search for: linscan

> Are you using the NightlyTest.pl file? To run the tester I don't use > gmake directly at all. Instead, I have this in my crontab: > > 5 3 * * * LLVM_LIB_SEARCH_PATH=/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs $HOME/llvm/utils/NightlyTest.pl -parallel -enable-linscan /home/vadve/shared/PublicCVS $HOME/buildtest-X86 $HOME/cvs/testresults-X86 > > When run like this, you can use -noexternals as you wish. Thank you fo information. I will used this code for tester run. Bu i have problem - script:termnated with error: "Cannot change directory to: llvm s...

...roving > register allocation for large blocks (which was the target in the case > of BigBlock). They are about register allocation for general purpose > CFGs. And they report improvements (greatly reduced number of > load/stores for spills) over linear scan. Sure, we won't keep the linscan algorithm. The current LLVM linscan implementation is not linear time in any case, and shares little with the commonly published algorithms. -Chris

...I must do for enable it Are you using the NightlyTest.pl file? To run the tester I don't use gmake directly at all. Instead, I have this in my crontab: 5 3 * * * LLVM_LIB_SEARCH_PATH=/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs $HOME/llvm/utils/NightlyTest.pl -parallel -enable-linscan /home/vadve/shared/PublicCVS $HOME/buildtest-X86 $HOME/cvs/testresults-X86 When run like this, you can use -noexternals as you wish. -Chris -- http://llvm.cs.uiuc.edu/ http://www.nondot.org/~sabre/Projects/

At this moment I use "gmake TEST=nightly" command line to run night tester What I must specific and where to disable SPEC run (i don't have it) I known about new NightlyTest.pl option: -noexternals But what I must do for enable it Vladimir

On Jul 29, 2009, at 3:14 PM, Roman Levenstein wrote: > Hi Lang, > > There are at least two projects that were using BigBlock, directly or > indirectly. Hi Roman, We have many plans to rip out linscan and replace it with something that handles large blocks better. That's not the issue :). The problem is that bigblock is unmaintained and bitrotted. Since noone is working on it, it doesn't work, and it is a maintenance burden, I think it should be removed. -Chris > > One...

...?id=5010. The basic description of the issue (from http://llvm.org/bugs/show_bug.cgi?id=4668#c5): "The fundamental problem is we can't spill a register once it's fixed to a physical register." >From discussion on IRC: [17:14] <_sabre_> efriedma: sounds like a RA bug in linscan [17:14] <_sabre_> unfortunately, it is a fairly well known bug that is difficult to fix [17:14] <_sabre_> it would be worth bringing up on llvmdev, there may be a good approach since I stopped looking at the regalloc [17:14] <_sabre_> in any case, its worth surfacing the issue so...

...Chris Lattner <clattner at apple.com>: > > On Jul 29, 2009, at 3:14 PM, Roman Levenstein wrote: > >> Hi Lang, >> >> There are at least two projects that were using BigBlock, directly or >> indirectly. > > Hi Roman, > > We have many plans to rip out linscan and replace it with something > that handles large blocks better.  That's not the issue :). As far as I know, we have these ambitious plans since a while (just X years), but there is no good replacement for a LLVM linear scan yet :) The alternatives are either slower at compile time or gene...

...Gen/LiveIntervals.cpp. > > As for CVS, I am not sure this should be fixed because there is > really no LLVM bug here :-) While this is not an LLVM bug, if Finn has run into it, other people probably will too. I've fixed it in CVS: if you update llvm/lib/CodeGen and rebuild, hopefully linscan will magically work for Finn now. :) -Chris

...te that it looks like your tree is a bit out of date. You might > > try updating and see if it helps the problem. > > I've updated and rebuild. The bug is still there and the output you've asked > for is attached. Okay, it looks like a bug in your code generator (though the linscan allocator should give a better assertion). Specifically, you never define the %reg1028 register. -Chris -- http://llvm.cs.uiuc.edu/ http://www.nondot.org/~sabre/Projects/

...using the NightlyTest.pl file? To run the tester I don't use > > gmake directly at all. Instead, I have this in my crontab: > > > > 5 3 * * * > LLVM_LIB_SEARCH_PATH=/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs > $HOME/llvm/utils/NightlyTest.pl -parallel -enable-linscan > /home/vadve/shared/PublicCVS $HOME/buildtest-X86 $HOME/cvs/testresults-X86 > > > > When run like this, you can use -noexternals as you wish. > Thank you fo information. > > I will used this code for tester run. > > Bu i have problem - script:termnated with error: &...

