Mercurial > hg > truffle
graph
-
add a new simple CheckCast elimination phaseFri, 11 May 2012 16:00:00 +0200, by Lukas Stadler
-
let StampFactory.orTypes use RiResolvedType.leastCommonAncestorFri, 11 May 2012 15:59:11 +0200, by Lukas Stadler
-
disable frame state cullingFri, 11 May 2012 15:26:46 +0200, by Lukas Stadler
-
use exactType and assumptions to canonicalize ReadHubNodeFri, 11 May 2012 15:17:53 +0200, by Lukas Stadler
-
use dependencies collection instead of explicit guard fieldsFri, 11 May 2012 15:17:17 +0200, by Lukas Stadler
-
alphabetically sort debug values for Metric and TimeFri, 11 May 2012 14:39:58 +0200, by Lukas Stadler
-
MergeThu, 10 May 2012 14:24:25 +0200, by Lukas Stadler
-
MergeWed, 09 May 2012 12:11:36 +0200, by Lukas Stadler
-
cull unnecessary framestates at mergesWed, 09 May 2012 12:10:35 +0200, by Lukas Stadler
-
add field name to AccessFieldNodes in visualizerWed, 09 May 2012 10:43:05 +0200, by Lukas Stadler
-
canonicalize more integer operations: (a+1)+2 to a+3 and (a*2)*2 to a*4Wed, 09 May 2012 10:42:03 +0200, by Lukas Stadler
-
made graph marking stateless (in the graph) to ensure recursive marking is safeFri, 11 May 2012 14:11:56 +0200, by Doug Simon
-
moved checkcast lowering into LoweringPhase and added -G:HIRLowerCheckcast option to enable it (disabled by default) as it is not yet stableFri, 11 May 2012 12:20:36 +0200, by Doug Simon
-
slightly better implementation of NodeIterable.isEmpty()Fri, 11 May 2012 11:57:29 +0200, by Doug Simon
-
added javadoc to ValueProxyNodeFri, 11 May 2012 11:56:58 +0200, by Doug Simon
-
synchronized mxtool with MaxineFri, 11 May 2012 11:27:53 +0200, by Doug Simon
-
removed stateAfter field from FixedWithNextNode - now only StateSplit implementations have this fieldFri, 11 May 2012 11:26:39 +0200, by Doug Simon
-
rename: megamorph -> megamorphicFri, 11 May 2012 10:57:50 +0200, by Doug Simon
-
used more specific BeginNode subclass where appropriateFri, 11 May 2012 10:13:55 +0200, by Doug Simon
-
made CheckCastNode be a FixedNode instead of a BooleanNodeThu, 10 May 2012 00:36:12 +0200, by Doug Simon
-
created StartNode and DispatchBeginNode for clearer visualization of these special nodes in the visualizerThu, 10 May 2012 00:33:58 +0200, by Doug Simon
-
removed TypeCheckNodeWed, 09 May 2012 22:35:44 +0200, by Doug Simon
-
moved conversion of type profiles into hints for type check instructions from front end to lowering phaseWed, 09 May 2012 22:21:58 +0200, by Doug Simon
-
added javadoc to GuardNodeWed, 09 May 2012 16:39:52 +0200, by Doug Simon
-
if -G:Dump option is used, the compiler queues are given 5 seconds to finish any pending tasks (including graph dumping) before the VM exitsWed, 09 May 2012 16:21:03 +0200, by Doug Simon
-
build JaCoCo exclude list by scanning for snippetsTue, 08 May 2012 22:51:36 +0200, by Doug Simon
-
lowering checkcasts with Java snippetsTue, 08 May 2012 21:32:35 +0200, by Doug Simon
-
separated the notion of has-side-effect from may-have-frame-stateTue, 08 May 2012 20:17:30 +0200, by Doug Simon
-
expanded formatting of array CiConstants to show contents of the arrayTue, 08 May 2012 16:10:00 +0200, by Doug Simon
-
pulled AbstractStateSplit into FixedWithNextNode and made it no longer implement StateSplit. StateSplit is now implemented only by nodes that really have side effects.Tue, 08 May 2012 15:52:01 +0200, by Doug Simon
-
lowering checkcasts with Java snippets (incomplete)Mon, 07 May 2012 11:40:44 +0200, by Doug Simon
-
moved creation of a RiCodeInfo to the runtime side of the CRIMon, 07 May 2012 11:34:16 +0200, by Doug Simon
-
added thread name to debug info messageMon, 07 May 2012 11:08:58 +0200, by Doug Simon
-
canonicalize load from a compile-constant, immutable arrayFri, 04 May 2012 16:54:27 +0200, by Doug Simon
-
extended canonicalization so that a caller can specify a CiConstant value is an immutable object or arrayFri, 04 May 2012 16:52:58 +0200, by Doug Simon
-
removed product and debug configurations from hotspot CDT project configFri, 04 May 2012 16:38:38 +0200, by Doug Simon
-
Add verification in NullCheckNode and AccessFieldNodeFri, 04 May 2012 12:08:05 +0200, by Gilles Duboscq
-
Fix : in peeling, while merging early exits, transfer all anchored nodes from the exit to the mergeFri, 04 May 2012 12:08:00 +0200, by Gilles Duboscq
-
Fix : There can be proxies in the virtual chainThu, 03 May 2012 15:43:03 +0200, by Gilles Duboscq
-
MergeThu, 03 May 2012 15:19:16 +0200, by Gilles Duboscq
-
Fix mergeThu, 03 May 2012 15:19:11 +0200, by Gilles Duboscq
-
MergeThu, 03 May 2012 11:05:46 +0200, by Gilles Duboscq
-
rename CiKind.toUnboxedJavaClass to toBoxedJavaClassFri, 27 Apr 2012 13:13:06 +0200, by Gilles Duboscq
-
MergeFri, 27 Apr 2012 12:14:47 +0200, by Gilles Duboscq
-
Fix a peeling-escape analysis related problem that was causin LinearScan bailouts (merge virtual chains when merging ends of the peeled code)Fri, 27 Apr 2012 12:14:38 +0200, by Gilles Duboscq
-
added mechanism (enabled by -G:PICache and -G:PiFilter) for saving/loading method profiling info to/from diskThu, 03 May 2012 13:39:45 +0200, by Doug Simon
-
fixed bug in use of CheckcastMinHintHitProbabilityWed, 02 May 2012 18:23:12 +0200, by Doug Simon
-
moved RiResolvedMethod.dumpProfile() to CiUtil.profileAsString()Wed, 02 May 2012 18:04:36 +0200, by Doug Simon
-
added GraalOptions.CheckcastMinHintHitProbability to better guide use of hints for checkcastsWed, 02 May 2012 17:09:00 +0200, by Doug Simon
-
test case for the InstanceOfNode canonicalizer fixWed, 02 May 2012 16:32:00 +0200, by Lukas Stadler
-
MergeWed, 02 May 2012 15:08:41 +0200, by Lukas Stadler
-
changes to canonicalizer log outputWed, 02 May 2012 14:56:07 +0200, by Lukas Stadler
-
use PiNodes instead of CheckCastNodes to pin inlining receivers, remove emitCode flagWed, 02 May 2012 14:55:51 +0200, by Lukas Stadler
-
fix canonicalization of InstanceOfNodeWed, 02 May 2012 14:53:20 +0200, by Lukas Stadler
-
change PiNode to anchor on any FixedNode, not only BeginNodesWed, 02 May 2012 14:47:04 +0200, by Lukas Stadler
-
re-enable graph caching and the priority compilation queueWed, 02 May 2012 14:45:56 +0200, by Lukas Stadler