similar to: [LLVMdev] LTO, plugins and binary sizes

> The full patches I used are attached. I hope to get the non-hackish > bits reviewed, starting by the fix to 8313. attached now. Cheers, -- Rafael Ávila de Espíndola -------------- next part -------------- A non-text attachment was scrubbed... Name: llvm.patch Type: text/x-patch Size: 25795 bytes Desc: not available URL:

On Oct 9, 2010, at 10:51 AM, Rafael Espindola wrote: >> The full patches I used are attached. I hope to get the non-hackish >> bits reviewed, starting by the fix to 8313. > > attached now. Wow, nice results. Please split the patches out according to what they do. Is there any impact on clang compile time by building with LTO? What about build times? :) -Chris

Hi Rafael, > I did the testing on linux x86-64. Gcc is the 4.4.4 included with > Fedora 13. The results are: > > gcc -O3: 32MB > gcc -Os: 25MB > clang lto -Os: 22MB how does clang -Os without lto do? Also, what about gcc-4.5 with -flto? Ciao, Duncan.

> how does clang -Os without lto do? Not sure. Should be able to try it next Tuesday. >  Also, what about gcc-4.5 with -flto? Last time I tried gcc could not build itself with LTO, so I am not all that excited to try it :-) > Ciao, > > Duncan. Cheers, -- Rafael Ávila de Espíndola

I have gold from cvs, and llvm-gcc/llvm from svn. My platform is Ubuntu 9.04 x86_64. I've configured llvm-gcc with "--enable-multilib". When compiling the LTO example from http://llvm.org/docs/LinkTimeOptimization.html I have no problems with 64-bit compilation, but get the following with -m32. luked at node:~/tests/lto$ make USE32=1 llvm-gcc -m32 -c -o main.o main.c llvm-gcc

Hello, I down loaded "llvm-gcc4.2-2.7-x86_64-linux" and I built Spec2006 with it. It is great and except for one benchmark the rest work fine. I want to build them with LTO now. I followed the directions in " http://llvm.org/docs/GoldPlugin.html" and built ar, nm-new, and ld-new in binutils. I also built libLLVMgold.so. I get the following error: llvm-gcc: -use-gold-plugin,

On 30 August 2010 12:43, Luke Dalessandro <luked at cs.rochester.edu> wrote: > I have gold from cvs, and llvm-gcc/llvm from svn. My platform is > Ubuntu 9.04 x86_64. I've configured llvm-gcc with "--enable-multilib". > > When compiling the LTO example from > http://llvm.org/docs/LinkTimeOptimization.html I have no problems with > 64-bit compilation, but get

On 21 July 2010 22:14, Reza Yazdani <ry.yazdani at gmail.com> wrote: > Hello, > > I down loaded "llvm-gcc4.2-2.7-x86_64-linux" and I built Spec2006 with it. > It is great and except for one benchmark the rest work fine. I want to build > them with LTO now. I followed the directions in > "http://llvm.org/docs/GoldPlugin.html" and built ar, nm-new, and

I run the tests again. This time with only the more interesting candidates and with the constructor alias optimization enabled. The gcc used is the one with fedora 13 (4.4.4) The size of the final clang binary in (KiB) when compiled on linux x86-64 was gcc -Os 23073 clang -Os 25251 clang -Os lto 18295 Time to "link" the final clang binary with lto when the gold plugin was

Hello, With GCC, it is possible to compile GIMPLE from an object file back to assembly. With Dragonegg, it seems to be a NOP. (See below for a comparison.) This appears it could be a nice way to mix the strengths of GCC and LLVM. If it worked.. Should it? Thanks, Marcus [mdaniels at dn002 dragonegg]$ cat hello.c #include <stdio.h> int hello () { printf ("Hello

Hola a todos. Soy un poco torpe manejando cadenas de texto, así que os pido ayuda. Tengo un vector de texto de este tipo datos$tipo [1] m.1.p.Álava m.1.p.Albacete [3] m.2.p.Alicante m.1.p.Almería [5] m.3.p.Asturias m.1.p.Ávila [7] m.1.p.Badajoz m.1.p.Baleares (Illes) [9] m.1.p.Barcelona m.1.p.Burgos [11] m.1.p.Cáceres m.1.p.Cádiz Y quiero extraer el

Is there a canonical way to check if all elements of a vector or matrix are the same? Solutions below work, but look hackish to me. > x <- rep(1, 10) > all(x == x[1]) # == operator does not provide for small differences [1] TRUE > isTRUE(all.equal(x, rep(x[1], length(x)))) # ugly [1] TRUE Best, Vincent -- Vincent Goulet, Associate Professor ??cole d'actuariat

On 11-06-02 6:56 AM, Anton Korobeynikov wrote: > Hi Chip, > >> Because of all this, it's hard to reconstruct the SEH information from >> the MachineMove array. I have thought about adding a new array specific >> to SEH information, but I'm not sure how you guys would feel about that. >> Any ideas on how to solve this problem? > Same problem with

Hi Everyone, Chris suggested[1] I should ask for feedback as to whether this is a desired feature before I put too much effort into it, so here goes: I would like to be able to export a symbol that is inside an LLVM structure. This is possible on ELF targets[2], and the attached proof- of-concept patch to AsmWriter makes it work (although in a hackish way that I am NOT suggesting be

Hello, I have a few models -- book, cd, dvd -- for which I''d like to have an abstract base superclass to hold some common stuff. That abstract class, I was thinking, would inherit from ActiveRecord. Didn''t work, though, and looking around, I found this: <http://wiki.rubyonrails.com/rails/pages/HowtoMakeAbstractModel>

All, This is probably a bit more of a mac question, but here goes. (Admission: I posted this to ruby-talk as well....) I am trying to use rdoc -d to generate class diagrams on a mac running Tiger (OS X 10.4) for a rails app. I managed to get dot installed, but now am running into problems with dot looking for libraries that either do not exist or are apparently in the wrong place. I probably made

On Jun 2, 2011, at 6:20 AM, Rafael Ávila de Espíndola wrote: > On 11-06-02 6:56 AM, Anton Korobeynikov wrote: >> Hi Chip, >> >>> Because of all this, it's hard to reconstruct the SEH information from >>> the MachineMove array. I have thought about adding a new array specific >>> to SEH information, but I'm not sure how you guys would feel about

Hi. We have some interest in using samba as a file server on a small device, but we have some problems with low amount of space on our target machine, the question is how low can we go? So far we have just made a quick samba packages that requires 22mb of space, however since we are not experts on samba, it could be us that are stupid :) What we want to have is a: Samba server that is

I've attached a less-hackish implementation of this. This includes the following modifications: - getSupportsOverlappingAliases() method on TargetMachine which returns whether the target supports multiple symbols to the same object. This returns false in the superclass and needs to be explicitly overridden for each target to enable it. - An implementation of this method in

2011/10/21 Rafael Ávila de Espíndola <rafael.espindola at gmail.com>: >> Could be, echristo, bigcheese, would this be correct for Mach-O and COFF? > > bigcheese noted on IRC that the test crashes the COFF emitter. For some > reason I am always getting > >        movl    $256, %eax              ## imm = 0x100 >        calll   *%eax > > on darwin already IIRC, we