...ble it > > Are you using the NightlyTest.pl file? To run the tester I don't use > gmake directly at all. Instead, I have this in my crontab: > > 5 3 * * * LLVM_LIB_SEARCH_PATH=/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs $HOME/llvm/utils/NightlyTest.pl -parallel -enable-linscan /home/vadve/shared/PublicCVS $HOME/buildtest-X86 $HOME/cvs/testresults-X86 > > When run like this, you can use -noexternals as you wish. > For note: I am use suggested in http://llvm.cs.uiuc.edu/docs/TestingGuide.html#run method: "For example, you could run the nightly tester tests u...

Hi! I'm developing a register allocator that works iteratively. It spills some virtual registers on each iteration until all the rest have physical ones assigned. How can I spill some live intervals at the end of each iteration with new live intervals having correct weights? Thanks. -------------- next part -------------- An HTML attachment was scrubbed... URL:

...e linear scan allocator. It is also iterative: it uses the spiller interface to insert spill code, which creates (unspillable) intervals for the spill code it inserts. > How can I spill some live intervals at the end of each iteration with new > live intervals having correct weights? The linscan allocator inserts spill code with infinite weight, take a look at how it works. -Chris -- http://nondot.org/sabre/ http://llvm.org/

...ne::addInstSelector method, add it to the pass mgr right after your isel. > And BTW, it seems to me that currently new RA passes are not allowed > to derive from the existing ones. If it is correct, why so? Wouldn't it > be nice? I'm not sure what you mean. We don't expose linscan through a public header, but a pass in the same .cpp file could subclass it. We haven't had a need to do this yet, so we don't have the provisions to do it. -Chris -- http://nondot.org/sabre/ http://llvm.org/

...lvm-gcc/bin/gcc > 5) A copy of the Makefile.config file that is created in > the object root directory after running configure. ------------- - attached. Chris wrote in a followup: > Can you send the output of 'llc -o - foo.bc -debug -print-machineinstrs'? Attached as "linscan". (But added the "-regalloc=linearscan" to provoke the error). For completenes, I also attach the a.out.bc bytecode, although I expect it can be generated from a.out.ll. -------------- That's it. I want to thank you very much for offering this kind of support. I hope it won&...

...using the NightlyTest.pl file? To run the tester I don't use > > gmake directly at all. Instead, I have this in my crontab: > > > > 5 3 * * * > LLVM_LIB_SEARCH_PATH=/home/vadve/lattner/local/x86/llvm-gcc/bytecode-libs > $HOME/llvm/utils/NightlyTest.pl -parallel -enable-linscan > /home/vadve/shared/PublicCVS $HOME/buildtest-X86 $HOME/cvs/testresults-X86 > > > > When run like this, you can use -noexternals as you wish. > Thank you fo information. > > I will used this code for tester run. > > Bu i have problem - script:termnated with error: &gt...

...a few test failure from the Frontend{C,C++} directories. I suppose this is because I have llvm-gcc4.0 2.1 installed, which might need to be upgraded. Is there a specific version of llvm-gcc that is prefered for nightly tests? Shall I go with the latest stable one? 4.0 or 4.2? What does -enable-linscan do? It's not referenced anywhere in NewNightlyTest.pl, but it's given as example in [2]. I will also take the freedom to disable the nightly test if I need the CPU cycles for something else during a night. jlh [1] http://llvm.org/nightlytest/test.php?machine=288&night=6315 [2] http:...

...On Sun, 2004-09-05 at 10:01, Vladimir Merzliakov wrote: > > x86 FreeBSD: > > * hasn't run with changes yet > > I manually start script. I use this options (+ -verbose now for testing) > > /home/wanderer/pkg/build/llvm/src/llvm/utils/NightlyTest.pl -parallel -enable-linscan > -noexternals -noremove :pserver:anon at llvm-cvs.cs.uiuc.edu:2401/var/cvs/llvm > /home/wanderer/pkg/build/llvm/night/build > /home/wanderer/pkg/build/llvm/night/testresults-X86-FreeBSD > > Log: > INITIALIZED > CVS Root = :pserver:anon at llvm-cvs.cs.uiuc.edu:2401/var/cvs...

> x86 FreeBSD: > * hasn't run with changes yet I manually start script. I use this options (+ -verbose now for testing) /home/wanderer/pkg/build/llvm/src/llvm/utils/NightlyTest.pl -parallel -enable-linscan -noexternals -noremove :pserver:anon at llvm-cvs.cs.uiuc.edu:2401/var/cvs/llvm /home/wanderer/pkg/build/llvm/night/build /home/wanderer/pkg/build/llvm/night/testresults-X86-FreeBSD Log: INITIALIZED CVS Root = :pserver:anon at llvm-cvs.cs.uiuc.edu:2401/var/cvs/llvm-test BuildDir = /home/wandere...

Hi, >> have not found any easy way to do it yet. For post-RA pass it is >> very easy and supported, but for pre-RA or post-code-selection - it >> is non obvious. > > I suggest a third approach: [snip] > 3. Write a machine function pass that runs before the RA that > translates these instructions into libcalls or other integer ops. > This would lower the f64