search for: tsoft

After I manually modified .ll file deleting %0 I am getting this error from llvm-as: lvm-as: my.ll:43:3: error: instruction expected to be numbered '%0' %1 = tail call noalias i8* @malloc(i32 8) nounwind ; <i8*> [#uses=1] ^ So now I have to also rename all other %<NUM> variables? Why llvm is so strict with naming? I think it should just accept whatever names are there

Török Edwin wrote: > Are you using 2.6 or 2.7, 32-bit or 64-bit? > I use 2.7 on i386. lli has debug asserts enabled, but I guess this shouldn't matter for JIT code speed. jit: 11.32 real exe: 7.64 user Both have -O3 option. Speed should be the same. Yuri

On Wed, Apr 28, 2010 at 10:48 AM, Yuri <yuri at tsoft.com> wrote: > Jeffrey Yasskin wrote: >> >> We currently use two different techniques to get whole libraries >> included into the .so. On Linux (gnu ld and gold), we pass >> --whole-archive, while on OSX we pass -all_load. Which ld does FreeBSD >> use? If you can...

Jeffrey Yasskin wrote: > We currently use two different techniques to get whole libraries > included into the .so. On Linux (gnu ld and gold), we pass > --whole-archive, while on OSX we pass -all_load. Which ld does FreeBSD > use? If you can rearrange tools/llvm-shlib/Makefile until it works for > you and send us a patch, we can probably get it working for llvm-2.8. > ld on

I am new to LLVM and couldn't find any llvm-user list, so I am posting my user question here, sorry. I am trying to create a simple "puts" call accepting the static string, with the code below. The last line (CallInst::Create) fails with an assert: "Calling a function with a bad signature!" Because the type of function is void(u8*) and the argument supplied is:

On Thu, Apr 29, 2010 at 6:31 PM, Jeffrey Yasskin <jyasskin at google.com> wrote: > On Wed, Apr 28, 2010 at 10:48 AM, Yuri <yuri at tsoft.com> wrote: >> Jeffrey Yasskin wrote: >>> >>> We currently use two different techniques to get whole libraries >>> included into the .so. On Linux (gnu ld and gold), we pass >>> --whole-archive, while on OSX we pass -all_load. Which ld does FreeBSD >&...

On Wed, Apr 28, 2010 at 9:23 AM, Yuri <yuri at tsoft.com> wrote: > Pekka Jääskeläinen wrote: >> >> On 04/28/2010 11:42 AM, Yuri wrote: >>> >>> Looks like this is platform dependent. >> >> Yep, works fine here on Debian 5.0/x86_64 with gcc 4.3.2. >> >> Did you build with make REQUIRES_RTTI=1...

On 04/28/2010 11:42 AM, Yuri wrote: > Looks like this is platform dependent. Yep, works fine here on Debian 5.0/x86_64 with gcc 4.3.2. Did you build with make REQUIRES_RTTI=1 which is required as of LLVM 2.7 to get RTTI for LLVM symbols? -- Pekka

On Wed, Apr 28, 2010 at 1:42 AM, Yuri <yuri at tsoft.com> wrote: > Jeffrey Yasskin wrote: >> >> I just tried it again on trunk (not 2.7) on OSX 10.5, and it works. >> What platform are you on? Does it work with another version of gcc? >> > > Looks like this is platform dependent. libLLVM-2.7.so created has size onl...

Yuri <yuri at tsoft.com> writes: >> With >> >> time lli -O3 fib.bc 45 >> >> you are measuring the time lli takes optimizing the LLVM code, >> generating the native code and, finally, executing it. If you add to >> this the debug asserts, it is not surprising that lli ends...

Hi If I am linking a large library with my app, is there a way I can eliminate all the unused functions used in the library as part of the linking ? thanks shrey

shreyas krishnan wrote: > Hi > If I am linking a large library with my app, is there a way I can > eliminate all the unused functions used in the library as part of the > linking ? > If you are talking about static libraries (.a), this is done automatically by linker. You shouldn't have to worry about this. If you are talking about the shared library (.so) -- it will be

Pekka Jääskeläinen wrote: > On 04/28/2010 11:42 AM, Yuri wrote: >> Looks like this is platform dependent. > > Yep, works fine here on Debian 5.0/x86_64 with gcc 4.3.2. > > Did you build with make REQUIRES_RTTI=1 which is required > as of LLVM 2.7 to get RTTI for LLVM symbols? > No. If this is a hack to get the command 'g++ -shared -o libLLVM-2.7.so -lXXX -lXXX

Jeffrey Yasskin wrote: > I just tried it again on trunk (not 2.7) on OSX 10.5, and it works. > What platform are you on? Does it work with another version of gcc? > Looks like this is platform dependent. libLLVM-2.7.so created has size only 4145 bytes and all llvm symbols are missing. It's empty because there are no .o modules included into it, only libraries through -l options

...such line in each group: # many similar lines deleted Nov 6 12:08:32 wheat fetchnews[13617]: comp.std.c++: considering articles 177500 - 177504 Nov 6 12:08:34 wheat fetchnews[13617]: comp.std.c++: 5 articles fetched (to 6683), 0 killed Nov 6 12:08:34 wheat fetchnews[13617]: clamping maxage for tsoft.general to global expire 50 Nov 6 12:08:34 wheat fetchnews[13617]: clamping maxage for comp.parallel to global expire 50 Nov 6 12:08:34 wheat fetchnews[13617]: comp.parallel: no new articles Nov 6 12:08:34 wheat fetchnews[13617]: clamping maxage for news.announce.newusers to global expire 50 No...

...e-shared was _created_ in 2.7. If you were passing it in 2.6 or earlier, it wasn't doing what you expected. I just tried it again on trunk (not 2.7) on OSX 10.5, and it works. What platform are you on? Does it work with another version of gcc? On Tue, Apr 27, 2010 at 4:06 PM, Yuri <yuri at tsoft.com> wrote: > > When I enabled this option I am getting errors, see below. > Is there any way to fix this? There should be an option to have most of > the code in shared library. > > Yuri > > --- error log using gcc-4.5.0 on 4 cpus with configure options: > --enable-as...

When I enabled this option I am getting errors, see below. Is there any way to fix this? There should be an option to have most of the code in shared library. Yuri --- error log using gcc-4.5.0 on 4 cpus with configure options: --enable-assertions --enable-optimized --enable-shared --prefix=/usr/local/llvm --- llvm[2]: Compiling llc.cpp for Release build gmake[2]: Entering directory

Jeffrey Yasskin wrote: >> Would you try the patch at >> http://codereview.appspot.com/download/issue968046_1.diff? It should >> make the BSDs fall into the same path as Linux, and since you use gnu >> ld, that should work for you. >> > > Ping? > Sorry for the delay. Here is what I am getting after applying the patch: gmake[2]: Leaving directory

On Tue, May 4, 2010 at 5:31 PM, Yuri <yuri at tsoft.com> wrote: > Jeffrey Yasskin wrote: >>> >>> Would you try the patch at >>> http://codereview.appspot.com/download/issue968046_1.diff? It should >>> make the BSDs fall into the same path as Linux, and since you use gnu >>> ld, that should work for...

I am trying to build llvm on freebsd-8.0 on sparc64 CPU. I patched sources because 'sparc' enum value in include/llvm/ADT/Triple.h conflicts with gcc typedef 'sparc'. This allowed llvm itself to build successfully. But there is an error during gcc frontend compile: Did not get a target machine! Triplet is sparc64-unknown-freebsd8.0 I think this is because in many places, like in