Silvio Ricardo Cordeiro
2010-Jun-03 02:24 UTC
[LLVMdev] Using benchmarks from SPEC (gcc, bzip2...)
I'm planning on implementing a register allocator on LLVM and I've noticed that the LLVM test suite supports SPEC's CINT2006 benchmarks with some makefiles. Since SPEC seems to be way too expensive, but some (all?) of its programs are available for free, I was thinking of making my own SPEC-based test suite. How hard would it be for me to integrate whole-program benchmarks in LLVM? Thanks in advance. -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20100602/033880f8/attachment.html>
On Wed, Jun 2, 2010 at 7:24 PM, Silvio Ricardo Cordeiro <silvioricardoc at gmail.com> wrote:> I'm planning on implementing a register allocator on LLVM and I've noticed > that the LLVM test suite supports SPEC's CINT2006 benchmarks with some > makefiles. Since SPEC seems to be way too expensive, but some (all?) of its > programs are available for free, I was thinking of making my own SPEC-based > test suite. > How hard would it be for me to integrate whole-program benchmarks in LLVM? > Thanks in advance.LLVM already has a whole-program benchmarks; see http://llvm.org/docs/TestingGuide.html#testsuiterun . -Eli
Maybe Matching Threads
- [LLVMdev] NumLoads/NumStores for linearscan?
- [LLVMdev] NumLoads/NumStores for linearscan?
- [LLVMdev] Register allocation marking spills (Re: NumLoads/NumStores for linearscan?)
- [LLVMdev] [llvm-testresults] bwilson__llvm-gcc_PROD__i386 nightly tester results
- [LLVMdev] NumLoads/NumStores for linearscan?