[vector] Test failing: compiler/vectorapi/VectorMaskCastTest.java.

[Paul Sandoz]

Hi,

Reporting that compiler/vectorapi/VectorMaskCastTest.java is failing on AVX512 with the options -XX:UseAVX=3 -XX:+UnlockDiagnosticVMOptions -XX:+UseKNLSetting.

I already reported this directly to Jatin.

Paul.


# JRE version: Java(TM) SE Runtime Environment (18.0+1) (fastdebug build 18-panama+1-60)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (fastdebug 18-panama+1-60, mixed mode, sharing, compressed oops, compressed class ptrs, g1 gc, linux-amd64)
# Problematic frame:
# V  [libjvm.so+0x148375a]  Matcher::Label_Root(Node const*, State*, Node*, Node*&)+0x4ea

…

Host: XXXXXX, Intel(R) Xeon(R) Gold 6354 CPU @ 3.00GHz, 12 cores, 46G, Oracle Linux Server release 7.9
Time: Wed Sep 29 08:41:07 2021 UTC elapsed time: 2.835986 seconds (0d 0h 0m 2s)

…

Current CompileTask:
C2:   2836 1086             compiler.vectorapi.VectorMaskCastTest::testInt256ToDouble512 (27 bytes)

Stack: [0x00007fed388f9000,0x00007fed389fa000],  sp=0x00007fed389f4b60,  free space=1006k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
V  [libjvm.so+0x148375a]  Matcher::Label_Root(Node const*, State*, Node*, Node*&)+0x4ea
V  [libjvm.so+0x148387b]  Matcher::Label_Root(Node const*, State*, Node*, Node*&)+0x60b
V  [libjvm.so+0x14841ac]  Matcher::match_tree(Node const*)+0x20c
V  [libjvm.so+0x1492ff5]  Matcher::xform(Node*, int)+0xf55
V  [libjvm.so+0x149c06a]  Matcher::match()+0x490a
V  [libjvm.so+0xa7859c]  Compile::Code_Gen()+0x9c
V  [libjvm.so+0xa84d02]  Compile::Compile(ciEnv*, ciMethod*, int, bool, bool, bool, bool, bool, DirectiveSet*)+0x1652
V  [libjvm.so+0x89cca6]  C2Compiler::compile_method(ciEnv*, ciMethod*, int, bool, DirectiveSet*)+0x646
V  [libjvm.so+0xa94e89]  CompileBroker::invoke_compiler_on_method(CompileTask*)+0xc39
V  [libjvm.so+0xa95b68]  CompileBroker::compiler_thread_loop()+0x598
V  [libjvm.so+0x191dcbc]  JavaThread::thread_main_inner()+0x27c
V  [libjvm.so+0x19242f0]  Thread::call_run()+0x100
V  [libjvm.so+0x1601f44]  thread_native_entry(Thread*)+0x104