search for: graal

Hey Scott, There has been quite a lot of research on capture tracking (aka escape analysis) for Java and other dynamic languages. See e.g.: https://wiki.openjdk.java.net/display/HotSpot/EscapeAnalysis http://docs.oracle.com/javase/7/docs/technotes/guides/vm/performance-enhancements-7.html http://dl.acm.org/citation.cfm?doid=320384.320386 Nuno -----Original Message----- From: Scott Egerton via

...r general design thing to keep in mind: Many complex >>> problems look simple once you find the right way to slice the problem. :) >>> >>> One really interesting approach I'd recommend you read is the "partial >>> escape analysis" stuff done by the Graal compiler project. It has a >>> lot of parallels to our mayBeCapturedBefore. One reasonable starting >>> point is: >>> https://wiki.openjdk.java.net/display/Graal/Graal+Partial+Escape+Analysis. >>> >>> I *think* the best paper starting point might be &q...

Hi everybody! I'm Josef and I'm working at Oracle Labs on Sulong [1,2], the LLVM IR execution engine in GraalVM [3]. In addition to executing bare bitcode files, Sulong also accepts ELF files with embedded bitcode sections. Therefore, it would be great if LLD in (Full)LTO mode would support embedding bitcode sections to the resulting object file. Is that something that would be considered useful and wo...

...- ISDP: In-Situ Data Processing technologies http://users.nccs.gov/~sklasky/isdp - LSAP2012: 4th Workshop on Large-scale Systems and Applications Performance http://www.lsap2012.org/ - MapReduce'12: The Third Int'l Workshop on MapReduce and its Applications http://graal.ens-lyon.fr/mapreduce - VTDC-2012: 6th Int'l Workshop on Virtualization Technologies in Distributed Computing http://www.grid-appliance.org/wiki/index.php/VTDC12 -- ================================================================= Ioan Raicu, Ph.D. Assistant Professor, Illinois In...

...- ISDP: In-Situ Data Processing technologies http://users.nccs.gov/~sklasky/isdp - LSAP2012: 4th Workshop on Large-scale Systems and Applications Performance http://www.lsap2012.org/ - MapReduce'12: The Third Int'l Workshop on MapReduce and its Applications http://graal.ens-lyon.fr/mapreduce - VTDC-2012: 6th Int'l Workshop on Virtualization Technologies in Distributed Computing http://www.grid-appliance.org/wiki/index.php/VTDC12 -- ================================================================= Ioan Raicu, Ph.D. Assistant Professor, Illinois In...

.... A patchpoint could do this, and so could spilling all > state to the stack before the divide - but both are very heavy hammers > that are sure to be more expensive than just doing a branch. > This isn't necessarily as expensive as you might believe. I'd recommend reading the Graal project papers on this topic. Whether deopt or branching is more profitable *in this case*, I can't easily say. I'm not yet to the point of being able to run that experiment. We can argue about what "should" be better all we want, but real performance data is the only way to...

Hi everyone, we started a new open source project Sulong: https://github.com/graalvm/sulong. Sulong is a LLVM IR interpreter with JIT compilation running on top of the JVM. By using the Truffle framework, it implements speculative optimizations such as inlining of function pointer calls through AST rewriting. One area of our research is to provide alternative ways of executing L...

...d spilling all >>> state to the stack before the divide - but both are very heavy >>> hammers that are sure to be more expensive than just doing a branch. >>> >> This isn't necessarily as expensive as you might believe. I'd >> recommend reading the Graal project papers on this topic. >> >> Whether deopt or branching is more profitable *in this case*, I can't >> easily say. I'm not yet to the point of being able to run that >> experiment. We can argue about what "should" be better all we want, >>...

...ould do this,  > and so could spilling all state to the stack before the divide - but both  > are very heavy hammers that are sure to be more expensive than just doing a  > branch.  >  > This isn't necessarily as expensive as you might believe. I'd recommend  > reading the Graal project papers on this topic.  >  > Whether deopt or branching is more profitable *in this case*, I can't easily  > say. I'm not yet to the point of being able to run that experiment. We can  > argue about what "should" be better all we want, but real performance data ...

...e the divide - but both are >>>>> very heavy hammers that are sure to be more expensive than just >>>>> doing a branch. >>>>> >>>> This isn't necessarily as expensive as you might believe. I'd >>>> recommend reading the Graal project papers on this topic. >>>> >>>> Whether deopt or branching is more profitable *in this case*, I >>>> can't easily say. I'm not yet to the point of being able to run >>>> that experiment. We can argue about what "should" b...

...tack before the divide - >> but both >> > are very heavy hammers that are sure to be more expensive than just >> doing a >> > branch. >> > >> > This isn't necessarily as expensive as you might believe. I'd recommend >> > reading the Graal project papers on this topic. >> > >> > Whether deopt or branching is more profitable *in this case*, I >> can't easily >> > say. I'm not yet to the point of being able to run that experiment. >> We can >> > argue about what "should&quo...

...int could do this, > and so could spilling all state to the stack before the divide - but both > are very heavy hammers that are sure to be more expensive than just doing a > branch. > > This isn't necessarily as expensive as you might believe. I'd recommend > reading the Graal project papers on this topic. > > Whether deopt or branching is more profitable *in this case*, I can't easily > say. I'm not yet to the point of being able to run that experiment. We can > argue about what "should" be better all we want, but real performance data &g...

Hey folks, After a long amount of discussion both offline and on, I put a pass/intrinsic to add extended SSA up at http://reviews.llvm.org/D29316. Sean asked me to share it more broadly on llvm-dev, so here you go :) For those not familiar with extended-SSA, it's described in the paper "ABCD: Eliminating Array Bounds Checks on Demand". There is a very large amount of explanation

As the discussion has progressed and I've spent more time thinking about the topic, I find myself less and less enthused about the current proposal. I'm in full support of having idiomatic ways to express safe division, but I'm starting to doubt that using an intrinsic is the right way at the moment. One case I find myself thinking about is how one would combine profiling