IRC.git
14 years agochanges to analysis..
bdemsky [Tue, 16 Jun 2009 22:08:00 +0000 (22:08 +0000)]
changes to analysis..

14 years agomore work on mlp system
jjenista [Tue, 16 Jun 2009 21:47:49 +0000 (21:47 +0000)]
more work on mlp system

14 years agoFixes for bugs in regard to null analysis results for nodes outside of the root SESE...
jjenista [Tue, 16 Jun 2009 21:46:56 +0000 (21:46 +0000)]
Fixes for bugs in regard to null analysis results for nodes outside of the root SESE scope

14 years agoforgot a file
adash [Tue, 16 Jun 2009 00:13:08 +0000 (00:13 +0000)]
forgot a file

14 years agoclean up code
adash [Tue, 16 Jun 2009 00:10:02 +0000 (00:10 +0000)]
clean up code
bug fix for sorting

14 years agochanges
bdemsky [Mon, 15 Jun 2009 23:49:31 +0000 (23:49 +0000)]
changes

14 years agochanges:
bdemsky [Mon, 15 Jun 2009 23:09:28 +0000 (23:09 +0000)]
changes:
(1) Add support for synchronized blocks
(2) Fix analysis bug
(3) Start support for delaying operations until commit

14 years agoMake SESE root spliced into IR graph just as any other SESE, make sure MLP analysis...
jjenista [Mon, 15 Jun 2009 22:51:52 +0000 (22:51 +0000)]
Make SESE root spliced into IR graph just as any other SESE, make sure MLP analysis simply ignores the few IR nodes outside of root SESE scope.  Problem: this implementation doesn't transform all return nodes into a single exit point yet

14 years agoadd readme file
adash [Mon, 15 Jun 2009 21:46:54 +0000 (21:46 +0000)]
add readme file

14 years agofiles modified for Bayes benchmark
adash [Mon, 15 Jun 2009 21:08:55 +0000 (21:08 +0000)]
files modified for Bayes benchmark

14 years agoforgot a file
adash [Mon, 15 Jun 2009 21:05:27 +0000 (21:05 +0000)]
forgot a file

14 years ago add new benchmark : Bayes
adash [Mon, 15 Jun 2009 20:59:12 +0000 (20:59 +0000)]
 add new benchmark : Bayes
 Will clean up files later, at least have a good copy that compiles

14 years agowork stealing schedule system cleanly runs highly parallel workload at appropriate...
jjenista [Fri, 12 Jun 2009 22:39:04 +0000 (22:39 +0000)]
work stealing schedule system cleanly runs highly parallel workload at appropriate load factor

14 years agonew changes to common library files and bug fixes to Genome(still uses hashtable...
adash [Fri, 12 Jun 2009 01:17:44 +0000 (01:17 +0000)]
new changes to common library files and bug fixes to Genome(still uses hashtable and Strings)

14 years agobug fix for comparisons
bdemsky [Fri, 12 Jun 2009 00:41:05 +0000 (00:41 +0000)]
bug fix for comparisons

14 years agoNeed a double-ended queue for work-stealing algorithm, extended Runtime//Queue to...
jjenista [Thu, 11 Jun 2009 18:13:20 +0000 (18:13 +0000)]
Need a double-ended queue for work-stealing algorithm, extended Runtime//Queue to have the functionality.  Found some bugs in Queue with pointers that were never exposed when using Queue as singly-linked list, added fixes and my test files

14 years agomore mlp code, needs dependency counts
jjenista [Wed, 10 Jun 2009 16:55:20 +0000 (16:55 +0000)]
more mlp code, needs dependency counts

14 years agoAdded ByteArray implementation.
afedward [Wed, 10 Jun 2009 01:56:47 +0000 (01:56 +0000)]
Added ByteArray implementation.

14 years agomore updates towards implementing SESE runtime
jjenista [Tue, 9 Jun 2009 22:01:38 +0000 (22:01 +0000)]
more updates towards implementing SESE runtime

14 years agocreate new Genome directory and add files
adash [Tue, 9 Jun 2009 01:31:41 +0000 (01:31 +0000)]
create new Genome directory and add files
clean up the comments
change C longs -> ints
works fine for singleTM version with all compiler optimizations

14 years agobug fix for alokika
bdemsky [Mon, 8 Jun 2009 21:38:09 +0000 (21:38 +0000)]
bug fix for alokika

14 years agoadd makefile
adash [Wed, 3 Jun 2009 21:11:23 +0000 (21:11 +0000)]
add makefile

14 years agobug fix
bdemsky [Wed, 3 Jun 2009 20:19:36 +0000 (20:19 +0000)]
bug fix

14 years ago add common Library files
adash [Wed, 3 Jun 2009 01:24:48 +0000 (01:24 +0000)]
 add common Library files

14 years agoForgot the actual hashtable.java.
afedward [Tue, 2 Jun 2009 21:20:23 +0000 (21:20 +0000)]
Forgot the actual hashtable.java.

14 years agoAdding Hashtable implementation.
afedward [Tue, 2 Jun 2009 21:19:57 +0000 (21:19 +0000)]
Adding Hashtable implementation.

14 years ago*** empty log message ***
navid [Tue, 2 Jun 2009 21:13:45 +0000 (21:13 +0000)]
*** empty log message ***

14 years ago*** empty log message ***
navid [Tue, 2 Jun 2009 21:13:38 +0000 (21:13 +0000)]
*** empty log message ***

14 years ago*** empty log message ***
navid [Tue, 2 Jun 2009 21:13:33 +0000 (21:13 +0000)]
*** empty log message ***

14 years ago*** empty log message ***
navid [Tue, 2 Jun 2009 21:13:26 +0000 (21:13 +0000)]
*** empty log message ***

14 years agofix random class
bdemsky [Tue, 2 Jun 2009 07:48:18 +0000 (07:48 +0000)]
fix random class

14 years agosupport to print out size of garbage
bdemsky [Tue, 2 Jun 2009 01:55:20 +0000 (01:55 +0000)]
support to print out size of garbage

14 years agobug fix for barriers...should be able to use joptimize with barriers now
bdemsky [Tue, 2 Jun 2009 01:38:52 +0000 (01:38 +0000)]
bug fix for barriers...should be able to use joptimize with barriers now

14 years agoModified to allow byte[] casting
afedward [Mon, 1 Jun 2009 23:59:37 +0000 (23:59 +0000)]
Modified to allow byte[] casting

14 years agoabout to make lots of changes to system, just committing this stable compilation...
jjenista [Mon, 1 Jun 2009 23:09:53 +0000 (23:09 +0000)]
about to make lots of changes to system, just committing this stable compilation that is benign to other compiler uses to reduce conflicts

14 years agoworking version of SSCA2,
adash [Sat, 30 May 2009 01:39:13 +0000 (01:39 +0000)]
working version of SSCA2,
gives out of memory errors for scale s > 17
does not use the joptimize flag for optimization

14 years agoadd math operations for multicore version
jzhou [Fri, 29 May 2009 01:15:25 +0000 (01:15 +0000)]
add math operations for multicore version

14 years agoMaking necessary changes to have latest working version of benchmark, was missing...
afedward [Wed, 27 May 2009 01:37:36 +0000 (01:37 +0000)]
Making necessary changes to have latest working version of benchmark, was missing some files.

14 years agoStable compile, seg fault in MLP mode.
jjenista [Tue, 26 May 2009 20:55:08 +0000 (20:55 +0000)]
Stable compile, seg fault in MLP mode.

14 years agoStable, not functional. Builds a switch case for unique SESE class invocation, inclu...
jjenista [Tue, 26 May 2009 18:23:53 +0000 (18:23 +0000)]
Stable, not functional.  Builds a switch case for unique SESE class invocation, including whether it has the root SESE for a parent or not

14 years agoadd support for 56 cores in multicore runtime
jzhou [Tue, 26 May 2009 17:12:59 +0000 (17:12 +0000)]
add support for 56 cores in multicore runtime

14 years agoAdd new file and debugging SSCA bug
adash [Tue, 26 May 2009 16:29:02 +0000 (16:29 +0000)]
Add new file and debugging SSCA bug

