search for: weigert

...#39;t measure the performance but i would assume that the boost developers had a good reason for using assembler in this context. will llvm ever support inline assembly? is there anybody who is working on that? thanks, stefan. On Tuesday 21 July 2009 01:54:20 pm Vladimir Prus wrote: > Stefan Weigert wrote: > > hi, > > > > when using the execution engine (no matter, if JIT or Interpreter) i get > > the following assertion as soon as i use boost::shared_ptr: > > > > /build/buildd/llvm-2.5/lib/Target/X86/X86CodeEmitter.cpp:522: > > void<unnamed>::Em...

hi, when using the execution engine (no matter, if JIT or Interpreter) i get the following assertion as soon as i use boost::shared_ptr: /build/buildd/llvm-2.5/lib/Target/X86/X86CodeEmitter.cpp:522: void<unnamed>::Emitter::emitInstruction(const llvm::MachineInstr&, const llvm::TargetInstrDesc*): Assertion `0 && "JIT does not support inline asm!\n"' failed.

Stefan Weigert wrote: > hi, > > when using the execution engine (no matter, if JIT or Interpreter) i get the > following assertion as soon as i use boost::shared_ptr: > > /build/buildd/llvm-2.5/lib/Target/X86/X86CodeEmitter.cpp:522: > void<unnamed>::Emitter::emitInstruction(const llv...

Hi Stefan, On Thu, 2007-03-01 at 23:26 +0100, Stefan Weigert wrote: > hello, > > i am new to llvm and can't manage to compile the gcc backend. Okay. Did you read: http://llvm.org/docs/GettingStartedGuide.html ? > > my steps: > i compile the llvm-source first as suggested in "README.LLVM" > that works fine and without an...

hello, i am new to llvm and can't manage to compile the gcc backend. my steps: i compile the llvm-source first as suggested in "README.LLVM" that works fine and without any issues. then i make a new dir called build. from there i do the following: stefan at ubuntu:~/programs/llvm/build$ ../llvm-gcc4-1.9.source/configure --enable-languages=c,c++ --prefix=/usr/local/

hello mailinglist, i currently try to compile ffmpeg on a 64bit fedora-core 8 machine using llvm2.3. the "good" things first: everything works fine on a 32bit ubuntu-machine using llvm-2.3 (actually i do not know if that is grace to ubuntu or grace to 32bit) when using the normal gcc on the 64bit fc8 machine, everything works too. now the "bad" thing: it does not work on fc8

Hello, have a short look at the following simple c-prog: #include <stdlib.h> int main(int argc, char* argv[]) { void* buf = malloc(10 * sizeof(char*)); /* do sth with buf, so that it is not "optimized away" */ return 0; } When you compile this using a plain llvm-gcc, the call to libc-malloc is left inside.But compiling it with -O2 alters the call to llvm::MallocInst. Now

hello, after a lot of tries i could assemble a pretty small version of what either triggers a bug in LLVM or is caused by misuse of LLVM on my side. i would be very happy, if someone with more experience with the JIT could have a look on my code. the use-case: a class in a bitcode-file, loaded (jitted) by an app at runtime the app calls a static create inside the bitcode and then tries to

dear community, sorry for posting again a beginner's question but i can't find a solution on my own. i installed all llvm and llvm-gcc4 in a directory called "~/programs/llvm/install" in my home folder. i can build a test c - file as follows: ~/programs/llvm/install/bin/gcc -v -o test test.c Using built-in specifications. Target: i686-pc-linux-gnu Configured with: