Skip to content

Improve the vectorizedHashCode Intrinsic for P9+ #21338

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 10 commits into from
Closed

Improve the vectorizedHashCode Intrinsic for P9+ #21338

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
8b2a5bc
Fast-Path String and Vector Hash Code Methods on Power
luke-li-2003 Jan 31, 2025
5e60c2c
Improve the vectorizedHashCode Intrinsic for P10
luke-li-2003 Mar 11, 2025
170cee1
Enable P10
luke-li-2003 Mar 12, 2025
b803789
remove trailing whitespaces
luke-li-2003 Mar 12, 2025
561bc85
special2
luke-li-2003 Mar 14, 2025
bf4b8ed
quickfix
luke-li-2003 Mar 14, 2025
e2823a1
modified impl
luke-li-2003 Mar 14, 2025
73a1f63
guard for special case
luke-li-2003 Mar 17, 2025
5b5afc6
modernisation
luke-li-2003 Apr 24, 2025
11c66d1
bug fix
luke-li-2003 Apr 29, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 4 additions & 14 deletions runtime/compiler/optimizer/InlinerTempForJ9.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5592,26 +5592,16 @@ TR_J9InlinerPolicy::supressInliningRecognizedInitialCallee(TR_CallSite* callsite
return false;
}
case TR::java_lang_String_hashCodeImplDecompressed:
/*
* X86 and z want to avoid inlining both java_lang_String_hashCodeImplDecompressed and java_lang_String_hashCodeImplCompressed
* so they can be recognized and replaced with a custom fast implementation.
* Power currently only has the custom fast implementation for java_lang_String_hashCodeImplDecompressed.
* As a result, Power only wants to prevent inlining of java_lang_String_hashCodeImplDecompressed.
* When Power gets a fast implementation of TR::java_lang_String_hashCodeImplCompressed, this case can be merged into the case
* for java_lang_String_hashCodeImplCompressed instead of using a fallthrough.
*/
if (!TR::Compiler->om.canGenerateArraylets() && !TR::Compiler->om.isOffHeapAllocationEnabled() &&
comp->target().cpu.isPower() && comp->target().cpu.isAtLeast(OMR_PROCESSOR_PPC_P8) && comp->target().cpu.supportsFeature(OMR_FEATURE_PPC_HAS_VSX) && !comp->compileRelocatableCode())
{
return true;
}
// Intentional fallthrough here.
case TR::java_lang_String_hashCodeImplCompressed:
// X86, Z and Power have custom fast implementations for these 2 methods
// so their inlining should be avoided.
{
if (comp->cg()->getSupportsInlineStringHashCode())
{
return true;
}
break;
}
case TR::jdk_internal_util_ArraysSupport_vectorizedHashCode:
{
if (comp->cg()->getSupportsInlineVectorizedHashCode())
Expand Down
14 changes: 14 additions & 0 deletions runtime/compiler/p/codegen/J9CodeGenerator.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -123,6 +123,20 @@ J9::Power::CodeGenerator::initialize()
if (!comp->getOption(TR_DisableReadMonitors))
cg->setSupportsReadOnlyLocks();

if (!TR::Compiler->om.canGenerateArraylets()
&& !TR::Compiler->om.isOffHeapAllocationEnabled()
&& comp->target().cpu.isAtLeast(OMR_PROCESSOR_PPC_P8)
&& comp->target().cpu.supportsFeature(OMR_FEATURE_PPC_HAS_VSX)
&& !comp->compileRelocatableCode()
#ifdef J9VM_OPT_JITSERVER
&& !comp->isOutOfProcessCompilation()
#endif
)
{
cg->setSupportsInlineStringHashCode();
cg->setSupportsInlineVectorizedHashCode();
}

static bool disableTLHPrefetch = (feGetEnv("TR_DisableTLHPrefetch") != NULL);

// Enable software prefetch of the TLH and configure the TLH prefetching
Expand Down
Loading