search for: speedups

A couple more data points. Current llvm 3.7svn with the two outstanding OPENMP patches can build the openmp support in gdl 0.9.5 (which completely passes its test suite) and apbs 1.4.1's limited openmp support. On Sat, May 2, 2015 at 11:11 PM, Jack Howarth < howarth.mailing.lists at gmail.com> wrote: > On a positive note, current llvm 3.7svn with the two outstanding > OPENMP

TLDR - The proposal below is intended to allow inlining of larger callees when such inlining is expected to reduce the dynamic instructions count. Proposal ------------- LLVM inlines a function if the size growth (in the given context) is less than a threshold. The threshold is increased based on certain characteristics of the called function (inline keyword and the fraction of vector

Just nitpicking: 1) DI(F) should include a component that estimate the epi/prologue cost (frameSetupCost) which InlinedDF does not have 2) The speedup should include callsite cost associated with 'C' (call instr, argument passing): Speedup(F,C) = (DI(F) + CallCost(C) - InlinedDF(F,C))/DI(F). Otherwise the proposal looks reasonable to me. David On Thu, Jul 30, 2015 at 2:25 PM,

On a dry run, rsync displays a speedup value calculated from the total size of the source file data and the amount of data sent over the connection, but this value is meaningless and grossly misleading because the file data is not sent over the connection. Example: [matt@mattlaptop2 test]$ rsync -avi -n ~/eclipse/releases/eclipse-SDK-3.3-linux-gtk.tar.gz . sending incremental file list

Hi Jack, thanks for splitting out what the effects of LLVM's / GCC's vectorizers is. On 01/06/13 21:34, Jack Howarth wrote: > On Sat, Jun 01, 2013 at 06:45:48AM +0200, Duncan Sands wrote: >> >> These results are very disappointing, I was hoping to see a big improvement >> somewhere instead of no real improvement anywhere (except for gas_dyn) or a >> regression

You mean the input test data? I was testing performance using our offline perf suite (which is a ton of out of tree code), so it’s not something I can share, but I imagine any similar test suite put through a typical compiler pipeline will exercise ADCE in similar ways. ADCE’s cost is pretty much the sum of (cost of managing the set) + (cost of eraseinstruction), which in our case turns out to be

Jack, Can you please file a bug report and attach the BC files for the major loops that we miss ? Thanks, Nadav On Jun 2, 2013, at 1:27, Duncan Sands <duncan.sands at gmail.com> wrote: > Hi Jack, thanks for splitting out what the effects of LLVM's / GCC's vectorizers > is. > > On 01/06/13 21:34, Jack Howarth wrote: >> On Sat, Jun 01, 2013 at 06:45:48AM +0200,

...39;clang -O3' against 'clang -O3 -mllvm -vectorize'? Yes. [I've tested the current patch directly using opt -vectorize -unroll-allow-partial; for running the test suite I recompiled llvm/clang to hardcode the options as I wanted them]. > > > The largest three performance speedups are: > > SingleSource/Benchmarks/BenchmarkGame/puzzle - 59.2% speedup > > SingleSource/UnitTests/Vector/multiplies - 57.7% speedup > > SingleSource/Benchmarks/Misc/flops-7 - 50.75% speedup > > > > The largest three performance slowdowns are: > > MultiSource/Bench...

Unless you suppress all output with the -q option, rsync will always print an initial "building/receiving file list ... done" line. In my opinion, this is a bit superfluous. When I want to see a progress indication, I can use the --progress option. Another issue is the 3-line transfer speed and speedup factor report at the end. So every rsync invocation produces at least four lines of

...w much of the runtime overhead is due to the unrolling (produces more code that needs to be optimized) and which part is due to vectorization. The same counts for the speedup. How much is caused by unrolling and how much is actually caused by your pass. >>> The largest three performance speedups are: >>> SingleSource/Benchmarks/BenchmarkGame/puzzle - 59.2% speedup >>> SingleSource/UnitTests/Vector/multiplies - 57.7% speedup >>> SingleSource/Benchmarks/Misc/flops-7 - 50.75% speedup >>> >>> The largest three performance slowdowns are: >>>...

Hallo short question how is the Syntax for interactivity.patch ext3-dir-speedup.patch ? patch -p0 ext3-dir-speedup.patch doesnt work -- Frank

On May 14, 2010, at 11:24 AM, Jan Voung wrote: > I'm sure this has been asked many times, but is there current work on decreasing the time taken by the DAG-based instruction selector, or the other phases of llc? I am just beginning to dive into LLVM, and I am interested in compile-time reductions that do not reduce code quality dramatically. For example, simply switching on

On May 18, 2010, at 12:07 PM, Jan Voung wrote: > Here are some recent stats of the fast vs local vs linear scan at O0 on "opt -std-compile-opts" processed bitcode files. The fast regalloc is still certainly faster at codegen than local with such bitcode files. Let me know if the link doesn't work: > >

I'm syncing from a USB disk to my hard disk like this: rsync -vr /path/to/usb/disk/dir/ /path/to/hard/disk/dir/ But the speedup is always 0.99 which I think means it is just copying the files each time instead of syncing them. What could be wrong? - Grant

Actually this kind of opportunities, as outlined bellow, was one of my contrived motivating example for fast-math. But last year we don't see such opportunities in real applications we care about. t1 = x1/y ... t2 = x2/y. I think it is better to be taken care by GVN/PRE -- blindly convert x/y => x *1/y is not necessarily beneficial. Or maybe we can blindly perform such

I am trying to copy an 8 gig file. After first copying it over I try and rsync it again. There are no changes so why is it copying the file over again? Here is the output of what I am trying to do. And what does speedup mean? [root@VMWARE Storage]# rsync --verbose --recursive -t --stats -- progress ntapps.dsk vmware2::home/ [root@VMWARE Storage]# rsync --verbose --recursive -t --stats --

I did something similar for dominators, for GVN, etc. All see significant speedups. However, the answer i got back when i mentioned this was "things like ptrset and densemap should only have a small performance difference from side data when used and sized right", and i've found this to mostly be true after looking harder. In the case you are looking at, i see: -...

...ch will probably > work for you. Hey Hal, those are great news. Especially as the numbers seem to show that vectorization has a significant performance impact. What did you compare exactly. 'clang -O3' against 'clang -O3 -mllvm -vectorize'? > The largest three performance speedups are: > SingleSource/Benchmarks/BenchmarkGame/puzzle - 59.2% speedup > SingleSource/UnitTests/Vector/multiplies - 57.7% speedup > SingleSource/Benchmarks/Misc/flops-7 - 50.75% speedup > > The largest three performance slowdowns are: > MultiSource/Benchmarks/MiBench/security-rijndae...

On May 17, 2010, at 9:09 PM, Rafael Espindola wrote: >> The fast and local register allocators are meant to be used on unoptimized code, a 'Debug build'. While they do work on optimized code, they do not give good results. Their primary goal is compile time, not code quality. > > Yes, we have a somewhat uncommon use case. It is fine to spend time > optimizing bitcode (LTO

Hi, the combination of --keep-dirlinks and --delete (--delete-during, --delete-delay) erroneously deletes symlinks to directories on the receiver. With --delete-before and --delete-after it works as expected. (rsync version 3.0.3 protocol version, on Debian Lenny) I don't think that this behaviour is intended. Details/example: - Host 1 directory-structure: drwx------ ./1