| Wed, 19 Mar 2014 19:02:29 +0100 |
Josef Eisl |
Baseline emit return.
|
| Wed, 19 Mar 2014 17:07:22 +0100 |
Josef Eisl |
Baseline compiler emitPrologue.
|
| Wed, 19 Mar 2014 15:22:14 +0100 |
Josef Eisl |
merge
|
| Tue, 18 Mar 2014 09:56:20 +0000 |
Niclas Adlertz |
LIRFrameStateBuilder added
|
| Wed, 19 Mar 2014 15:14:37 +0100 |
Josef Eisl |
Create Baseline version of doBlock in LIRGenerator.
|
| Wed, 19 Mar 2014 15:13:13 +0100 |
Josef Eisl |
Track AbstractBlock in LIRGenerator.
|
| Wed, 19 Mar 2014 13:35:13 +0100 |
Josef Eisl |
Introduce LIRGeneratorCommon and NodeBasedLIRGenerator interfaces.
|
| Tue, 18 Mar 2014 14:48:39 +0100 |
Josef Eisl |
Introduce LIRGenResLowerable.
|
| Tue, 18 Mar 2014 13:49:11 +0100 |
Josef Eisl |
Decoupled LIRGenerator and LIRGenerationResult.
|
| Tue, 18 Mar 2014 10:28:27 +0100 |
Josef Eisl |
LIRGenerator: add getter for DebugInfoBuilder and CallingConvention.
|
| Mon, 17 Mar 2014 19:18:35 +0100 |
Josef Eisl |
LIR renamed setter and getter functions.
|
| Mon, 17 Mar 2014 19:11:27 +0100 |
Josef Eisl |
Introduce LIRGenerationResultBase.
|
| Mon, 17 Mar 2014 18:36:58 +0100 |
Josef Eisl |
Introduce LIRGenerationResult interface.
|
| Thu, 13 Mar 2014 13:28:09 +0100 |
Josef Eisl |
Restrict access to members of LIRGenerator.
|
| Wed, 12 Mar 2014 20:40:01 +0100 |
Josef Eisl |
Introduce AbstractControlFlowGraph.
|
| Thu, 27 Mar 2014 13:20:41 +0100 |
Roland Schatz |
Add assertions to ensure that the deoptimizing FrameState is unique in LIR generation.
|
| Tue, 25 Mar 2014 23:52:57 +0100 |
Doug Simon |
fixed findbugs issue
|
| Tue, 25 Mar 2014 18:53:34 +0100 |
Roland Schatz |
Introduce separate interfaces for deoptimizing nodes that deopt to a state before, during or after their execution.
|
| Thu, 20 Mar 2014 15:57:03 -0700 |
Tom Rodriguez |
initial support for memory arithmetic on x86
|
| Thu, 20 Mar 2014 11:48:39 +0100 |
Josef Eisl |
Rename ValueNode kind() to getKind().
|
| Tue, 11 Mar 2014 16:55:57 +0100 |
Thomas Wuerthinger |
Encapsulate LIR class fields.
|
| Tue, 11 Mar 2014 16:16:24 +0100 |
Thomas Wuerthinger |
Remove reference to graph in LIRGenerator.
|
| Wed, 05 Mar 2014 20:12:46 +0100 |
Josef Eisl |
Remove LIRGenerator from LinearScan.
|
| Fri, 21 Feb 2014 19:35:08 +0100 |
Roland Schatz |
Fix wrong kind in LIRGenerator.
|
| Fri, 21 Feb 2014 13:04:58 +0100 |
Roland Schatz |
Make conversion from Stamp to PlatformKind extensible by backend.
|
| Tue, 18 Feb 2014 21:55:26 -0800 |
twisti |
added Arrays.equals substitutions
|
| Wed, 12 Feb 2014 10:23:21 -0800 |
Christian Wimmer |
Allow disabling of redundant move elimination
|
| Fri, 07 Feb 2014 17:39:32 +0100 |
Gilles Duboscq |
Setup the OSR calling convention before calling compileGraph rather than patching it in the LIRGenerator
|
| Fri, 07 Feb 2014 15:20:59 +0100 |
Roland Schatz |
Use branch probability for emitting conditional jump.
|
| Fri, 07 Feb 2014 15:03:21 +0100 |
Roland Schatz |
Remove unused 'negated' arguments.
