 128, 4096, and 8192 element arrays of all primitive types shows a speedup of around 4.2× on x86, 2.7× on Power, and 4.1× on Z. Benchmark arrays of 1 or 2 elements shows no significant slowdown, and in some cases a speedup (notably, I saw a speedup of 3× comparing two element arrays of long).
Full benchmarking results here, for the curious: arraycompare-results.zip

Fixes: #15204

r30shah

@Spencer-Comin, Can you upload a sample log file printing trees before and after this particular optimization?

runtime/compiler/codegen/J9RecognizedMethodsEnum.hpp

runtime/compiler/optimizer/J9RecognizedCallTransformer.cpp

Spencer-Comin · 2023-02-06T22:00:52Z

@Spencer-Comin, Can you upload a sample log file printing trees before and after this particular optimization?

vectorizedMismatch.log.txt

runtime/compiler/optimizer/J9RecognizedCallTransformer.cpp

runtime/compiler/optimizer/J9RecognizedCallTransformer.hpp

runtime/compiler/p/codegen/J9CodeGenerator.cpp

Spencer-Comin · 2023-02-09T15:28:15Z

runtime/compiler/x/codegen/J9CodeGenerator.cpp

@@ -134,6 +134,12 @@ J9::X86::CodeGenerator::initialize()
      cg->setSupportsBDLLHardwareOverflowCheck();
      }

+   static char *disableInlineVectorizedMismatch = feGetEnv("TR_disableInlineVectorizedMismatch");


@0xdaryl I've only considered arraycmp support in enabling this acceleration on x. Is there anything else to take into consideration here?

r30shah

Changes looks OK to me apart from couple of minor nitpicks in comments. @0xdaryl Can I request you to give your review as well to this change?

runtime/compiler/optimizer/J9RecognizedCallTransformer.hpp

runtime/compiler/z/codegen/J9CodeGenerator.cpp

r30shah · 2023-02-16T15:48:24Z

@Spencer-Comin Do we have any functional tests for these changes ? OR we should contribute one ?

Spencer-Comin · 2023-02-16T18:28:22Z

Nothing I see in OpenJ9. There are JDK tests that use Arrays.mismatch and Arrays.compare.

$ grep -rnw `pwd` -e 'Arrays\.compare'
jdk11/test/jdk/java/net/httpclient/http2/server/Http2TestServerConnection.java:309:        if (Arrays.compare(clientPreface, bytes) != 0) {
jdk11/test/jdk/java/util/Arrays/ArraysEqCmpTest.java:755:        Assert.assertTrue(Arrays.compare(a, b) < 0);
jdk11/test/jdk/java/util/Arrays/ArraysEqCmpTest.java:756:        Assert.assertTrue(Arrays.compare(b, a) > 0);
jdk11/test/jdk/java/util/Arrays/ArraysEqCmpTest.java:1037:                        testNPE(() -> Arrays.compare(o1, o2, o3));
jdk11/test/jdk/java/util/Arrays/ArraysEqCmpTest.java:1042:                    testNPE(() -> Arrays.compare(o1, 0, 0, o2, 0, 0, o3));
jdk11/test/jdk/java/nio/file/Files/ReadWriteString.java:218:        assertTrue((Arrays.compare(bytes, bytes2) == 0), "The bytes should be the same");
jdk11/test/jdk/com/sun/crypto/provider/Cipher/DES/DESKeyCleanupTest.java:70:        } while (Arrays.compare(zeros, array) != 0);
jdk11/test/jdk/com/sun/crypto/provider/Cipher/PBE/PBEKeyCleanupTest.java:94:        } while (Arrays.compare(zeros, array) != 0);
jdk11/test/lib/jdk/test/lib/OSVersion.java:89:        return Arrays.compare(this.versionTokens, o.versionTokens);
jdk11/test/lib/jdk/test/lib/containers/cgroup/MetricsTesterCgroupV1.java:434:        if (Arrays.compare(oldVal, newVal) != 0) {
jdk11/test/lib/jdk/test/lib/containers/cgroup/MetricsTesterCgroupV1.java:446:        if (Arrays.compare(oldVal, newVal) != 0) {
jdk11/test/lib/jdk/test/lib/containers/cgroup/MetricsTesterCgroupV1.java:456:        if (Arrays.compare(oldVal, newVal) != 0) {
jdk11/test/lib/jdk/test/lib/containers/cgroup/MetricsTesterCgroupV1.java:468:        if (Arrays.compare(oldVal, newVal) != 0) {
jdk11/test/lib/jdk/test/lib/containers/cgroup/MetricsTesterCgroupV2.java:357:        if (Arrays.compare(oldVal, newVal) != 0) {
jdk11/test/lib/jdk/test/lib/containers/cgroup/MetricsTesterCgroupV2.java:367:        if (Arrays.compare(oldVal, newVal) != 0) {
jdk11/test/lib/jdk/test/lib/containers/cgroup/MetricsTesterCgroupV2.java:377:        if (Arrays.compare(oldVal, newVal) != 0) {
jdk11/test/lib/jdk/test/lib/containers/cgroup/MetricsTesterCgroupV2.java:387:        if (Arrays.compare(oldVal, newVal) != 0) {

$ grep -rnw `pwd` -e 'Arrays\.mismatch'
jdk11/test/jdk/tools/jmod/hashes/HashesOrderTest.java:150:                int i = Arrays.mismatch(buffer1, 0, nRead1, buffer2, 0, nRead2);
jdk11/test/jdk/tools/jlink/JLinkReproducibleTest.java:134:                int i = Arrays.mismatch(buffer1, 0, nRead1, buffer2, 0, nRead2);
jdk11/test/jdk/java/util/Arrays/ArraysEqCmpTest.java:1038:                        testNPE(() -> Arrays.mismatch(o1, o2, o3));
jdk11/test/jdk/java/util/Arrays/ArraysEqCmpTest.java:1043:                    testNPE(() -> Arrays.mismatch(o1, 0, 0, o2, 0, 0, o3));
jdk11/test/jdk/java/rmi/testlibrary/TestSocketFactory.java:832:            long index = Arrays.mismatch(actual, expected);
jdk11/test/jdk/javax/xml/crypto/dsig/Versions.java:56:                int i = Arrays.mismatch(buffer1, 0, nRead1, buffer2, 0, nRead2);
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressBooleanArrayCopy.java:41:            int m = Arrays.mismatch(test, 0, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressBooleanArrayCopy.java:53:            int m = Arrays.mismatch(test, r, r+len,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressBooleanArrayCopy.java:62:            int m = Arrays.mismatch(test, 0, r,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressBooleanArrayCopy.java:69:            int m = Arrays.mismatch(test, r + len, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressIntArrayCopy.java:41:            int m = Arrays.mismatch(test, 0, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressIntArrayCopy.java:53:            int m = Arrays.mismatch(test, r, r+len,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressIntArrayCopy.java:62:            int m = Arrays.mismatch(test, 0, r,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressIntArrayCopy.java:69:            int m = Arrays.mismatch(test, r + len, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressLongArrayCopy.java:41:            int m = Arrays.mismatch(test, 0, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressLongArrayCopy.java:53:            int m = Arrays.mismatch(test, r, r+len,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressLongArrayCopy.java:62:            int m = Arrays.mismatch(test, 0, r,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressLongArrayCopy.java:69:            int m = Arrays.mismatch(test, r + len, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressCharArrayCopy.java:41:            int m = Arrays.mismatch(test, 0, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressCharArrayCopy.java:53:            int m = Arrays.mismatch(test, r, r+len,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressCharArrayCopy.java:62:            int m = Arrays.mismatch(test, 0, r,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressCharArrayCopy.java:69:            int m = Arrays.mismatch(test, r + len, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressObjectArrayCopy.java:41:            int m = Arrays.mismatch(test, 0, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressObjectArrayCopy.java:53:            int m = Arrays.mismatch(test, r, r+len,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressObjectArrayCopy.java:62:            int m = Arrays.mismatch(test, 0, r,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressObjectArrayCopy.java:69:            int m = Arrays.mismatch(test, r + len, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressByteArrayCopy.java:41:            int m = Arrays.mismatch(test, 0, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressByteArrayCopy.java:53:            int m = Arrays.mismatch(test, r, r+len,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressByteArrayCopy.java:62:            int m = Arrays.mismatch(test, 0, r,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressByteArrayCopy.java:69:            int m = Arrays.mismatch(test, r + len, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressFloatArrayCopy.java:41:            int m = Arrays.mismatch(test, 0, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressFloatArrayCopy.java:53:            int m = Arrays.mismatch(test, r, r+len,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressFloatArrayCopy.java:62:            int m = Arrays.mismatch(test, 0, r,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressFloatArrayCopy.java:69:            int m = Arrays.mismatch(test, r + len, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressShortArrayCopy.java:41:            int m = Arrays.mismatch(test, 0, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressShortArrayCopy.java:53:            int m = Arrays.mismatch(test, r, r+len,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressShortArrayCopy.java:62:            int m = Arrays.mismatch(test, 0, r,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressShortArrayCopy.java:69:            int m = Arrays.mismatch(test, r + len, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressDoubleArrayCopy.java:41:            int m = Arrays.mismatch(test, 0, size,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressDoubleArrayCopy.java:53:            int m = Arrays.mismatch(test, r, r+len,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressDoubleArrayCopy.java:62:            int m = Arrays.mismatch(test, 0, r,
jdk11/test/hotspot/jtreg/compiler/arraycopy/stress/StressDoubleArrayCopy.java:69:            int m = Arrays.mismatch(test, r + len, size,

