Mercurial > hg > graal-compiler
graph
-
Add graal-specific option -XX:+GPUOffload to trigger all GPU offload functionality. Cleanup function to decide GPU offload. Exclude GPU sources from nongraal builds.Wed, 22 Jan 2014 16:06:47 -0500, by S.Bharadwaj Yadavalli
-
fix invalid MemoryPhiNodes at loop beginsWed, 22 Jan 2014 15:27:31 +0100, by Lukas Stadler
-
cyclic graph verificationWed, 22 Jan 2014 15:27:31 +0100, by Lukas Stadler
-
fix GuardNode canonicalizationWed, 22 Jan 2014 14:03:47 +0100, by Lukas Stadler
-
added library sources (if available) to generated Eclipse attach launcherWed, 22 Jan 2014 13:39:45 +0100, by Doug Simon
-
Support narrow oops in reference maps.Wed, 22 Jan 2014 11:44:27 +0100, by Roland Schatz
-
Refactor reference maps.Wed, 22 Jan 2014 11:27:28 +0100, by Roland Schatz
-
split a negated guard with a ShortCircuitOr condition into two guardsWed, 22 Jan 2014 10:51:32 +0100, by Lukas Stadler
-
GuardLowering: do not use trapping accesses for guards with a speculationTue, 21 Jan 2014 18:35:46 +0100, by Gilles Duboscq
-
Sweeper: make sure we only run the sweeper on java threadsTue, 21 Jan 2014 18:35:03 +0100, by Gilles Duboscq
-
added list to PTXHotSpotBackend for keeping a valid installed kernel (and its associated nmethods) aliveTue, 21 Jan 2014 15:26:38 +0100, by Doug Simon
-
Truffle: fix non-local control flow using exceptionsTue, 21 Jan 2014 12:14:38 +0100, by Andreas Woess
-
IGV: throw graph reference away when GraphNode component is closedTue, 21 Jan 2014 00:20:33 +0200, by Bernhard Urban
-
IGV: fix nulling out reference in rootnodeMon, 20 Jan 2014 20:17:05 +0100, by Bernhard Urban
-
Merge.Mon, 20 Jan 2014 22:30:19 +0100, by Doug Simon
-
In a Graal-enabled VM, run the Sweeper even if the current thread is not a compiler thread.Mon, 20 Jan 2014 19:50:17 +0100, by Gilles Duboscq
-
Avoid segfault in InstanceKlass::oop_print_on's String special case when the value array is null.Mon, 20 Jan 2014 19:49:47 +0100, by Gilles Duboscq
-
Deoptimization: only process the pendign speculation for Graal-compiled methodsMon, 20 Jan 2014 19:48:53 +0100, by Gilles Duboscq
-
moved @Ignore to the correct method; try better to force a GC in testMon, 20 Jan 2014 19:12:01 +0100, by Doug Simon
-
Merge.Mon, 20 Jan 2014 18:25:17 +0100, by Christian Humer
-
Truffle: fixed failing FrameTest.Mon, 20 Jan 2014 16:33:40 +0100, by Christian Humer
-
Truffle-DSL: added default buffering to all java source code that is written. JavaFileObject#openWriter may return a non-buffered writer.Mon, 20 Jan 2014 15:23:10 +0100, by Christian Humer
-
SL: another minor FrameDescriptor fix.Mon, 20 Jan 2014 15:18:39 +0100, by Christian Humer
-
Merge.Mon, 20 Jan 2014 14:22:56 +0100, by Christian Humer
-
SL: required fixes for moving FrameDescriptor to the RootNode.Fri, 17 Jan 2014 17:10:46 +0100, by Christian Humer
-
Ruby: required fixes for moving FrameDescriptor to the RootNode.Mon, 20 Jan 2014 13:44:54 +0100, by Christian Humer
-
Truffle API-change: FrameDescriptors are now stored in the RootNode in a final field instead of the CallTarget.Fri, 17 Jan 2014 17:06:08 +0100, by Christian Humer
-
Merge.Fri, 17 Jan 2014 16:55:59 +0100, by Christian Humer
-
Truffle: extend inlining interface with custom reprofile counters.Thu, 16 Jan 2014 15:20:17 +0100, by Christian Humer
-
IGV: remove listener when closing the top componentMon, 20 Jan 2014 18:22:02 +0200, by Bernhard Urban
-
IGV: avoid potential memory leakMon, 20 Jan 2014 16:09:11 +0100, by Bernhard Urban
-
IGV: explicitly avoid serialization of top componentMon, 20 Jan 2014 16:09:10 +0100, by Bernhard Urban
-
PTX wrapper maintains a reference to the HotSpotNmethod object for the installed GPU code so that code invalidation can be tested for in the wrapperMon, 20 Jan 2014 16:29:10 +0100, by Doug Simon
-
added check for invalidated code in executeCompiledMethodVarargsMon, 20 Jan 2014 16:27:24 +0100, by Doug Simon
-
changed return type of HotSpotCodeCacheProvider.addExternalMethod to be HotSpotNmethodMon, 20 Jan 2014 16:26:39 +0100, by Doug Simon
-
fixed warningMon, 20 Jan 2014 16:25:14 +0100, by Doug Simon
-
refined class path used by unittest to only include dependencies of the tests that will be runMon, 20 Jan 2014 16:25:04 +0100, by Doug Simon
-
adapt tests for new IfNode and MergeNode simplificationsFri, 17 Jan 2014 11:19:50 +0100, by Lukas Stadler
-
additional simplifications for IfNode and MergeNodeFri, 17 Jan 2014 09:38:39 +0100, by Lukas Stadler
-
allow multiple ReturnNodes per graphWed, 15 Jan 2014 16:11:56 +0100, by Lukas Stadler
-
expose memory map merging functionality in FloatingReadPhaseWed, 15 Jan 2014 16:08:57 +0100, by Lukas Stadler
-
add IterableNodeType to ReturnNodeWed, 15 Jan 2014 16:07:01 +0100, by Lukas Stadler
-
fix Class.isInstance substitutionFri, 17 Jan 2014 09:36:26 +0100, by Lukas Stadler
-
remove LoadFieldNodes without usagesWed, 15 Jan 2014 17:00:40 +0100, by Lukas Stadler
-
change label of ParameterNodeWed, 15 Jan 2014 16:03:07 +0100, by Lukas Stadler
-
fix compile error in method.cppMon, 20 Jan 2014 09:21:39 +0100, by Lukas Stadler
-
Allow TruffleOptions to be set using system properties.Sun, 19 Jan 2014 18:18:42 +0000, by Chris Seaton
-
Use .equals instead of == to compare ConstantFri, 17 Jan 2014 23:16:59 +0100, by Gilles Duboscq
-
schedule lambda method compilation and execution on GPU (PTX) when possible; fix a couple of bugs.Fri, 17 Jan 2014 16:03:13 -0500, by S.Bharadwaj Yadavalli
-
fixed problem in netbeansinit that generated invalid annotation processor paths for projectsFri, 17 Jan 2014 17:26:47 +0100, by Doug Simon
-
Use a more meaningfull Compilationresult/HotSpotNMethod name in the TruffleCompilerImplFri, 17 Jan 2014 16:36:51 +0100, by Gilles Duboscq
-
Rename OptimizeGuardAnchors to OptimizeGuardAnchorsPhaseFri, 17 Jan 2014 16:23:18 +0100, by Gilles Duboscq
-
Make optimal anchor point available ouside OptimizeGuardAnchors phaseFri, 17 Jan 2014 16:22:26 +0100, by Gilles Duboscq
-
Keep speculation and debug id when canonicalizing DynamicDeoptimizeNodeFri, 17 Jan 2014 16:13:05 +0100, by Gilles Duboscq
-
Rename more 'speculation id' to 'debug id'. Add decodeDebugId to MetaAccessProviderFri, 17 Jan 2014 16:13:02 +0100, by Gilles Duboscq
-
Tightly pack data references that are smaller than 8 bytes.Fri, 17 Jan 2014 15:46:30 +0100, by Roland Schatz