Sebastien DELDON-GNB
2012-Jul-27 16:36 UTC
[LLVMdev] ACE claims better result than LLVM for ARM 9 ?
ACE issued following PR: http://www.ace.nl/news/aces-cosy-compiler-framework-outperforms-llvm-arm9-processor Weird that they don't give any number and use ARM 9, do they mean cortex-a9 ?
Anton Korobeynikov
2012-Jul-27 17:31 UTC
[LLVMdev] ACE claims better result than LLVM for ARM 9 ?
> ACE issued following PR: > http://www.ace.nl/news/aces-cosy-compiler-framework-outperforms-llvm-arm9-processor > Weird that they don't give any number and use ARM 9, do they mean cortex-a9 ?Why don't ask them? ARM9 is either ARMv4 (ARM9) or ARMv5 (ARM9E), so, not even close to ARMv7 :) Also ask them about time machine they used - "latest LLVM" at 12 July wasn't 3.0, really. -- With best regards, Anton Korobeynikov Faculty of Mathematics and Mechanics, Saint Petersburg State University
They're also using 3.0, which is the release from December 2011. On 27 July 2012 17:36, Sebastien DELDON-GNB <sebastien.deldon at st.com> wrote:> ACE issued following PR: > > http://www.ace.nl/news/aces-cosy-compiler-framework-outperforms-llvm-arm9-processor > Weird that they don't give any number and use ARM 9, do they mean > cortex-a9 ? > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20120727/3fa10ed4/attachment.html>
Chris Lattner
2012-Jul-27 17:37 UTC
[LLVMdev] ACE claims better result than LLVM for ARM 9 ?
On Jul 27, 2012, at 9:36 AM, Sebastien DELDON-GNB wrote:> ACE issued following PR: > http://www.ace.nl/news/aces-cosy-compiler-framework-outperforms-llvm-arm9-processor > Weird that they don't give any number and use ARM 9, do they mean cortex-a9 ?It's impossible to say. This sort of marketing statement is impossible to refute, because there are no details. Who knows whether they compared against LLVM fairly with it tuning for the right architecture, had frame pointers set the same way, etc... -Chris ps, it's hard to take them seriously when they make factually incorrect statements like LLVM 3.0 being the latest, and give no way to validate their claims.
I guess we should be flattered that companies are using LLVM as the baseline for their bogus assertions. Evan On Jul 27, 2012, at 10:37 AM, Chris Lattner <clattner at apple.com> wrote:> > On Jul 27, 2012, at 9:36 AM, Sebastien DELDON-GNB wrote: > >> ACE issued following PR: >> http://www.ace.nl/news/aces-cosy-compiler-framework-outperforms-llvm-arm9-processor >> Weird that they don't give any number and use ARM 9, do they mean cortex-a9 ? > > It's impossible to say. This sort of marketing statement is impossible to refute, because there are no details. Who knows whether they compared against LLVM fairly with it tuning for the right architecture, had frame pointers set the same way, etc... > > -Chris > > ps, it's hard to take them seriously when they make factually incorrect statements like LLVM 3.0 being the latest, and give no way to validate their claims. > > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev