IRC.git
13 years agofix: handle the case that TASK doesn't have any heap conflicts
yeom [Fri, 8 Apr 2011 17:17:06 +0000 (17:17 +0000)]
fix: handle the case that TASK doesn't have any heap conflicts

13 years agochanges
stephey [Fri, 8 Apr 2011 06:33:01 +0000 (06:33 +0000)]
changes

13 years agochanged number of RCR workers to matchooo workers. The new speed up is either 13x...
stephey [Fri, 8 Apr 2011 06:31:48 +0000 (06:31 +0000)]
changed number of RCR workers to matchooo workers. The new speed up is either 13x or 6x...... 13x is running the single version but for some odd reason, it garbage collects... the 6x comes from running the rcrpointer version with depth of 0 which SHOULD call no traverser and spawn no sese (other than the parent).

I think it could be faster still, but a bug in the conflict graph is preventing me from adding more seses to speed things up.

13 years agovoronoi benchmark. On dc-11, the computation part runs 3x faster with a computation...
stephey [Fri, 8 Apr 2011 05:39:01 +0000 (05:39 +0000)]
voronoi benchmark. On dc-11, the computation part runs 3x faster with a computation depth of 3.

It could probably do better but for some reason, when I add another sese (see Vertex.java line 115 where it's commented out), RuntimeConflictResolver.java crashes on line 506 (cannot find a conflict graph for a given sese).

13 years agofixes -nostalltr flag: if it is valid to prune heap examiners, turning off both of...
yeom [Fri, 8 Apr 2011 03:45:29 +0000 (03:45 +0000)]
fixes -nostalltr flag: if it is valid to prune heap examiners, turning off both of task & stallsite examiners

13 years agoEnable 1,2,4,8,16,32,50 cores' execution
jzhou [Fri, 8 Apr 2011 03:42:44 +0000 (03:42 +0000)]
Enable 1,2,4,8,16,32,50 cores' execution

13 years agoTODO: move record pool to a thread-local thing and never deallocate it. Issue is...
jjenista [Thu, 7 Apr 2011 02:33:53 +0000 (02:33 +0000)]
TODO: move record pool to a thread-local thing and never deallocate it.  Issue is that a child can retire, then its parent, then deallocate the child's entire mem pool... then the child's traverser finally gets around to the record

13 years agobug fix, at task exit we are looking for sources that cause virtual read of soemthing...
jjenista [Thu, 7 Apr 2011 02:32:00 +0000 (02:32 +0000)]
bug fix, at task exit we are looking for sources that cause virtual read of soemthing the task wrote, don't just look at parent and siblings, but also the parent's parent and the parent's sibling, etc.

13 years agochange on the labyrinth run script: making memory interleaved on all cpus as like...
yeom [Thu, 7 Apr 2011 02:24:50 +0000 (02:24 +0000)]
change on the labyrinth run script: making memory interleaved on all cpus as like other runscripts

13 years agohack seems to help a little...power isn't going to have good performance...
bdemsky [Thu, 7 Apr 2011 01:21:33 +0000 (01:21 +0000)]
hack seems to help a little...power isn't going to have good performance...

13 years agoD2 bug
bdemsky [Wed, 6 Apr 2011 23:33:51 +0000 (23:33 +0000)]
D2 bug

13 years agofix SOR bug
bdemsky [Wed, 6 Apr 2011 21:55:59 +0000 (21:55 +0000)]
fix SOR bug

13 years agobug and remove debug code
bdemsky [Wed, 6 Apr 2011 21:15:22 +0000 (21:15 +0000)]
bug and remove debug code

13 years agoget back some of the speed we are losing from bug fixes...
bdemsky [Wed, 6 Apr 2011 20:20:12 +0000 (20:20 +0000)]
get back some of the speed we are losing from bug fixes...

13 years agobug fix...wasn't adding out of context nodes that get summarized to set of ndoes...
bdemsky [Wed, 6 Apr 2011 20:01:46 +0000 (20:01 +0000)]
bug fix...wasn't adding out of context nodes that get summarized to set of ndoes that exist in context...

13 years agochange benchmark to run once, not up to seven times
jjenista [Wed, 6 Apr 2011 19:41:47 +0000 (19:41 +0000)]
change benchmark to run once, not up to seven times

13 years agoalter benchmark to run once, not up to seven times
jjenista [Wed, 6 Apr 2011 19:39:10 +0000 (19:39 +0000)]
alter benchmark to run once, not up to seven times

13 years agoget debug statements out of single version
jjenista [Wed, 6 Apr 2011 19:15:39 +0000 (19:15 +0000)]
get debug statements out of single version

13 years agoChanges
jzhou [Wed, 6 Apr 2011 16:50:39 +0000 (16:50 +0000)]
Changes

13 years agobug fix...things are slower to compile...:(
bdemsky [Wed, 6 Apr 2011 08:11:57 +0000 (08:11 +0000)]
bug fix...things are slower to compile...:(

13 years agobug fix for refCount in SMFEStates. I'm not sure if this fix covers all cases (it...
stephey [Wed, 6 Apr 2011 07:33:22 +0000 (07:33 +0000)]
bug fix for refCount in SMFEStates. I'm not sure if this fix covers all cases (it seems like it should to me)...

13 years agobug fix
bdemsky [Wed, 6 Apr 2011 03:31:46 +0000 (03:31 +0000)]
bug fix

13 years agobug fix and other change
bdemsky [Wed, 6 Apr 2011 03:24:48 +0000 (03:24 +0000)]
bug fix and other change

13 years agochecked in new version of bm
bdemsky [Wed, 6 Apr 2011 03:22:55 +0000 (03:22 +0000)]
checked in new version of bm

13 years agobug fix, have to dereference this pointer
jjenista [Wed, 6 Apr 2011 01:18:17 +0000 (01:18 +0000)]
bug fix, have to dereference this pointer

13 years agochanges.
yeom [Wed, 6 Apr 2011 00:49:05 +0000 (00:49 +0000)]
changes.

13 years agochanges: handle the case that there is only one work item after enqueue, then it...
yeom [Wed, 6 Apr 2011 00:39:23 +0000 (00:39 +0000)]
changes: handle the case that there is only one work item after enqueue, then it reports nothing and makes the worker steal from other queues.

13 years agobug fix...code that was never exercised...
bdemsky [Tue, 5 Apr 2011 23:03:42 +0000 (23:03 +0000)]
bug fix...code that was never exercised...

13 years agostatic initializers cause a problem for DFJ, easy enough to work around them
jjenista [Tue, 5 Apr 2011 23:02:07 +0000 (23:02 +0000)]
static initializers cause a problem for DFJ, easy enough to work around them

13 years agoversion of delaunay that has a spin bug
jjenista [Tue, 5 Apr 2011 21:50:53 +0000 (21:50 +0000)]
version of delaunay that has a spin bug

13 years agoget this debug print out of here
jjenista [Tue, 5 Apr 2011 21:50:25 +0000 (21:50 +0000)]
get this debug print out of here

13 years agogc bug fix
bdemsky [Tue, 5 Apr 2011 20:45:02 +0000 (20:45 +0000)]
gc bug fix

13 years agobug fix...
bdemsky [Tue, 5 Apr 2011 20:44:49 +0000 (20:44 +0000)]
bug fix...

13 years agoupdates.
yeom [Tue, 5 Apr 2011 20:26:12 +0000 (20:26 +0000)]
updates.

13 years agoat task exit, a task should acquire any out-set variables that arent already it the...
jjenista [Tue, 5 Apr 2011 17:40:14 +0000 (17:40 +0000)]
at task exit, a task should acquire any out-set variables that arent already it the task record.  These could be from static or dynamic sources, so some extra analysis info should be saved to generate the correct copy statements.  make sure static and dynamic tracking variables are generated too

13 years agospeculative version of delaunay works single-threaded just fine
jjenista [Mon, 4 Apr 2011 20:43:53 +0000 (20:43 +0000)]
speculative version of delaunay works single-threaded just fine

13 years agochanges on run scripts
yeom [Mon, 4 Apr 2011 18:34:18 +0000 (18:34 +0000)]
changes on run scripts

13 years agodfj version of delaunay refinement executes the original algorithm with workers=1...
jjenista [Mon, 4 Apr 2011 18:29:21 +0000 (18:29 +0000)]
dfj version of delaunay refinement executes the original algorithm with workers=1, but set to 2 another malformed triangle appears...

13 years ago#ifdef this out if not threads
bdemsky [Mon, 4 Apr 2011 01:41:40 +0000 (01:41 +0000)]
#ifdef this out if not threads

13 years agobug fix on the oooJava queue: when tries to add a vector item, get rid of finished...
yeom [Mon, 4 Apr 2011 01:14:06 +0000 (01:14 +0000)]
bug fix on the oooJava queue: when tries to add a vector item, get rid of finished item in front of it.

13 years agokeep a copy of the original delaunay refinement algorithm for comparison, and dfj...
jjenista [Mon, 4 Apr 2011 00:19:05 +0000 (00:19 +0000)]
keep a copy of the original delaunay refinement algorithm for comparison, and dfj version is crashing on nodes (triangles) with zero neighbors, but original configured graphs are the same size...

13 years agolock bug for gc...jin, make sure tilera gc doesn't have same bug...need to update...
bdemsky [Mon, 4 Apr 2011 00:06:45 +0000 (00:06 +0000)]
lock bug for gc...jin, make sure tilera gc doesn't have same bug...need to update thread lock pointer to locks at end of gc

13 years agobug fixes
bdemsky [Sun, 3 Apr 2011 23:43:43 +0000 (23:43 +0000)]
bug fixes

13 years agobug...
bdemsky [Sun, 3 Apr 2011 23:21:01 +0000 (23:21 +0000)]
bug...

13 years agougle hacks to get around compiler bugss...
bdemsky [Sun, 3 Apr 2011 21:52:03 +0000 (21:52 +0000)]
ugle hacks to get around compiler bugss...

13 years agobugs in buildflat...break optimization phase
bdemsky [Sun, 3 Apr 2011 21:41:10 +0000 (21:41 +0000)]
bugs in buildflat...break optimization phase

13 years agochanges: build conflict graphs that have line number information of stall sites
yeom [Sun, 3 Apr 2011 21:13:35 +0000 (21:13 +0000)]
changes: build conflict graphs that have line number information of stall sites

13 years agosimilar bugs to rcr
bdemsky [Sun, 3 Apr 2011 20:29:39 +0000 (20:29 +0000)]
similar bugs to rcr

13 years agoreworking parallel implementation, both single and DFJ runs build a zero-node configu...
jjenista [Sun, 3 Apr 2011 18:52:23 +0000 (18:52 +0000)]
reworking parallel implementation, both single and DFJ runs build a zero-node configuration right now, whoops

13 years agoAdd a warning when a method call does not resolve to a defined method
jjenista [Sun, 3 Apr 2011 18:51:44 +0000 (18:51 +0000)]
Add a warning when a method call does not resolve to a defined method

13 years agochanges
bdemsky [Sun, 3 Apr 2011 08:56:04 +0000 (08:56 +0000)]
changes

13 years agochanges to get pass specJBB validation
bdemsky [Sun, 3 Apr 2011 06:23:47 +0000 (06:23 +0000)]
changes to get pass specJBB validation

13 years agospecjbb build on intel now...
bdemsky [Sun, 3 Apr 2011 05:46:53 +0000 (05:46 +0000)]
specjbb build on intel now...

13 years agoadd wait/notify
bdemsky [Sun, 3 Apr 2011 05:38:59 +0000 (05:38 +0000)]
add wait/notify

13 years agochange
bdemsky [Sun, 3 Apr 2011 05:26:46 +0000 (05:26 +0000)]
change

13 years agochange
bdemsky [Sun, 3 Apr 2011 04:23:36 +0000 (04:23 +0000)]
change

13 years agoadd ieeeremainder
bdemsky [Sun, 3 Apr 2011 04:22:56 +0000 (04:22 +0000)]
add ieeeremainder

13 years agomake this work with MGC
bdemsky [Sun, 3 Apr 2011 04:21:45 +0000 (04:21 +0000)]
make this work with MGC

13 years agofix array
bdemsky [Sun, 3 Apr 2011 04:12:23 +0000 (04:12 +0000)]
fix array

13 years agosmall bug
bdemsky [Sun, 3 Apr 2011 04:09:24 +0000 (04:09 +0000)]
small bug

13 years agoadd enwmethos
bdemsky [Sun, 3 Apr 2011 03:56:30 +0000 (03:56 +0000)]
add enwmethos

13 years agochanges for correctness
bdemsky [Sun, 3 Apr 2011 03:56:14 +0000 (03:56 +0000)]
changes for correctness

13 years agotrying to build specjbb on intel
bdemsky [Sun, 3 Apr 2011 03:17:22 +0000 (03:17 +0000)]
trying to build specjbb on intel

13 years agochanges
bdemsky [Sun, 3 Apr 2011 03:17:05 +0000 (03:17 +0000)]
changes

13 years agofixed benchmark
bdemsky [Sat, 2 Apr 2011 23:26:43 +0000 (23:26 +0000)]
fixed benchmark

13 years agoreally, really nasty bug...see page 8-9 of vol 3A of intel processor manual for x86...
bdemsky [Sat, 2 Apr 2011 22:53:33 +0000 (22:53 +0000)]
really, really nasty bug...see page 8-9 of vol 3A of intel processor manual for x86 memory reordering...

13 years agonew information...x86 allows reads to be reordered with earlier writes to different...
bdemsky [Sat, 2 Apr 2011 22:49:40 +0000 (22:49 +0000)]
new information...x86 allows reads to be reordered with earlier writes to different locations....

add mbarrier for situations where we can't allow this to happen...

13 years agomissing barrier
bdemsky [Sat, 2 Apr 2011 21:03:59 +0000 (21:03 +0000)]
missing barrier

13 years agobug fix
bdemsky [Sat, 2 Apr 2011 20:22:58 +0000 (20:22 +0000)]
bug fix

13 years agostuff for running set up
jjenista [Sat, 2 Apr 2011 15:12:31 +0000 (15:12 +0000)]
stuff for running set up

13 years agomodified algorithm for dfj style parallelism
jjenista [Sat, 2 Apr 2011 15:06:07 +0000 (15:06 +0000)]
modified algorithm for dfj style parallelism

13 years agostatic variables now actually work...we don't want static variables here i think...
bdemsky [Sat, 2 Apr 2011 04:05:53 +0000 (04:05 +0000)]
static variables now actually work...we don't want static variables here i think though...

13 years agoadd RCR to the global gc initializer...bad if we have globals though
bdemsky [Sat, 2 Apr 2011 04:02:42 +0000 (04:02 +0000)]
add RCR to the global gc initializer...bad if we have globals though

13 years agoBug fixes and add some code for easy debugging
jzhou [Sat, 2 Apr 2011 03:39:06 +0000 (03:39 +0000)]
Bug fixes and add some code for easy debugging

13 years agobug fix for yeom...can you try to see if power works now?
bdemsky [Sat, 2 Apr 2011 03:38:01 +0000 (03:38 +0000)]
bug fix for yeom...can you try to see if power works now?

13 years agobuild pruned graphs that encode conflict information...
bdemsky [Fri, 1 Apr 2011 21:52:15 +0000 (21:52 +0000)]
build pruned graphs that encode conflict information...
double periphery nodes may contain conflicts...conflict effects are put in brackets...
edges that have conflicts are dashed

13 years agoThese are the files that were giving me trouble last night. The single version runs...
stephey [Fri, 1 Apr 2011 19:55:05 +0000 (19:55 +0000)]
These are the files that were giving me trouble last night. The single version runs fine, the rcr version errors out (either it's a RCR build problem, an RCR runtime error, or it screws with the data of the actual program and causes it to error out at Cavity.build...). This may indicate that there's something wrong with our runtime...

13 years agomake hashCode a native method for java lang Object...
bdemsky [Fri, 1 Apr 2011 02:10:17 +0000 (02:10 +0000)]
make hashCode a native method for java lang Object...

13 years agomove hashCode method to native method
bdemsky [Fri, 1 Apr 2011 02:03:53 +0000 (02:03 +0000)]
move hashCode method to native method

13 years agobug fix for stephen
bdemsky [Fri, 1 Apr 2011 00:20:42 +0000 (00:20 +0000)]
bug fix for stephen

13 years agoonly print out line numbers in debug mode
yeom [Thu, 31 Mar 2011 22:53:12 +0000 (22:53 +0000)]
only print out line numbers in debug mode

13 years agobring last changes before executing benchmakrs
yeom [Thu, 31 Mar 2011 21:45:51 +0000 (21:45 +0000)]
bring last changes before executing benchmakrs

13 years agochanges: reorganizes debug messages
yeom [Thu, 31 Mar 2011 19:08:32 +0000 (19:08 +0000)]
changes: reorganizes debug messages

13 years agoadd a new compiler flag -nolock: turning off synchronization lock
yeom [Thu, 31 Mar 2011 18:43:42 +0000 (18:43 +0000)]
add a new compiler flag -nolock: turning off synchronization lock

13 years agochanges for better debug messages
yeom [Thu, 31 Mar 2011 17:46:41 +0000 (17:46 +0000)]
changes for better debug messages

13 years agochanges.
yeom [Thu, 31 Mar 2011 17:26:22 +0000 (17:26 +0000)]
changes.

13 years agoAdded in semi-trivial locations (to test it). However, can't confirm results because...
stephey [Thu, 31 Mar 2011 09:52:47 +0000 (09:52 +0000)]
Added in semi-trivial locations (to test it). However, can't confirm results because the old make rcr takes a while (I didn't let it finish, but it was over 5 mins...) and the new one crashes with a NULLPOINTEREXCEPTION

13 years agoTried to squeeze out more performance by changing the LinkedLIsts in the Delaunay...
stephey [Thu, 31 Mar 2011 09:04:04 +0000 (09:04 +0000)]
Tried to squeeze out more performance by changing the LinkedLIsts in the Delaunay port to vectors (which is closer to the original implementation of ArrayLists). Seems not to make an appreciable difference though...

