Re: Events output fail in libjava
Revital1 Eres wrote: > Hello, > > I am testing some patch on r130269, running on ppc64-linux and I get > the following events output fail in libjava with the new patch applied > (the fail does not appear in the vanilla version). The error does not > seem to be related to the patch I am testing, but to make sure I am not > mistaken, I appreciate another opinion regarding this. btw - I saw this > fail on different patches after testing them and compare the result to > the vanilla version. > > This is the diff of the libjava.log files of the gcc version w and w\o > the patch: > ... > created JVMTI environment #1 > created JVMTI environment #2 > setting callbacks for envs > -RequestedEvents: > +RequestedEvents: VMInit, > enable VM_INIT for env0, env1, env2 > ... > -PASS: events output > ... > +FAIL: events output > > It is expected that 'events output' will FAIL on a multi-lib system when testing against a non-default ABI. See: http://gcc.gnu.org/ml/gcc-testresults/2007-11/msg00952.html If this is not the case, then you may have a problem. David Daney
Events output fail in libjava
Hello, I am testing some patch on r130269, running on ppc64-linux and I get the following events output fail in libjava with the new patch applied (the fail does not appear in the vanilla version). The error does not seem to be related to the patch I am testing, but to make sure I am not mistaken, I appreciate another opinion regarding this. btw - I saw this fail on different patches after testing them and compare the result to the vanilla version. This is the diff of the libjava.log files of the gcc version w and w\o the patch: ... created JVMTI environment #1 created JVMTI environment #2 setting callbacks for envs -RequestedEvents: +RequestedEvents: VMInit, enable VM_INIT for env0, env1, env2 ... -PASS: events output ... +FAIL: events output Thanks, Revital
Re: undocumented optimization options
Mark Mitchell <[EMAIL PROTECTED]> wrote on 12/11/2007 18:25:03: > Richard Guenther wrote: > > > For options that are not enabled by default at any -O level the > documentation > > should be a help for the user to decide whether to enable or not enable it. > > So, yes, if there are non-standard terms used or if there are not obvious > > side-effects (like performing function versioning for ipa-cp) > those should be > > documented. > > > > For example googling for "interprocedural constant propagation" gives you > > an idea what it does by citing papers. Googling for "matrix flattening > > transposing" instead gives you only patches ;) > > To echo Richard: this is a user manual, so we don't need to say exactly > what the option does, but we have to say when you should use it. :-) In > other words, we certainly don't want to say anything about SSA names or > whatever, but we do want to explain what it's for. For interprocedural > constant propagation, maybe something like: > > == > This optimization analyzes the program to determine when values passed > to, or return from, functions are constants and then optimizes > accordingly. This optimization can substantially increase performance > if the application has lots of functions that return constants, but, > because this optimization can create multiple copies of code fragments, > it may significantly increase code size. > == > Thanks. Described these two optimizations in more detail. O.K now? Thanks, Razya 2007-11-20 Razya Ladelsky <[EMAIL PROTECTED]> * doc/invoke.texi (fipa-cp, fipa-matrix-reorg): Add documentation. > -- > Mark Mitchell > CodeSourcery > [EMAIL PROTECTED] > (650) 331-3385 x713 Index: invoke.texi === --- invoke.texi (revision 130311) +++ invoke.texi (working copy) @@ -328,7 +328,8 @@ -ffunction-sections -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm @gol -fgcse-sm -fif-conversion -fif-conversion2 -finline-functions @gol -finline-functions-called-once [EMAIL PROTECTED] @gol --finline-small-functions -fipa-pta -fipa-pure-const -fipa-reference @gol +-finline-small-functions -fipa-cp -fipa-marix-reorg -fipa-pta @gol +-fipa-pure-const -fipa-reference @gol -fipa-type-escape -fivopts -fkeep-inline-functions -fkeep-static-consts @gol -fmerge-all-constants -fmerge-constants -fmodulo-sched @gol -fmodulo-sched-allow-regmoves -fmove-loop-invariants -fmudflap @gol @@ -5715,6 +5716,29 @@ @opindex fipa-pta Perform interprocedural pointer analysis. [EMAIL PROTECTED] -fipa-cp [EMAIL PROTECTED] fipa-cp +Perform interprocedural constant propagation. +This optimization analyzes the program to determine when values passed +to functions are constants and then optimizes accordingly. +This optimization can substantially increase performance +if the application has constants passed to functions, but +because this optimization can create multiple copies of functions, +it may significantly increase code size. + [EMAIL PROTECTED] -fipa-matrix-reorg [EMAIL PROTECTED] fipa-matrix-reorg +Perform matrix flattening and transposing. +Matrix flattening tries to replace a m-dimensional matrix +with its equivalent n-dimensional matrix, where n < m. +This reduces the level of indirection needed for accessing the elements +of the matrix. The second optimization is matrix transposing that +attemps to change the order of the matrix's dimensions in order to +improve cache locality. +Both optimizations need fwhole-program flag. +Transpoing is enabled only if profiling information is avaliable. + + @item -ftree-sink @opindex ftree-sink Perform forward store motion on trees. This flag is =
Re: undocumented optimization options
On Nov 20, 2007 12:32 PM, Razya Ladelsky <[EMAIL PROTECTED]> wrote: > Mark Mitchell <[EMAIL PROTECTED]> wrote on 12/11/2007 18:25:03: > > > > Richard Guenther wrote: > > > > > For options that are not enabled by default at any -O level the > > documentation > > > should be a help for the user to decide whether to enable or not > enable it. > > > So, yes, if there are non-standard terms used or if there are not > obvious > > > side-effects (like performing function versioning for ipa-cp) > > those should be > > > documented. > > > > > > For example googling for "interprocedural constant propagation" gives > you > > > an idea what it does by citing papers. Googling for "matrix > flattening > > > transposing" instead gives you only patches ;) > > > > To echo Richard: this is a user manual, so we don't need to say exactly > > what the option does, but we have to say when you should use it. :-) In > > other words, we certainly don't want to say anything about SSA names or > > whatever, but we do want to explain what it's for. For interprocedural > > constant propagation, maybe something like: > > > > == > > This optimization analyzes the program to determine when values passed > > to, or return from, functions are constants and then optimizes > > accordingly. This optimization can substantially increase performance > > if the application has lots of functions that return constants, but, > > because this optimization can create multiple copies of code fragments, > > it may significantly increase code size. > > == > > > > Thanks. > Described these two optimizations in more detail. > O.K now? Ok. Thanks! Richard. > Thanks, > Razya > > 2007-11-20 Razya Ladelsky <[EMAIL PROTECTED]> > > * doc/invoke.texi (fipa-cp, fipa-matrix-reorg): Add documentation.
Re: SIMD-enabled and -lpthread incompatible?
Daniel Verkamp wrote: > The MMX registers are aliased onto the x87 floating-point registers, > so they should be saved and restored correctly regardless. That's not true of all CPU implementations. If you want to be sure your code will run only on the CPUs where they are distinct, don't take that precaution.
bug with -fleading-underscore doubling underscore in some cases on templates
hi, we have found a case where g++ will generate template instantiations with two underscore prefixes (std::string, "__ZNSs7reserveEm" instead of just "_ZNSs7reserveEm") while other .cpp files will compile just fine. the template stabs sections are correctly setup as "_ZNSs7reserveEm" but the assembly code that calls the functions have two underscores instead so the linker cannot find the template and complain of undefined reference to __X I tried moving things around in that particular cpp file to make the bug go away (and isolate the case) to no avail. and std::string is the only template that seem to cause the behaviour i386-unknown-elf-g++ (GCC) 4.1.1 -- Stephane Hockenhull SSC-Studios.com
Re: bug with -fleading-underscore doubling underscore in some cases on templates
On Tue, Nov 20, 2007 at 01:59:23PM -0500, Stephane Hockenhull wrote: > hi, > > we have found a case where g++ will generate template instantiations with two > underscore prefixes (std::string, "__ZNSs7reserveEm" instead of just > "_ZNSs7reserveEm") while other .cpp files will compile just fine. Please report a bug: http://gcc.gnu.org/bugs.html -- Daniel Jacobowitz CodeSourcery
Re: Build failure of trunk on i686-pc-gnu-linux
Hello! On Mon, Nov 19, 2007 at 05:22:52PM -0800, Andrew Pinski wrote: > On 11/19/07, Thomas Schwinge <[EMAIL PROTECTED]> wrote: > > I'm glad that nobody answered. I've been doing nothing wrong. And also > > it's not someone else's fault. But let me tell you what in fact was > > going wrong. > > I think the reason why nobody answered is because we are all busy > doing real work. Now I'm not sure about how to take that one. I for sure didn't mean to blame or even insult someone of you GCC people. But then, why do you now in turn think I wasn't trying to get ``real work'' done? > > I had a some-months-old checkout of GCC trunk on my disk. Now I needed > > such a checkout and thought: ``Good that the old one -- which I only used > > once for compiling and didn't do any modifications to it -- is still > > laying around: I'll simply use that one after having ran a ``svn update'' > > in there.'' I also ran a ``svn diff''. No modifications shown. Also, I > > attempted a ``svn status'', but as that one -- apart from completely > > saturating the system's disk I/O (Pentium M with 1.7 GHz, 768 MiB of RAM) > > -- didn't give any output after having it had running for half a minute, > > I terminated it again. After all, both ``svn update'' and ``svn diff'' > > had made me think that the tree was as clean as it could be. (SVN was > > made for easy transition from CVS, right?) > > I use svn stat all the time without any issues. It does take over a > half a minute for some of my local trees (they are not FSF GCC checks > and include a lot more files) but it is reasonable. On a freshly booted, otherwise idle system: #v+ $ time svn status M libgcc/config.host M gcc/config.gcc M gcc/unwind-generic.h M gcc/config/i386/gnu.h M gcc/config/gnu.h real2m17.461s user0m2.972s sys 0m2.736s #v- Subsequent runs, as long as the data is still being cached, are much faster, of course. > What version of svn are you using? It might be something which is > already fixed in a newer version of svn. This is what Ubuntu feisty ships, which is version 1.4.3 (r23084), Ubuntu package version 1.4.3dfsg1-1ubuntu1.2. Regards, Thomas signature.asc Description: Digital signature