14 years agoNot functional, but stable--sese invocation code pulls from an SESErecord
jjenista [Sun, 24 May 2009 16:35:04 +0000 (16:35 +0000)]
Not functional, but stable--sese invocation code pulls from an SESErecord

14 years agoWorking version of SSCA2 benchmark for kernel 1
adash [Thu, 21 May 2009 23:43:44 +0000 (23:43 +0000)]
Working version of SSCA2 benchmark for kernel 1
-singleTM flag problems for more than 1 thread still exists, some problem in TRANSREAD, Amem_cpy()

14 years agoRemoving ~ files
afedward [Tue, 19 May 2009 23:18:28 +0000 (23:18 +0000)]
Removing ~ files

14 years agoGenome benchmark works for less than enormous test cases.
afedward [Tue, 19 May 2009 23:14:24 +0000 (23:14 +0000)]
Genome benchmark works for less than enormous test cases.

Runs into memory allocation error for test cases of -n16777216. (see ReadMe for argument info)

14 years agoUpdates to support Tilera multicore version runtime
jzhou [Tue, 19 May 2009 22:47:31 +0000 (22:47 +0000)]
Updates to support Tilera multicore version runtime

14 years agoBug fixes for generating all execution paths for SESE methods up to and including...
jjenista [Tue, 19 May 2009 20:24:46 +0000 (20:24 +0000)]
Bug fixes for generating all execution paths for SESE methods up to and including an SESE exit node

14 years agochanges for multicore version runtime according to latest update of class library
jzhou [Sat, 16 May 2009 00:47:37 +0000 (00:47 +0000)]
changes for multicore version runtime according to latest update of class library

14 years agouse int and float data type for arrays/variables (Compatible with size used in C...
adash [Sat, 16 May 2009 00:39:38 +0000 (00:39 +0000)]
use int and float data type for arrays/variables (Compatible with size used in C version)

14 years agoupdates for multicore runtime to support Tilera
jzhou [Sat, 16 May 2009 00:33:57 +0000 (00:33 +0000)]
updates for multicore runtime to support Tilera

14 years agoChecking in changes. Genome still not functional, but getting close.
afedward [Fri, 15 May 2009 22:15:47 +0000 (22:15 +0000)]
Checking in changes.  Genome still not functional, but getting close.

SingleTM error is revealed in Sequencer.java.  Sequencer.run() function reveals transactional errors in atomic blocks, especially as it applies to lines 251-271.

14 years agochanges for alex
bdemsky [Fri, 15 May 2009 22:05:27 +0000 (22:05 +0000)]
changes for alex

14 years agosmall change for better messages
bdemsky [Fri, 15 May 2009 21:15:03 +0000 (21:15 +0000)]
small change for better messages

14 years agochanges
bdemsky [Fri, 15 May 2009 21:06:41 +0000 (21:06 +0000)]
changes

14 years agostable for non-sese projects, an sese can have partial code
jjenista [Fri, 15 May 2009 17:16:57 +0000 (17:16 +0000)]
stable for non-sese projects, an sese can have partial code

14 years agochange for debug mode
bdemsky [Fri, 15 May 2009 17:07:58 +0000 (17:07 +0000)]
change for debug mode

14 years agochanges to use floats
bdemsky [Fri, 15 May 2009 09:29:13 +0000 (09:29 +0000)]
changes to use floats

14 years agochange
bdemsky [Fri, 15 May 2009 06:04:44 +0000 (06:04 +0000)]
change

14 years agodouble -> float
bdemsky [Fri, 15 May 2009 05:55:36 +0000 (05:55 +0000)]
double -> float

14 years agowrapper classes
bdemsky [Fri, 15 May 2009 05:31:35 +0000 (05:31 +0000)]
wrapper classes

14 years agothis change isn't all that well tested...
bdemsky [Fri, 15 May 2009 05:31:06 +0000 (05:31 +0000)]
this change isn't all that well tested...
but it lets us declare intwrapper, doublewrapper types which work just like the primitives but are really objects behind the scenes

14 years agonew files
adash [Fri, 15 May 2009 00:37:00 +0000 (00:37 +0000)]
new files

14 years agouse sese pretty id and id to generate method names, param struct and local struct...
jjenista [Thu, 14 May 2009 22:34:21 +0000 (22:34 +0000)]
use sese pretty id and id to generate method names, param struct and local struct that is unique--keep a mapping of seses to the bogus flat method for retrieval in order to reproduce those names at call sites.  Also some bug fixes

14 years agoLots of changes that aren't working yet, so new methods are commented and existing...
jjenista [Thu, 14 May 2009 21:37:42 +0000 (21:37 +0000)]
Lots of changes that aren't working yet, so new methods are commented and existing code altered for new stuff is checked in.  Won't affect other compiler modes, and checks in stuff to reduce conflicts while getting the sese's working.

14 years agochanges
bdemsky [Thu, 14 May 2009 21:36:53 +0000 (21:36 +0000)]
changes

14 years agomore efficiency hacks...we're looking fields for values we already know, and doing...
bdemsky [Thu, 14 May 2009 20:18:07 +0000 (20:18 +0000)]
more efficiency hacks...we're looking fields for values we already know, and doing arithmetic just to undo it..

14 years agomore changes...fix soft abort case...it needs to let other transactions through...
bdemsky [Thu, 14 May 2009 02:53:52 +0000 (02:53 +0000)]
more changes...fix soft abort case...it needs to let other transactions through...

14 years agochange
bdemsky [Thu, 14 May 2009 02:28:02 +0000 (02:28 +0000)]
change

14 years ago Add new files
adash [Thu, 14 May 2009 01:35:52 +0000 (01:35 +0000)]
 Add new files

14 years agoslightly better locks for stm...
bdemsky [Thu, 14 May 2009 01:06:14 +0000 (01:06 +0000)]
slightly better locks for stm...

14 years agoUpdates to support multicore version runtime of Tilera. Also fix a bug in BuildCodeMu...
jzhou [Wed, 13 May 2009 23:14:44 +0000 (23:14 +0000)]
Updates to support multicore version runtime of Tilera. Also fix a bug in BuildCodeMulticore.java: it faulty omitted codes which repeatedly transfer objects to the same core. In current scheduling, to send 4 objs to 2 cores, it could be specified as an array [0,0,1,1] indicating the target cores for each object. Originaly the array was generated as [0,0,1,null] which caused missing of objects.

14 years agomake atomic work in normal java mode...it just uses coarse locks
bdemsky [Wed, 13 May 2009 08:05:01 +0000 (08:05 +0000)]
make atomic work in normal java mode...it just uses coarse locks

14 years agochanges
bdemsky [Wed, 13 May 2009 03:06:41 +0000 (03:06 +0000)]
changes

14 years agovarious bug fixes
bdemsky [Wed, 13 May 2009 02:13:05 +0000 (02:13 +0000)]
various bug fixes

14 years ago add new files
adash [Wed, 13 May 2009 01:50:19 +0000 (01:50 +0000)]
 add new files

14 years agochanges to compile
bdemsky [Wed, 13 May 2009 01:27:33 +0000 (01:27 +0000)]
changes to compile

14 years agotry to simplify garbage collector lock
bdemsky [Wed, 13 May 2009 01:22:50 +0000 (01:22 +0000)]
try to simplify garbage collector lock

14 years agochanges to build slightly more optimal code
bdemsky [Wed, 13 May 2009 00:11:58 +0000 (00:11 +0000)]
changes to build slightly more optimal code

14 years agochanges:
bdemsky [Tue, 12 May 2009 21:47:46 +0000 (21:47 +0000)]
changes:
1) bug fix
2) support for fast memcpy...contact me for objects

14 years agosmall changes
bdemsky [Tue, 12 May 2009 03:28:59 +0000 (03:28 +0000)]
small changes
plus bugfixes...one race condition on riskyflag vs lock fixed

14 years agobug fix: acquire lock before copying the object into cache
adash [Tue, 12 May 2009 00:41:54 +0000 (00:41 +0000)]
bug fix: acquire lock before copying the object into cache

14 years agobug fix: changes to make things safe for the GC
adash [Mon, 11 May 2009 23:09:36 +0000 (23:09 +0000)]
bug fix: changes to make things safe for the GC

