similar to: [LLVMdev] llvm ir vs dag

On Apr 24, 2013, at 6:27 PM, Reed Kotler <rkotler at mips.com> wrote: > I would really push towards doing this in LLVM IR as the next step. What makes you say that? > It's possible that what you are proposing is the right "long term" solution but I think it's not a good evolutionary approach; it's more revolutionary. Doing this in LLVM IR seems like a major

On 04/24/2013 07:39 PM, Chris Lattner wrote: > On Apr 24, 2013, at 6:27 PM, Reed Kotler <rkotler at mips.com> wrote: >> I would really push towards doing this in LLVM IR as the next step. > What makes you say that? > Partly for the reasons Dan stated. For me, the IR is definitely way more friendly too and not tangled up in lots of undocumented obscurity as selection DAG is

How can I custom lower the ISD::STORE? I am using -enable-legalize-types and trying to customize most of our operations in xxxTargetLowering::ReplaceNodeResults(...) There are hooks to get trunk-store and indexed-store customized, But I can't get regular STORE customized... Any suggestions? Thanks Alireza Moshtaghi Senior Software Engineer Development Systems, Microchip Technology

On 04/24/2013 05:26 PM, Chris Lattner wrote: > On Apr 24, 2013, at 5:01 PM, Dan Gohman <dan433584 at gmail.com> wrote: >> In the spirit of the (long-term) intent to migrate away from the SelectionDAG framework, it is desirable to implement legalization passes as discrete passes. Attached is a patch which implements the beginning of a new type legalization pass, to help motivate

Is there a way to find the IR that corresponds to a callee DAG node? In other words to get the function definition, attributes, etc.? Tia. Reed

To generate code for a target which doesn't have immediate constant as instruction operand, do I (the target specific back-end, XXXTarget) need to provide code to break up the SDNode with constant (like ISD::ADD $reg1, #1) to 2 SDNodes (ISD::LOAD $reg2, #1; ISD::ADD $reg1, $reg2) in my XXXTargetLowering::LowerOperation, or LLVM target-independent framework can do such transformation

On 04/28/2013 09:49 PM, reed kotler wrote: > Is there a way to find the IR that corresponds to a callee DAG node? > > In other words to get the function definition, attributes, etc.? > > Tia. > > Reed It looks like you can do this by: 1) getParent will return the Module that some global value is in 2) in Module, getFunction will return the function

Hi, Today I downloaded LLVM 1.8 (until yesterday I used 1.7) and I tried to build it (Slackware 10.2 with default GCC 3.3.6). However, the build process fails each time it reaches lib/Target/XXX/XXXISelDAGToDAG.cpp for all backends (XXX=X86, ARM, ...), e.g. like: /.../llvm-build/lib/Target/ARM/ARMGenDAGISel.inc:145: error: syntax error before `{' token ... [lots of other errors

Hi David, > Previously, I had been testing with only one routine per test .ll file, but I > thought I'd reached a point where I could test multiple operations at once and > understand the output. The odd part about this is that the likelihood of seeing > the above assertion scales with the number of functions in the .ll file. If I > have one or two functions, I never see it.

Hello, I've made some fair progress on a target for 6502 family CPUs recently, but I've run into an error I'm not sure how to address. I've ruminated over it for about a week now, trying various things and not having any success. It seems to scale with the number of routines in my .ll file, which I am trying to run through llc. I get the following stack dump from an assertion:

Thanks for the suggestion, Duncan. I recently figured out that it had to do with how I was removing the pseudo instruction in my overridden expandPostRAPseudo() implementation. // member function's signature bool TheInstrInfo::expandPostRAPseudo(MachineBasicBlock::iterator p_mi) // works bb.erase(p_mi); // produces the assertion / memory leak. p_mi->removeFromParent(); I should have

On 04/24/2013 07:39 PM, Chris Lattner wrote: > On Apr 24, 2013, at 6:27 PM, Reed Kotler <rkotler at mips.com> wrote: >> I would really push towards doing this in LLVM IR as the next step. > > What makes you say that? > >> It's possible that what you are proposing is the right "long term" solution but I think it's not a good evolutionary approach;

On 25/07/16 16:31, Data Control Systems - Mike Elkevizth wrote: > Hi Mark, > > I'm not sure why a DC ignores the "winbind use default domain = yes" > setting. Its not the only setting that a DC ignores and the only real hint > of DCs acting weird is the line in the introduction of the wiki about > setting Samba up as a DC that calls these "idiosyncrasies in

removeFromParent just unlinks it from the basic block and returns the removed instruction. It does not delete it. On Sun, Jan 20, 2013 at 9:54 PM, David Waggoner <mathonnapkins at gmail.com>wrote: > > Thanks for the suggestion, Duncan. > > I recently figured out that it had to do with how I was removing the > pseudo instruction in my overridden expandPostRAPseudo()

Cool! This is great news. I feel like this information should be in our documentation somewhere. Could you start a new file ContinuousIntegration.rst and use this content to seed it? This new page would also be a good place to mention some LLVM idiosyncrasies like smooshlab being Apple-internal but still reporting via IRC; these things have not had a good place to be put yet. AFAIK currently our

On Mon, June 29, 2015 02:14, Sorin Srbu wrote: OS 6? > > Please note: I'm not criticizing, just curious about the argument > behind using a regular OS to do firewall-stuff. > Maintenance. A consistent set of expectations does wonders for debugging odd-ball occurrences. Why learn the idiosyncrasies of two distros when one suffices? Just start with a minimal CentOS install on

Hi Reed, Sent from my iPad On Apr 25, 2013, at 3:46 PM, Reed Kotler <rkotler at mips.com> wrote: > On 04/24/2013 07:39 PM, Chris Lattner wrote: >> On Apr 24, 2013, at 6:27 PM, Reed Kotler <rkotler at mips.com> wrote: >>> I would really push towards doing this in LLVM IR as the next step. >> >> What makes you say that? >> >>> It's

Hi, The `system.time(expr)' command provide 3 different times for evaluating the expression `expr'; the first two are user and system CPUs and the third one is total elapsed time. Suppose I want to compare two different computational procedures for performing the same task, which component of `system.time' is most meaningful in the sense that it most accurately reflects the

On Fri, 15 Mar 2019 at 16:50, James Y Knight <jyknight at google.com> wrote: >> Ie. the current series is already dead, no matter what we do > > But this last statement seems odd. So far, there looks to be a fairly good consensus from a number of experienced llvm developers that the approach seems like a good idea, both on this thread, and from skimming the earlier threads you

Apologies for the limitations of the registration system. - No email is sent upon completion - If you get a page at the end that says you successfully registered, then you are good to go - There have been a few reported cases of the registration page hanging, if so try another browser such as Firefox. If you have questions mail me directly (dkipping at qualcomm.com). If the group wants, I can