The current implementation in LLVM is experimental at best. It probably should be ripped out. Evan On Oct 30, 2013, at 2:10 PM, mcrosier at codeaurora.org wrote:> I know nothing about this pass, but here are a few comments. > >> Does anyone know if the shrink wrap pass works for ARM architecture? > > First, how are you invoking it? > > Looks to be a command line option, so -mllvm -shrink-wrap should enable > it. Assuming you have a Debug build, you can see the debug output using > -debug-only=shrink-wrap. > > >> I tried it seems not working. Which architecture shrink warp was >> originally designed for? > > AFAICT, the implementation is target independent. > > Chad > > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
On 30 October 2013 20:08, Evan Cheng <evan.cheng at apple.com> wrote:> The current implementation in LLVM is experimental at best. It probably should be ripped out.Patch attached :-) Cheers, Rafael -------------- next part -------------- A non-text attachment was scrubbed... Name: t.patch Type: text/x-patch Size: 57169 bytes Desc: not available URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20131031/1d0dc79a/attachment.bin>
On 30 October 2013 22:34, Rafael Espíndola <rafael.espindola at gmail.com>wrote:> On 30 October 2013 20:08, Evan Cheng <evan.cheng at apple.com> wrote: > > The current implementation in LLVM is experimental at best. It probably > should be ripped out. > > Patch attached :-) >Thanks Rafael! It's a shame to see a feature go away that easily, but I also don't have man-power to maintain this. At least it was well isolated so that anyone can start over from that when needed. For my education, how bad was it? Was the problem in the implementation or in the edge case performance? Was it something wrong with the code or just under-tested? Maybe, if it's not too bad (just need more testing), than I could possibly convince someone to look at it? cheers, --renato -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20131030/18cc9313/attachment.html>
LGTM. Was that the last use of SparseBitVector? Thanks, /jakob> On 30 Oct 2013, at 22:34, Rafael Espíndola <rafael.espindola at gmail.com> wrote: > >> On 30 October 2013 20:08, Evan Cheng <evan.cheng at apple.com> wrote: >> The current implementation in LLVM is experimental at best. It probably should be ripped out. > > Patch attached :-) > > Cheers, > Rafael > <t.patch> > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev