Mercurial > hg > graal-jvmci-8
graph
-
changes to support Graal co-existing with the other HotSpot compiler(s) and being used for explicit compilation requests and code installation via the Graal APIWed, 12 Dec 2012 21:36:40 +0100, by Doug Simon
-
made initialization of GraphPrinterDumpHandler.sdf be lazy to workaround an initialization order issue during VM start upWed, 12 Dec 2012 15:46:11 +0100, by Doug Simon
-
fixed issues triggered when using the CountingProxy (i.e., -Dgraal.countcalls=true) or LoggingProxy (i.e., -Dgraal.debug=true) to analyze traffic across the VM/compiler boundaryTue, 11 Dec 2012 20:54:11 +0100, by Doug Simon
-
Merge.Tue, 11 Dec 2012 08:29:25 +0100, by Doug Simon
-
branch to isolate fix for perf. regressionTue, 11 Dec 2012 08:27:01 +0100, by Doug Simon
-
made the Graal implementation of the C++ AbstractCompiler class support native wrapper generationMon, 10 Dec 2012 22:19:07 +0100, by Doug Simon
-
removed declaration of temps (no longer) used by the monitorenter/monitorexits stubsFri, 07 Dec 2012 23:41:51 +0100, by Doug Simon
-
added TLAB fast refill stub for array allocationFri, 07 Dec 2012 23:38:09 +0100, by Doug Simon
-
removed unsafe injection of a new object constant into compiler graphFri, 07 Dec 2012 23:35:23 +0100, by Doug Simon
-
consolidated new_type_array and new_object_array stubs into one as there no difference between themFri, 07 Dec 2012 18:26:26 +0100, by Doug Simon
-
rename: RuntimeCall -> RuntimeCallTargetFri, 07 Dec 2012 16:01:38 +0100, by Doug Simon
-
added support for writing stubs in Java and wrote the TLAB fast refill stubFri, 07 Dec 2012 15:12:05 +0100, by Doug Simon
-
tightened first parameter type for LIRGeneratorTool.emitCall()Fri, 07 Dec 2012 11:14:12 +0100, by Doug Simon
-
made snippet template specialization support ConstantParameters of type WordFri, 07 Dec 2012 11:12:26 +0100, by Doug Simon
-
made word type rewriting do stamp inference so that word typed phis are handled correctlyFri, 07 Dec 2012 11:09:17 +0100, by Doug Simon
-
modified code installation to handle debug info that only contains oop maps but no frame informationFri, 07 Dec 2012 11:07:36 +0100, by Doug Simon
-
created shared, cached exceptions with empty stack traces for the implementation of the -G:+OmitHotExceptionStacktrace option. This also avoids the issue of having an object embedded in compiled code without an external strong reference to the same object (objects in compiled code are weak references in HotSpot)Thu, 06 Dec 2012 14:01:32 +0100, by Doug Simon
-
unify object reallocation code for c2 and GraalWed, 05 Dec 2012 13:57:00 +0100, by Lukas Stadler
-
sort the fields returned by getInstanceFields by offsetWed, 05 Dec 2012 13:56:56 +0100, by Lukas Stadler
-
Defer lowering to next iteration when the CFG was changed by a previous lowering so that no lastFixedNode is availableTue, 04 Dec 2012 11:05:01 -0800, by Christian Wimmer
-
Remove a usasge of getMirrorOrFailTue, 04 Dec 2012 11:02:56 -0800, by Christian Wimmer
-
Make method accessible from the outsideTue, 04 Dec 2012 10:53:19 -0800, by Christian Wimmer
-
Add accessorTue, 04 Dec 2012 10:52:35 -0800, by Christian Wimmer
-
When a hub has Kind.Object, the hub is never null. Preserve that information it the stamp of the LoadHubNode.Tue, 04 Dec 2012 10:51:46 -0800, by Christian Wimmer
-
Allow VM-specific lowering of FixedGuardNodeTue, 04 Dec 2012 10:50:12 -0800, by Christian Wimmer
-
Provide Thread context class loader to getMirrorOrFail. This is a workaround until the Graal API is complete enough so that getMirrorOrFail is not necessary at all.Tue, 04 Dec 2012 10:49:18 -0800, by Christian Wimmer
-
Assertion in Java code to prevent crash in native codeTue, 04 Dec 2012 10:46:21 -0800, by Christian Wimmer
-
Snippets cannot have speculative optimizations since they have to be valid for the entire run of the VMTue, 04 Dec 2012 10:44:41 -0800, by Christian Wimmer
-
removed all Graal modifications to ci and c1Mon, 03 Dec 2012 17:54:05 +0100, by Doug Simon
-
decoupled C++ Graal runtime from C1Mon, 03 Dec 2012 15:32:17 +0100, by Doug Simon
-
Remove overloading of filter(Class<F>) in NodeIterable<Node> returned by Graph.getNodes() : getNodes().filter(MergeNode.class) is not the same as getNodes(MergeNode.class)Mon, 03 Dec 2012 13:56:13 +0100, by Gilles Duboscq
-
graal.graph refactoringsMon, 03 Dec 2012 13:53:53 +0100, by Gilles Duboscq
-
make jtt test cases staticFri, 30 Nov 2012 11:39:05 -0800, by Christian Wimmer
-
LIR generation for BreakpointNode is not HotSpot specific and generally useful, so it should not be in HotSpot projectFri, 30 Nov 2012 11:22:50 -0800, by Christian Wimmer
-
make jtt test cases staticFri, 30 Nov 2012 11:21:54 -0800, by Christian Wimmer
-
MergeFri, 30 Nov 2012 13:46:37 +0100, by Gilles Duboscq
-
Update scala-benchmark dependencyThu, 15 Nov 2012 16:42:39 +0100, by Gilles Duboscq
-
MergeThu, 15 Nov 2012 14:21:50 +0100, by Gilles Duboscq
-
Factor out the precondition checking code in ArrayCopySnippetsTue, 13 Nov 2012 14:41:19 +0100, by Gilles Duboscq
-
simplified TLAB allocation (don't use + and -)Fri, 30 Nov 2012 12:08:38 +0100, by Lukas Stadler
-
MergeFri, 30 Nov 2012 11:13:36 +0100, by Gilles Duboscq
-
Do not print code installation failures from Java.Fri, 30 Nov 2012 11:11:13 +0100, by Gilles Duboscq
-
changed System.err output to TTY to avoid that DaCapo benchmarks fail digest validationFri, 30 Nov 2012 10:40:07 +0100, by Christian Haeubl
-
Merge.Fri, 30 Nov 2012 08:30:22 +0100, by Christian Haeubl
-
minor fixThu, 29 Nov 2012 17:19:13 +0100, by Christian Haeubl
-
Merge.Thu, 29 Nov 2012 16:20:42 +0100, by Christian Haeubl
-
fixed some issues that precluded compilation of Graal compiler methodsThu, 29 Nov 2012 15:44:49 +0100, by Christian Haeubl
-
print a message when optimistic optimizations are disabledThu, 29 Nov 2012 11:00:32 +0100, by Christian Haeubl
-
MergeThu, 29 Nov 2012 17:43:48 -0800, by Christian Wimmer
-
Make constructors of Constant private to ensure proper encapuslation of the type-overloaded primitive field; reduce to one constructor that sets all fields.Thu, 29 Nov 2012 17:43:09 -0800, by Christian Wimmer
-
Remove the ResolvedJavaType.Representation for static fields. It is not necessary to emit the constant holder for static fields in the GraphBuilder. Only the VM-specific lowering needs to create the constants, and it can access the VM-specific metadata objects directly.Thu, 29 Nov 2012 16:53:44 -0800, by Christian Wimmer
-
Eliminate bogus conversion that Constant.asLong() performed; simplify asXxx methods so that they do not have control flow.Thu, 29 Nov 2012 15:05:39 -0800, by Christian Wimmer
-
Merge.Thu, 29 Nov 2012 22:53:21 +0100, by Doug Simon
-
better encoding of dependencies during dependency recording by GraalThu, 29 Nov 2012 22:53:03 +0100, by Doug Simon
-
removed Graal specific initialization of the ciObjectFactoryThu, 29 Nov 2012 17:57:03 +0100, by Doug Simon
-
Cleanup of Kind class: remove use of Unsafe classThu, 29 Nov 2012 12:17:02 -0800, by Christian Wimmer
-
Cleanup of Kind class: remove isXxx methodsThu, 29 Nov 2012 11:27:23 -0800, by Christian Wimmer
-
Replace ResolvedJavaType.isAssignableTo with isAssignableFrom to be consistent with java.lang.ClassThu, 29 Nov 2012 10:10:03 -0800, by Christian Wimmer
-
Remove ResolvedJavaType.isClass()Thu, 29 Nov 2012 09:13:59 -0800, by Christian Wimmer
-
Merge.Thu, 29 Nov 2012 17:13:59 +0100, by Doug Simon
-
removed remaining use of the C++ "compiler interface" (i.e. ci) by Graal for installing dependenciesThu, 29 Nov 2012 17:13:13 +0100, by Doug Simon
-
removed unnecessary and incorrect initialization of ciEnv._dependenciesThu, 29 Nov 2012 10:18:32 +0100, by Doug Simon
-
For code installation failure, start the counter so that we don't get an error message if an installation fails shortly after the VM has startedThu, 29 Nov 2012 15:13:19 +0100, by Gilles Duboscq
-
Canoncialize UnsafeLoad/StoreThu, 29 Nov 2012 13:31:48 +0100, by Gilles Duboscq
-
Common out some parts of UnsafeLoad/Store in UnsafeAccessThu, 29 Nov 2012 13:24:08 +0100, by Gilles Duboscq
-
added metering of code installation failure rate to detect excessive failure caused by overly optimistic assumptionsWed, 28 Nov 2012 20:39:43 +0100, by Doug Simon
-
guard against nmethod allocation failure in implementation of PrintMachineCodeToFile optionWed, 28 Nov 2012 19:18:39 +0100, by Doug Simon
-
made HotSpotResolvedObjectType.findUniqueConcreteSubtype() more conservative for array types to fix issue with frequent invalidation of Graal compiled methodsWed, 28 Nov 2012 18:48:17 +0100, by Doug Simon
-
added toString() implementation to Assumption classesWed, 28 Nov 2012 18:45:54 +0100, by Doug Simon
-
MergeWed, 28 Nov 2012 14:28:53 +0100, by Gilles Duboscq
-
HotSpotResolvedJavaType is now the HotSpot implementation of ResolvedJavaType,Wed, 28 Nov 2012 14:07:25 +0100, by Gilles Duboscq
-
Remove unsafe mirror() method from HotSpotTypeUnresolved. Use a HotSpotMirrorHolder interface for HotSpotResolvedJavaType and HotSpotTypePrimitive which always hold valid mirrors.Wed, 28 Nov 2012 12:18:34 +0100, by Gilles Duboscq
-
readded IntrinsifyArrayCopyPhase that was removed accidentlyWed, 28 Nov 2012 14:01:17 +0100, by Christian Haeubl
-
Merge.Wed, 28 Nov 2012 09:01:39 +0100, by Christian Haeubl
-
disabled control-flow-based inliningWed, 28 Nov 2012 09:01:15 +0100, by Christian Haeubl
-
re-applied changeset d9930fec7553Tue, 27 Nov 2012 13:14:30 +0100, by Christian Haeubl
-
Merge.Tue, 27 Nov 2012 13:02:43 +0100, by Christian Haeubl
-
Merge.Tue, 27 Nov 2012 12:16:47 +0100, by Christian Haeubl
-
fixes after mergeTue, 27 Nov 2012 12:15:18 +0100, by Christian Haeubl
-
Merge.Tue, 27 Nov 2012 12:12:02 +0100, by Christian Haeubl
-
added some more assertions to ensure that metadata constant are not emitted inlineTue, 27 Nov 2012 10:49:12 +0100, by Christian Haeubl
-
bugfix to prevent metadata constants from being emitted inlineTue, 27 Nov 2012 10:41:42 +0100, by Christian Haeubl
-
fix to avoid emitting a range switch for non-integer keysMon, 26 Nov 2012 14:32:48 +0100, by Christian Haeubl
-
fixes after mergeMon, 26 Nov 2012 11:02:50 +0100, by Christian Haeubl
-
Merge.Mon, 26 Nov 2012 10:57:02 +0100, by Christian Haeubl