r30shah · 2023-02-24T21:29:29Z

@0xdaryl Can we get your review on these changes?

r30shah · 2023-03-07T16:49:25Z

@0xdaryl / @jdmpapin Can I request your review on this change from @Spencer-Comin ?

Spencer-Comin · 2023-03-15T14:33:43Z

With aarch64 now supporting arraycmp in eclipse/omr#6904 I've enabled this acceleration there too. @Akira1Saitoh could you review this?

Akira1Saitoh · 2023-03-16T08:57:31Z

@Spencer-Comin Thanks. Your change looks good to me.

0xdaryl

A couple of minor nits. Generally looks ok.

runtime/compiler/codegen/J9CodeGenerator.hpp

0xdaryl · 2023-03-22T14:29:03Z

Jenkins test sanity all jdk17

runtime/compiler/optimizer/J9RecognizedCallTransformer.hpp

Spencer-Comin · 2023-04-14T19:45:04Z

Converting this to draft while we resolve the size restriction on arraycmp.
See eclipse/omr#6951 for issue tracking progress on that, and #16662 (comment) for context on why this issue blocks this PR.

Spencer-Comin · 2023-06-19T20:09:51Z

I've rewritten this to use arraycmplen, so this PR is now dependent on eclipse/omr#6983 and indirectly #17382

r30shah

Already reviewed this, so just went through again, looks good to me. @jdmpapin as the PR for arrayCmpLen will be merged in coming days, requesting your review on this one.

runtime/compiler/optimizer/J9RecognizedCallTransformer.cpp

r30shah · 2023-09-01T12:18:53Z

runtime/compiler/z/codegen/J9CodeGenerator.cpp

@@ -124,6 +124,12 @@ J9::Z::CodeGenerator::initialize()
      cg->setSupportsInlineEncodeASCII();
      }

+   static bool disableInlineVectorizedMismatch = feGetEnv("TR_disableInlineVectorizedMismatch") != NULL;
+   if (cg->getSupportsArrayCmpLen() && !disableInlineVectorizedMismatch)


getSupportsArrayCmpLen would be disabled for arraylets and off-heap, right? Just confirming.

eclipse/omr#6983 enables arraycmplen in the same way arraycmp is enabled on each platform:

AArch64 - disabled for arraylets, also disabled by environment variable TR_aarch64DisableArrayCmpLen

Power - always enabled

x86 - disabled for arraylets

Z - always enabled

arraycmplen should probably be uniformly disabled for arraylets, and maybe by a single environment variable for all platforms. Same for arraycmp. ~~I'll add a commit to eclipse/omr#6983 to make arraycmp enablement more uniform and amend the commit introducing arraycmplen~~. Edit: To not put another obstacle to the beleaguered arraycmplen PRs, I'll make this change in its own PR (eclipse/omr#7108).

Not sure how off-heap will affect arraycmp(len) enablement, @VermaSh any input?

r30shah

LGTM, @jdmpapin Would appreciate if you can give your review on these changes as well and if OK, merge it.

r30shah · 2023-09-19T14:49:20Z

@Spencer-Comin seems like there has been merge conflicts. Can you resolve them ?

@jdmpapin As arraycmp opcode changes have been merged, Can we get your review on these changes?

runtime/compiler/optimizer/J9RecognizedCallTransformer.hpp

runtime/compiler/optimizer/J9RecognizedCallTransformer.cpp

runtime/compiler/optimizer/J9RecognizedCallTransformer.hpp

jdmpapin · 2023-09-19T19:48:13Z

Thanks! I suppose I should have also mentioned the places in the comment that show 64-bit shift amounts, since those are now inconsistent with the implementation

This patch adds ArraysSupport.vectorizedMismatch as a recognized method, and adds a SupportsInlineVectorizedMismatch flag to the code generator. This flag is set in Z, Power, aarch64 and x86 code generator initialization if the arraycmp opcode is supported and the TR_disableInlineVectorizedMismatch environment variable is not set. If the flag is set, vectorizedMismatch call nodes are transformed to a functionally equivalent tree that uses arraycmp. Fixes: eclipse-openj9#15204 Signed-off-by: Spencer Comin <spencer.comin@ibm.com>

Spencer-Comin · 2023-10-25T17:22:21Z

I've resolved the conflict, @jdmpapin is there anything left to do before this PR can begin the merge process?

jdmpapin · 2023-10-25T19:10:20Z

Due to the state of PR builds, I'm coordinating with Spencer to test this PR privately

jdmpapin · 2023-10-27T19:18:20Z

This has now passed JDK17 sanity.functional on Linux with x86-64, PPC64LE, z, and AArch64, with the exception of a number of unrelated CRIU test failures caused by machine configuration. Merging

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from 53e32e6 to 256dd74 Compare February 2, 2023 21:54

r30shah requested changes Feb 6, 2023

View reviewed changes

runtime/compiler/codegen/J9RecognizedMethodsEnum.hpp Show resolved Hide resolved

r30shah requested changes Feb 6, 2023

View reviewed changes

runtime/compiler/optimizer/J9RecognizedCallTransformer.cpp Outdated Show resolved Hide resolved

runtime/compiler/optimizer/J9RecognizedCallTransformer.cpp Show resolved Hide resolved

runtime/compiler/optimizer/J9RecognizedCallTransformer.cpp Outdated Show resolved Hide resolved

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch 2 times, most recently from ff8662b to 693be7e Compare February 7, 2023 22:10

r30shah requested changes Feb 8, 2023

View reviewed changes

runtime/compiler/optimizer/J9RecognizedCallTransformer.cpp Outdated Show resolved Hide resolved

runtime/compiler/optimizer/J9RecognizedCallTransformer.cpp Outdated Show resolved Hide resolved

runtime/compiler/optimizer/J9RecognizedCallTransformer.hpp Show resolved Hide resolved

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from 693be7e to 3fbd0ea Compare February 9, 2023 15:14

Spencer-Comin commented Feb 9, 2023

View reviewed changes

runtime/compiler/p/codegen/J9CodeGenerator.cpp Outdated Show resolved Hide resolved

Spencer-Comin commented Feb 9, 2023

View reviewed changes

VermaSh mentioned this pull request Feb 14, 2023

Update Accelerate ArraysSupport.vectorizedMismatch changes to work with off-heap #16717

Open

r30shah approved these changes Feb 15, 2023

View reviewed changes

runtime/compiler/optimizer/J9RecognizedCallTransformer.hpp Outdated Show resolved Hide resolved

runtime/compiler/z/codegen/J9CodeGenerator.cpp Outdated Show resolved Hide resolved

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from 3fbd0ea to f48cfcc Compare February 16, 2023 14:13

r30shah approved these changes Feb 16, 2023

View reviewed changes

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from f48cfcc to c27af73 Compare March 15, 2023 14:30

0xdaryl reviewed Mar 22, 2023

View reviewed changes

runtime/compiler/codegen/J9CodeGenerator.hpp Outdated Show resolved Hide resolved

runtime/compiler/codegen/J9CodeGenerator.hpp Outdated Show resolved Hide resolved

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from c27af73 to 5b0c361 Compare March 22, 2023 14:14

jdmpapin reviewed Mar 22, 2023

View reviewed changes

runtime/compiler/optimizer/J9RecognizedCallTransformer.hpp Outdated Show resolved Hide resolved

runtime/compiler/optimizer/J9RecognizedCallTransformer.hpp Show resolved Hide resolved

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch 4 times, most recently from c4373c8 to 5049d1e Compare March 29, 2023 20:19

Spencer-Comin mentioned this pull request Apr 14, 2023

Create arraycmplen opcode to replace arraycmp + arrayCmpLen flag eclipse/omr#6951

Closed

Spencer-Comin marked this pull request as draft April 14, 2023 19:42

knn-k added the comp:jit label May 24, 2023

r30shah mentioned this pull request Jun 15, 2023

Panama MemoryAccessAPI: Add jit intrinsic support for vectorizedMismatch #15204

Closed

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from 5049d1e to 1ff5372 Compare June 19, 2023 20:04

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from 1ff5372 to dd490e9 Compare August 29, 2023 17:26

r30shah reviewed Sep 1, 2023

View reviewed changes

This was referenced Sep 6, 2023

Create arraycmplen opcode to replace arraycmp + arrayCmpLen flag eclipse/omr#6983