Added Vector.clone() and VectorIterator

13 years agoBug fix: should use unsigned int instead of int for pointers in shared heap. Also...
jzhou [Thu, 31 Mar 2011 03:46:19 +0000 (03:46 +0000)]
Bug fix: should use unsigned int instead of int for pointers in shared heap. Also as we now use a big share array to hold mapping information, do not need to allocate extra memory for master core's local heap.

13 years agochanges: only apply the variable analysis on the method containing TASK
yeom [Thu, 31 Mar 2011 03:00:34 +0000 (03:00 +0000)]
changes: only apply the variable analysis on the method containing TASK

13 years agochange
yeom [Thu, 31 Mar 2011 02:52:15 +0000 (02:52 +0000)]
change

13 years agoAdded manual invocation of garbage collector via System.gc()
stephey [Thu, 31 Mar 2011 02:28:21 +0000 (02:28 +0000)]
Added manual invocation of garbage collector via System.gc()

13 years agoOptimized HashMap/HashSet and added System.gc() for manual garbage collection invokation
stephey [Thu, 31 Mar 2011 02:27:46 +0000 (02:27 +0000)]
Optimized HashMap/HashSet and added System.gc() for manual garbage collection invokation

13 years agoChanges/optimizations to benchmark.
stephey [Thu, 31 Mar 2011 02:25:53 +0000 (02:25 +0000)]
Changes/optimizations to benchmark.

13 years agochange due to changes of math class library
yeom [Thu, 31 Mar 2011 01:26:31 +0000 (01:26 +0000)]
change due to changes of math class library

13 years agobug fixes on OoOJava, now it works fine with all of benchmarks(but, Kmeans has lower...
yeom [Wed, 30 Mar 2011 23:12:40 +0000 (23:12 +0000)]
bug fixes on OoOJava, now it works fine with all of benchmarks(but, Kmeans has lower speedup 10.6x than 13.8x. still working on...)
-changes on the potential stall site analysis: propagating the status of callee's return node to the caller region and when current node has a status change, making following nodes updated to get a new potential stall site status.
-changes on liveness analysis of OoOJava: new analysis only covers the task region, not whole region of the flat method.
-changes on disjoint analysis: rather than using reachgraph's inAccessibleSet, using the result of brian's new accessible analysis.

13 years agoThe benchmark works as far as I can tell (passes internal tests). We just need to...
stephey [Wed, 30 Mar 2011 06:22:47 +0000 (06:22 +0000)]
The benchmark works as far as I can tell (passes internal tests). We just need to put in sese blocks now if that's the case.

... There are things preventing me from doing a precise check.
1) It appears that given the same input and everything, the while loop will inevitably run a different number of times for each run (this occurs in both the original and ported code) and pick a different order of triangles to process.
2) The differences in how we and Java handle collection/sets gives a different ordering of which triangles we process (which affects which triangles become bad at the end).

Aside from that, there seems to be some performance issues.
1) It seems that the original code benefits from branch prediction as successive, duplicate runs (inherent in the benchmark) yields faster and faster timings. Our run-time stays the same in successive runs. This is true even though the number of triangles processed stays relatively close to each other.
2) When processing the large.2 files (about 10k triangles), the difference between our first runtime and the original code's runtime is within 100ms of each other (out of roughly 1000ms), which sounds reasonable (note that on successive runs, the original code speeds up by a factor of 1.6).  However when we process the massive.2 files (roughly 100k triangles), our runtimes are double the original code's runtime (but successive runs of the original code yields a speedup of only 1.12. Whatever speedup it got on the large.2 file is amortized). Also, our ported code garbage collects and then crashes (out of memory) on the second run (there are 3 repeated runs).

This is as much as I can do at the moment, going to need some guidance on the performance issues mentioned above and also where to place the sese blocks.

13 years agoIt compiles and runs now... but it doesn't appear to be doing the right thing......
stephey [Wed, 30 Mar 2011 04:11:15 +0000 (04:11 +0000)]
It compiles and runs now... but it doesn't appear to be doing the right thing... It passes internal verification but appears to be taking fewer iterations than the original program...

13 years agoCompiles with a few warnings... Doesn't run yet though.
stephey [Tue, 29 Mar 2011 23:37:34 +0000 (23:37 +0000)]
Compiles with a few warnings... Doesn't run yet though.

13 years agoAdded an error case
stephey [Tue, 29 Mar 2011 22:54:40 +0000 (22:54 +0000)]
Added an error case