14 years agoGood cut of variable analysis
jjenista [Mon, 11 May 2009 22:41:22 +0000 (22:41 +0000)]
Good cut of variable analysis

14 years agoWhen we stall for some specific SESE, make other variables that come from that SESE...
jjenista [Mon, 11 May 2009 18:04:32 +0000 (18:04 +0000)]
When we stall for some specific SESE, make other variables that come from that SESE available as well

14 years agoChange isAvailable to notAvailable, updated computations for that analysis, use those...
jjenista [Fri, 8 May 2009 22:38:35 +0000 (22:38 +0000)]
Change isAvailable to notAvailable, updated computations for that analysis, use those results to prune waitSet for stalls--looks good\!

14 years agocompute when to write dynamic address of variables
jjenista [Fri, 8 May 2009 17:56:23 +0000 (17:56 +0000)]
compute when to write dynamic address of variables

14 years agosmall changes
bdemsky [Fri, 8 May 2009 05:39:45 +0000 (05:39 +0000)]
small changes

14 years agoAdd new benchmark:....some prelimnary files
adash [Fri, 8 May 2009 01:27:04 +0000 (01:27 +0000)]
Add new benchmark:....some prelimnary files

14 years agoupdates on Math operations for multicore version
jzhou [Thu, 7 May 2009 21:48:55 +0000 (21:48 +0000)]
updates on Math operations for multicore version

14 years agosmall changes
bdemsky [Thu, 7 May 2009 21:38:37 +0000 (21:38 +0000)]
small changes

14 years agothrow out commented code
adash [Thu, 7 May 2009 19:35:10 +0000 (19:35 +0000)]
throw out commented code

14 years agorace condition fixed: race due to thread rec becoming NULL before accessing its field...
adash [Thu, 7 May 2009 19:26:22 +0000 (19:26 +0000)]
race condition fixed: race due to thread rec becoming NULL before accessing its field/element

14 years agocommented out all functions by mistake
jzhou [Thu, 7 May 2009 18:35:05 +0000 (18:35 +0000)]
commented out all functions by mistake

14 years agoCurious. Found a bug in liveness where liveIn set becomes empty erroneously. While...
jjenista [Thu, 7 May 2009 18:28:08 +0000 (18:28 +0000)]
Curious.  Found a bug in liveness where liveIn set becomes empty erroneously.  While tracing bug back the output started producing the correct liveIn set, but the code diff confirms to me that I haven't made a functional change to the system.  I've been trying to understand what happened, but for now I'll check in this working, stable capture and keep an eye on liveness results in case the bug is just latent now.

14 years agoupdates for multicore version runtime, temporarily disable math operations in multico...
jzhou [Thu, 7 May 2009 18:28:06 +0000 (18:28 +0000)]
updates for multicore version runtime, temporarily disable math operations in multicore version

14 years agoadd frame for another pass to compute whether variables are available, then use to...
jjenista [Wed, 6 May 2009 23:08:02 +0000 (23:08 +0000)]
add frame for another pass to compute whether variables are available, then use to do computation for stalls

14 years agolocking technique for risky objects enabled
adash [Wed, 6 May 2009 21:06:38 +0000 (21:06 +0000)]
locking technique for risky objects enabled

14 years agoAdd CodePlan object for injecting code before or after a FlatNode
jjenista [Wed, 6 May 2009 18:27:25 +0000 (18:27 +0000)]
Add CodePlan object for injecting code before or after a FlatNode

14 years agoSeveral bug fixes. Now that tokens have a mutable element, reference var sets, they...
jjenista [Tue, 5 May 2009 22:08:19 +0000 (22:08 +0000)]
Several bug fixes.  Now that tokens have a mutable element, reference var sets, they must be deep copied when merging table to avoid altering two tables with one operation.  A bug in removal by temp prevented sese-var mapping from being updated also.  A change in table implementation makes a difference between public and private ops like add and remove.  Any public op calls the private version and then asserts consistency.  Private versions can be called in more complex operatios and don't necessarily have to leave the table consistent until the bigger op is finished.  Much easier to find source of bugs this way.