similar to: Proper way to do thinlto (from code)

You need to make sure that your bitcode file has a module summary index. You can do that either by creating a WriteBitcodePass with EmitSummaryIndex set to true or by calling WriteBitcodeToFile with a module summary index that you created yourself. It doesn't look like either of those options are hooked up to the c API, though. Peter On Fri, Mar 30, 2018, 00:50 Carlo Kok via llvm-dev

Thanks! I'll create a wrapper for WriteBitcodePass. Do I still need the PrepareForThinTLO pass or is the writeBitCodePass enough? On Fri, Mar 30, 2018, at 10:02, Peter Collingbourne wrote: > You need to make sure that your bitcode file has a module summary > index. You can do that either by creating a WriteBitcodePass with > EmitSummaryIndex set to true or by calling

I added your suggestion and am using this now llvm::legacy::FunctionPassManager *functionPassManager = new llvm::legacy::FunctionPassManager(Mod); llvm::PassRegistry &registry = *llvm::PassRegistry::getPassRegistry(); initializeScalarOpts(registry); functionPassManager->add( new llvm::TargetLibraryInfoWrapperPass(llvm::TargetLibraryInfoImpl(targetMachine->getTargetTriple())) );

Hi Richard, Thanks for working on this. Comments below. > I don't understand the basis for the last statement though. Do you mean > that you think most cases produce better code if scalarised at the SD stage > rather than at the IR level? Could you give an example? You presented an example that shows that scalarizing vectors allow further optimizations. But I don’t think that

Richard Sandiford <rsandifo at linux.vnet.ibm.com> writes: > Are you worried that adding it to PMB will increase compile time? > The pass exits very early for any target that doesn't opt-in to doing > scalarisation at the IR level, without even looking at the function. As an alternative, adding Scalarizer and InstCombine passes to SystemZPassConfig::addIRPasses() would probably

Hi, Our public library management software (PMB) runs on Apache/PHP/MySQL. It requires some additional PHP modules to run correctly, namely: 1) php-gd 2) php-yaz 3) php-xslt I've googled and fiddled around quite a bit, and come to the following conclusions: 1) php-gd can be installed from extra repos (rpmforge IIRC), so this one's no problem. 2) To install php-yaz, I have to

I'm using the terminal commands from http://appdb.winehq.org/objectManager.php?sClass=version&iId=19141 The LeagueofLegends.exe starts fine, but when it gets to the part where it starts Pando Media Booster, I get a Program Error saying "The program PMB.exe has encountered a serious problem and needs to close. We are sorry for the inconvenience." Under that it says: "This

Nadav Rotem <nrotem at apple.com> writes: >> I don't understand the basis for the last statement though. Do you mean >> that you think most cases produce better code if scalarised at the SD stage >> rather than at the IR level? Could you give an example? > > You presented an example that shows that scalarizing vectors allow > further optimizations. But I

Nadav Rotem <nrotem at apple.com> writes: > On Nov 14, 2013, at 2:32 PM, Richard Sandiford > <rsandifo at linux.vnet.ibm.com> wrote: >> Richard Sandiford <rsandifo at linux.vnet.ibm.com> writes: >>> Are you worried that adding it to PMB will increase compile time? >>> The pass exits very early for any target that doesn't opt-in to doing

I have some trouble making the SIMD vector length visible to the passes. My application is basically on the level of 'opt'. What I did in version 3.6 was functionPassManager->add(new llvm::TargetLibraryInfo(llvm::Triple(Mod->getTargetTriple()))); functionPassManager->add(new llvm::DataLayoutPass()); and then the -basicaa and -loop-vectorizer were able to vectorize the input

Hi: In my in-house transform passes I needs to properly lower Switch/Invoke and PhiNodes, prior to LLVM9 I just call createLowerSwitchPass() and run it. However in LLVM9 doing this would assert out with error message: &nbsp;&nbsp; Assertion failed: (Resolver &amp;&amp; "Pass has not been inserted into a PassManager object!"), function getAnalysis, file

On Jul 29, 2013, at 9:05 AM, Krzysztof Parzyszek <kparzysz at codeaurora.org> wrote: > On 7/16/2013 11:38 PM, Andrew Trick wrote: >> Since introducing the new TargetTransformInfo analysis, there has been some confusion over the role of target heuristics in IR passes. A few patches have led to interesting discussions. >> >> To centralize the discussion, until we get

Hi, I'm running a few PHP-based apps on our server (PMB, SPIP, Joomla, PHPMyAdmin), and I'm not always comforted about security. I don't know the details, but many a security expert frowns when it comes to PHP. Now I just stumbled over this: http://www.hardened-php.net/suhosin.127.html Has anyone already tried this out? An opinion about it? Is it worth it? Since I have to

Hi Richard, The discussion on llvmpipe is irrelevant. llvmpipe has its own pass manager and optimization pipe, it is not a C compiler. Nadav On Nov 15, 2013, at 3:26 AM, Richard Sandiford <rsandifo at linux.vnet.ibm.com> wrote: > Nadav Rotem <nrotem at apple.com> writes: >> On Nov 14, 2013, at 2:32 PM, Richard Sandiford >> <rsandifo at linux.vnet.ibm.com>

On Nov 14, 2013, at 2:32 PM, Richard Sandiford <rsandifo at linux.vnet.ibm.com> wrote: > Richard Sandiford <rsandifo at linux.vnet.ibm.com> writes: >> Are you worried that adding it to PMB will increase compile time? >> The pass exits very early for any target that doesn't opt-in to doing >> scalarisation at the IR level, without even looking at the function.

On 7/29/2013 6:28 PM, Andrew Trick wrote: > > You mean that LICM and Unswitching should be left for later? For the purpose of exposing scalar optimizations, I'm not sure I agree with that but I'd be interested in examples. Optimizations like LICM, and unswitching can potentially damage perfect nesting of loops. For example, consider this nest: for (i) { for (j) {

Hi, I would like to run the -O1 pass sequence followed by -reg2mem from an out of tree project which uses llvm. I am using the following code snippet to do so but in some cases, my method is also vectorising the code, which doesn't happen when running the same sequence(-S -O1 -reg2mem) through opt. Can someone please help me find what I am missing? Thanks, Akash. *

Hi guys, How's things going ? Got a bit of a weird one here that I've been unable to solve. I have a Panasonic PBX linked to a Sirrix Quad BRI card that is running in TE (ptp) mode in a Asterisk box - this then links through Internet to another Asterisk box via IAX2. When a user on the Panasonic PBX system dials the extension of my Sirrix Asterisk box, Asterisk answers and says

Hi, I have finished migrating all of the simplify-libcalls pass functionality into instcombine and functionattrs. Now I am ready to completely to remove what is left of the pass from the source tree. However, there are a few C API functions for creating and managing the pass: /** See llvm::PassManagerBuilder::DisableSimplifyLibCalls */ void

Hi, I would like to know how (if) I can extract some of the information from the summary of my nlme. at present, I get a summary looking something like this: > summary(fit.nlme) Nonlinear mixed-effects model fit by maximum likelihood Model: MLKYLD ~ W4(DIM, logA, B, C) Data: ADHIS.x0 AIC BIC logLik 265314 265401.6 -132647 Random effects: Formula: list(logA ~ 1 , B ~