similar to: [LLVMdev] Converting a i32 pointer to a vector of i32 ( C array to LLVM vector)

If what you are saying is that you know the array of i32 will always be 10 entries, make the function use a constant limit=10 to the loop. I.e Make the loop limit a constant and not a variable. On 11 October 2013 18:27, Matthieu Dubet <maattdd at gmail.com> wrote: > Hi, > > I'm creating a small function in LLVM which gets as a parameter an i32* > (this function is called

On 11 October 2013 18:27, Matthieu Dubet <maattdd at gmail.com> wrote: > How can I tell LLVM to consider this i32* as an <10 x i32> (and thus get > the performance improvements thanks to SIMD ..etc..) ? > Hi Matthieu, You shouldn't need to do anything, the vectorizer should spot that for you, if the machine you're compiling to has support for vector instructions.

Hi, Thank you for the information, So I'm now keeping the array as a pointer (i32*) but the vectorizer doesn't vectorize it . I've pasted the function code before and after optimization (and the list of optimization that I have activated) in this Gist : https://gist.github.com/maattd/7008683 Some "weird" fact of my LLVM code : * all variables (even the one used for the

On Wed, Oct 16, 2013 at 11:14:06AM -0400, Matthieu Dubet wrote: > Hi, > > Thank you for the information, > > So I'm now keeping the array as a pointer (i32*) but the vectorizer doesn't > vectorize it . > > I've pasted the function code before and after optimization (and the list > of optimization that I have activated) in this Gist : >

Both the SLP vectorizer and the Loop vectorizer support vectorizing pointers. The attached code looks like a candidate for the SLP-vectorizer. Can you run the SLP-vectorizer with the flag -mllvm -debug-only=SLP and attach the log ? I think that we are missing the pattern for the roots of the tree. Thanks, Nadav On Oct 16, 2013, at 5:28 PM, Tom Stellard <tom at stellard.net> wrote: >

Hi, I have a simple LLVM code which loop over an array and do a simple calculation, the bytecode is here : https://gist.github.com/maattd/7322927 This code is correctly vectorized with LLVM opt 3.3, however I can't get it to vectorize with call through the API. Is there a way to debug the LLVM vectorizer pass with the API ? Is there a way to just apply the exact same optimization passes as

Hello, I apologize for this question that may has been asked a lot of times but I could not go through it. I create a multivariate time series containing NA values. I want to compute a linear regression and obtain a time serie for both residuals and fitted values. I have tried the trick ts.intersect, without success. Could you help me out of this? #### Example: y<-ts(1:10+rnorm(10))

On Sun, Jan 13, 2013 at 2:10 PM, Matthieu Monrocq <matthieu.monrocq at gmail.com> wrote: > gcc 4.5, MSVC 10, clang 3.1 > - decltype v1.0 [1] + late specified return type > - lambda v1.0 [2] > - local types as template arguments > - r-value 2.0 [3] > - static_assert > - built-in type traits This isn't very encouraging. Anecdotally from what I've seen in LLD

Hello, I did not have any response of my first mail. Perhaps I did not be enough clear: Since I have installed the latest version of samba (2.2.7), I can't see user of my domain with the Win2K User management console. I have try to change several attribute in my user ( in particular acctFlags ) but I cannot manage with making my system works. Do you have any idea ? Thanks in advance

I am writing this letter and I hope that I'll have answers about the generation of smb.conf by the way of smb.conf.master with testparm command. Like is described on page 5 of Samba3-HOWTO.pdf. The situation is: - The manual of smb.conf says that the default value of the parameter ?name resolve order? is ?lmhosts host wins bcast?, but when I type the command like: # testparm -s

Hi everybody, I was wondering if there would be interest in having a hardware IP implementing Opus (hardware as in VHDL/Verilog description)? Does that make sense? Several companies that seem to have an interest in Opus (such as Skype, Mozilla, Broadcom, Orange, Huawei) could thus have a dedicated, efficient, low-power solution for phones and tablets. As LTE is being deployed en masse, it

http://bugs.freedesktop.org/show_bug.cgi?id=25966 Summary: nv25 : rxvt scrolling is very slow Product: xorg Version: 7.5 Platform: x86 (IA32) OS/Version: Linux (All) Status: NEW Severity: normal Priority: medium Component: Driver/nouveau AssignedTo: nouveau at lists.freedesktop.org

Hello, As I am using markdown to write documentation, I am often tempted to make links to terms in definition lists. Since this is not possible for now (at least in php markdown extra), I was wondering if this could be a common addition in major implementation supporting definition lists. the syntax would be term {#id} : definition This syntax tries to follows current convention and

Hi, We installed a samba4 AD controller using Gpo for a small group of users (5 users), everything is OK. (Samba4 beta 7 on Ubuntu 12.04) We installed another Samba4 AD controller as a BDC of the first one with the command "samba-tool domain join" with succes. After checking this new installation, we saw that the sysvol share was not replicated from the PDC and all the Policies are

greened at obbligato.org (David A. Greene) writes: > Actually git pull can sometimes get you into trouble. Probably git > fetch / git rebase is the better combination for LLVM. I don't get it. Doesn't "git pull --rebase" do exactly a fetch followed by a rebase? -- Matthieu Moy http://www-verimag.imag.fr/~moy/

On Mon, May 14, 2012 at 4:39 PM, John Criswell <criswell at illinois.edu>wrote: > On 5/14/12 4:32 AM, Umesh Kalappa wrote: > > Hi All , > > Was trying to build the LLVM src from > http://llvm.org/svn/llvm-project/llvm/branches/release_30 ,But unable to > build the same and clang poped up with below error . > > > First, it sounds like you're building LLVM

Hi all, I'm trying to build passes with the PassBuilder to optimize the result of MCJIT (I assume, this is a requirement for performance). So I do this: llvm::PassBuilder passBuilder; llvm::ModulePassManager modulePassManager = passBuilder.buildPerModuleDefaultPipeline(llvm::PassBuilder::OptimizationLevel::O3); llvm::ModuleAnalysisManager moduleAnalysisManager;

Hi We are about to publish a book, which contains figures made with R plots. An important detail that we did not take into account is that the book will not be printed in 4 colors (cmyk mode), but only 2 (black +"spotcolor"). The spotcolor we use is part of the big Pantone family. The problem is that both pdf() and postscript() offer either rgb or cmyk, but no spotcolors such as

Hi, I just fixed a couple of corner-cases relating to UpdateHostkeys in git HEAD and have enabled the option by default. IMO this protocol extension is important because it allows ssh clients to automatically migrate to the best available signature algorithms available on the server and supports our goal of deprecating RSA/SHA1 in the future. We would really appreciate your feedback on this

On Sun, Oct 04, 2020 at 10:50:32PM +1100, Damien Miller wrote: > On Sun, 4 Oct 2020, Matthieu Herrb wrote: > > > On Sun, Oct 04, 2020 at 09:24:12PM +1100, Damien Miller wrote: > > > On Sun, 4 Oct 2020, Damien Miller wrote: > > > > > > > No - I think you've stumbled on a corner case I hadn't anticipated. > > > > Does your configuration