|
| Mon, 13 Jan 2014 13:45:25 +0100 |
Lukas Stadler |
Merge
|
| Mon, 13 Jan 2014 13:20:30 +0100 |
Lukas Stadler |
rename LocalNode to ParameterNode
|
| Mon, 13 Jan 2014 11:28:25 +0100 |
Erik Eckstein |
consider register priority in rematerialization decision
|
| Tue, 07 Jan 2014 14:59:18 -0800 |
twisti |
String.equals method substitution
|
| Mon, 06 Jan 2014 17:19:18 -0800 |
Tom Rodriguez |
Embed compressed constants when possible and use more efficient patterns for encoding
|
| Fri, 03 Jan 2014 11:46:35 -0800 |
Christian Wimmer |
Bugfix: node successors are unordered, so emitting a jump to the first successor hits a random target
|
| Wed, 18 Dec 2013 09:02:01 +0100 |
Erik Eckstein |
enable rematerialization of constants in LinearScan, including a bug fix
|
| Mon, 16 Dec 2013 13:23:33 +0100 |
Doug Simon |
moved loads of constants out of loops
|
| Mon, 16 Dec 2013 08:20:36 +0100 |
Erik Eckstein |
Merge
|
| Fri, 13 Dec 2013 16:40:41 +0100 |
Erik Eckstein |
Merge
|
| Fri, 13 Dec 2013 08:38:52 +0100 |
Erik Eckstein |
added rematerialization of constants in LinearScan, but still disabled
|
| Fri, 13 Dec 2013 20:16:44 +0100 |
Doug Simon |
added comment explaining insertion of nop during commoning of constant loads
|
| Fri, 13 Dec 2013 14:41:59 +0100 |
Doug Simon |
removed support for external nodes (GRAAL-508)
|
| Fri, 13 Dec 2013 14:10:30 +0100 |
Doug Simon |
made commoning of loading constants non-configurable (GRAAL-508)
|
| Fri, 13 Dec 2013 13:32:11 +0100 |
Doug Simon |
the load of a constant is commoned to the nearest block dominating all usages (GRAAL-508)
|
| Wed, 11 Dec 2013 20:49:47 +0100 |
Doug Simon |
removed unused import
|
| Wed, 11 Dec 2013 15:59:40 +0100 |
Lukas Stadler |
rework of switch generation: move code into platform independent SwitchStrategy, add boolean switch strategy
|
| Mon, 09 Dec 2013 15:24:27 +0100 |
Doug Simon |
introduced BlockEndOp interface and require that every LIR block is terminated by such an operation
|
| Wed, 06 Nov 2013 16:01:03 +0100 |
Lukas Stadler |
make LIRGenerator.emitSequentialSwitch protected
|
| Wed, 06 Nov 2013 11:04:01 +0100 |
Lukas Stadler |
LIRGenerator: don't create end-of-block JumpOps for blocks that end in a FallThroughOp with a target
|
| Tue, 05 Nov 2013 19:03:37 +0100 |
Doug Simon |
moved a number of options to be declared closer to usage and cached some heavily accessed options in fields
|
| Sun, 03 Nov 2013 15:27:52 +0100 |
Doug Simon |
made ConstantNodes external to a Graph (GRAAL-508)
|
| Thu, 24 Oct 2013 14:57:57 +0200 |
Doug Simon |
made ConstantNodes (optionally) not record their usages (GRAAL-508)
|
| Tue, 15 Oct 2013 14:14:06 +0200 |
Doug Simon |
split HotSpotRuntime into separate provider implementations
|
| Sat, 12 Oct 2013 01:03:47 +0200 |
Doug Simon |
moved lookupForeignCall() from CodeCacheProvider to ForeignCallsProvider (GRAAL-511)
|
| Thu, 10 Oct 2013 16:14:55 +0200 |
Doug Simon |
made CodeCacheProvider independent of MetaAccessProvider (GRAAL-511)
|
| Fri, 04 Oct 2013 00:53:19 +0200 |
Doug Simon |
added block ids to Labels
|
| Thu, 26 Sep 2013 16:03:28 +0200 |
Gilles Duboscq |
SpeculationLog gives back the necessary Constant when calling maySpeculate. Remove the deoptimizationReasonfrom the DebugInfo
|
| Wed, 25 Sep 2013 13:05:16 +0200 |
Gilles Duboscq |
StubForeignCallNode is not a DeoptimizingNode
|
| Thu, 05 Sep 2013 16:36:13 -0700 |
Christian Wimmer |
Fixed registers that flow into a block need to be inputs of the LabelOp, otherwise the register allocator can insert spill moves before the definition of a fixed register.
|