Merged

Enable arraycmp and arraycmplen uniformly eclipse/omr#7108

Merged

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from dd490e9 to 517f288 Compare September 6, 2023 16:41

Spencer-Comin marked this pull request as ready for review September 7, 2023 21:08

r30shah approved these changes Sep 12, 2023

View reviewed changes

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from 517f288 to de7ee82 Compare September 19, 2023 17:02

jdmpapin reviewed Sep 19, 2023

View reviewed changes

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from de7ee82 to dcc15e3 Compare September 19, 2023 19:42

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from dcc15e3 to b83f994 Compare September 19, 2023 19:50

jdmpapin approved these changes Sep 19, 2023

View reviewed changes

BradleyWood mentioned this pull request Jul 13, 2023

Implement various intrinsics in the JIT #17778

Open

Spencer-Comin force-pushed the vectorizedMismatch-iselect branch from b83f994 to 0636989 Compare October 25, 2023 17:21

jdmpapin merged commit d62e757 into eclipse-openj9:master Oct 27, 2023
2 checks passed

tajila mentioned this pull request Oct 30, 2023

JDK21 jdk_lang_1_FAILED sun/misc/URLClassPath/LargeClasspathWithPkgPrefix.java Exception in thread "main" java.lang.StackOverflowError #18337

Closed

pshipton mentioned this pull request Oct 30, 2023

OpenJDK jdk/modules/etc/DefaultModules, jdk/modules/incubator/ImageModules - IOException: module-info.class not found #18371

Closed

ChengJin01 mentioned this pull request Feb 23, 2024

jdk_foreign_0_FAILED - java/foreign/TestHandshake.java #13211

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Accelerate ArraysSupport.vectorizedMismatch in IL #16662

Accelerate ArraysSupport.vectorizedMismatch in IL #16662

Spencer-Comin commented Feb 2, 2023 •

edited

r30shah left a comment

Spencer-Comin commented Feb 6, 2023 •

edited

Spencer-Comin Feb 9, 2023

r30shah left a comment

r30shah commented Feb 16, 2023

Spencer-Comin commented Feb 16, 2023

r30shah commented Feb 24, 2023

r30shah commented Mar 7, 2023

Spencer-Comin commented Mar 15, 2023

Akira1Saitoh commented Mar 16, 2023

0xdaryl left a comment

0xdaryl commented Mar 22, 2023

Spencer-Comin commented Apr 14, 2023

Spencer-Comin commented Jun 19, 2023

r30shah left a comment

r30shah Sep 1, 2023

Spencer-Comin Sep 6, 2023 •

edited

r30shah left a comment

r30shah commented Sep 19, 2023

jdmpapin commented Sep 19, 2023

Spencer-Comin commented Oct 25, 2023

jdmpapin commented Oct 25, 2023

jdmpapin commented Oct 27, 2023

Accelerate ArraysSupport.vectorizedMismatch in IL #16662

Accelerate ArraysSupport.vectorizedMismatch in IL #16662

Conversation

Spencer-Comin commented Feb 2, 2023 • edited

r30shah left a comment

Choose a reason for hiding this comment

Spencer-Comin commented Feb 6, 2023 • edited

Spencer-Comin Feb 9, 2023

Choose a reason for hiding this comment

r30shah left a comment

Choose a reason for hiding this comment

r30shah commented Feb 16, 2023

Spencer-Comin commented Feb 16, 2023

r30shah commented Feb 24, 2023

r30shah commented Mar 7, 2023

Spencer-Comin commented Mar 15, 2023

Akira1Saitoh commented Mar 16, 2023

0xdaryl left a comment

Choose a reason for hiding this comment

0xdaryl commented Mar 22, 2023

Spencer-Comin commented Apr 14, 2023

Spencer-Comin commented Jun 19, 2023

r30shah left a comment

Choose a reason for hiding this comment

r30shah Sep 1, 2023

Choose a reason for hiding this comment

Spencer-Comin Sep 6, 2023 • edited

Choose a reason for hiding this comment

r30shah left a comment

Choose a reason for hiding this comment

r30shah commented Sep 19, 2023

jdmpapin commented Sep 19, 2023

Spencer-Comin commented Oct 25, 2023

jdmpapin commented Oct 25, 2023

jdmpapin commented Oct 27, 2023

Spencer-Comin commented Feb 2, 2023 •

edited

Spencer-Comin commented Feb 6, 2023 •

edited

Spencer-Comin Sep 6, 2023 •

edited