-
minor changesMon, 26 Nov 2012 10:51:39 +0100, by Christian Haeubl
-
removed high-level interpreter code from HotSpotFri, 23 Nov 2012 14:08:03 +0100, by Christian Haeubl
-
fixes after mergeFri, 23 Nov 2012 12:06:31 +0100, by Christian Haeubl
-
Merge.Fri, 23 Nov 2012 11:50:27 +0100, by Christian Haeubl
-
fixed logging of inlining decisionsFri, 23 Nov 2012 11:40:17 +0100, by Christian Haeubl
-
first part of refactoring the InliningPhaseThu, 22 Nov 2012 13:30:14 +0100, by Christian Haeubl
-
merged inlining and intrinsification phasesThu, 15 Nov 2012 15:10:41 +0100, by Christian Haeubl
-
increased CompileThreshold to increase reliability of profiling informationThu, 15 Nov 2012 11:40:50 +0100, by Christian Haeubl
-
replaced MetaUtil.isJavaLangObject() with type.isClass(Object.class)Tue, 27 Nov 2012 22:28:19 +0100, by Doug Simon
-
rename: ResolvedJavaType.isArrayClass() -> ResolvedJavaType.isArray()Tue, 27 Nov 2012 22:26:18 +0100, by Doug Simon
-
removed ResolvedJavaType.toJava() and introduced ResolvedJavaType.isPrimitive()Tue, 27 Nov 2012 22:24:39 +0100, by Doug Simon
-
Merge.Tue, 27 Nov 2012 20:36:46 +0100, by Doug Simon
-
rename: findFieldWithOffset -> findInstanceFieldWithOffsetTue, 27 Nov 2012 20:34:15 +0100, by Doug Simon
-
added test for ResolvedJavaType.findFieldWithOffsetTue, 27 Nov 2012 20:32:52 +0100, by Doug Simon
-
Merge.Tue, 27 Nov 2012 18:33:13 +0100, by Doug Simon
-
removed usages of ResolvedJavaType.toJava() from bytecode interpreterTue, 27 Nov 2012 18:30:20 +0100, by Doug Simon
-
removed more usages of ResolvedJavaType.toJava()Tue, 27 Nov 2012 18:03:49 +0100, by Doug Simon
-
removed a usage of ResolvedJavaType.toJava()Tue, 27 Nov 2012 17:22:19 +0100, by Doug Simon
-
Make frame omission optionalTue, 27 Nov 2012 18:25:02 +0100, by Gilles Duboscq
-
MergeTue, 27 Nov 2012 17:33:52 +0100, by Lukas Stadler
-
virtualization of unsafeload and unsafestoreTue, 27 Nov 2012 17:32:40 +0100, by Lukas Stadler
-
remove createSignature from CompilerToVM interfaceTue, 27 Nov 2012 17:29:42 +0100, by Lukas Stadler
-
remove MonitorEnter interface from DimensionsNodeTue, 27 Nov 2012 13:44:16 +0100, by Lukas Stadler
-
VM support for deferred reads and writes: ScopeDesc, DebugInfo, DebugInfoRecorderTue, 27 Nov 2012 13:43:04 +0100, by Lukas Stadler
-
bug fix for HotSpotReslvedJavaType.findUniqueConcreteSubtype() on arrays of interfacesTue, 27 Nov 2012 16:37:02 +0100, by Doug Simon
-
added unit tests for JavaType, JavaMethod and ResolvedJavaMethodTue, 27 Nov 2012 16:09:05 +0100, by Doug Simon
-
improved documentation for ResolvedJavaType.findUniqueConcreteSubtype()Tue, 27 Nov 2012 15:10:50 +0100, by Doug Simon
-
MergeTue, 27 Nov 2012 13:03:08 +0100, by Gilles Duboscq
-
Avoid call to MetaUtil.format in GraphBuilder if it is not necessaryTue, 27 Nov 2012 11:29:06 +0100, by Gilles Duboscq
-
HotSpotSignature.getParameterType should not try eager resolving.Tue, 27 Nov 2012 11:08:02 +0100, by Gilles Duboscq
-
MergeTue, 27 Nov 2012 12:10:41 +0100, by Lukas Stadler
-
PEA: remove shouldRevirtualize logicTue, 27 Nov 2012 11:51:13 +0100, by Lukas Stadler
-
added unit tests for ResolvedJavaTypeTue, 27 Nov 2012 11:21:48 +0100, by Doug Simon
-
added more #ifdefs for GRAAL specific codeTue, 27 Nov 2012 11:18:34 +0100, by Doug Simon
-
Simplify how stamp is preserved after lowering of allocation nodesMon, 26 Nov 2012 19:30:54 -0800, by Christian Wimmer