search for: autotune

...Current architecture complexity requires fine tuning of compiler and runtime parameters to achieve full potential performance. Autotuning substantially improves default parameters in many scenarios but it is a costly process requiring a long iterative evaluation. We propose an automatic piecewise autotuner based on CERE (Codelet Extractor and REplayer). CERE decomposes applications into small pieces called codelets: each codelet maps to a loop or to an OpenMP parallel region and can be replayed as a standalone program. Codelet autotuning achieves better speedups at a lower tuning cost. By grouping c...

...plexity requires fine tuning of > compiler > and runtime parameters to achieve full potential performance. Autotuning > substantially improves default parameters in many scenarios > but it is a costly process requiring a long iterative evaluation. > We propose an automatic piecewise autotuner based on CERE (Codelet > Extractor and REplayer). CERE decomposes applications into small > pieces called codelets: each codelet maps to a loop or to an OpenMP > parallel region and can be replayed as a standalone program. > Codelet autotuning achieves better speedups at a lower tuning...

Hello everybody. Last monday I sent an email to the LLVM dev mailing list saying that I was looking forward to working on these GSoC projects: *Advanced Heuristics for Ordering Compiler Optimization Passes* and *Machine learning and compiler optimizations: using inter-procedural analysis to select optimization* I currently do an undergraduate research on compiler autotuning of Rust code, more

Awesome, thanks! As per your suggestion, I read the description of these two projects: Advanced Heuristics for Ordering Compiler Optimization Passes Machine learning and compiler optimizations: using inter-procedural analysis to select optimizations and they are amazing! Indeed, they are very close to my interest in autotuning. I didn't see them on the list before. If I choose to focus on

Andre, Are you going to have time to MFC things from -current for auto-tuning -stable before 9.2? I fear (maybe unnecessarily?) that we are about to ship yet another release that can't do basic 10gigE when sufficient memory exists. If you don't have time, then let me know and I'll see what I can do. -- Alfred Perlstein VP Software Engineering, iXsystems

Hello! My name is Emanuel and I am an undergraduate student from Brazil (at the University of São Paulo) wanting to participate in this years GSoC on LLVM. Specifically, on the "Improve parallelism-aware analyses and optimizations" project. I currently do research on autotuning of LLVM IR optimization passes and I am sitting for a class about parallel computing, but I have been studying

Hi, I am planning to begin a project to explore the space of tuning LLVM internals in an effort to increase performance. I am wondering if anyone can point to me any parameterizations, heuristics, or priorities functions within LLVM that can be tuned/adjusted. So far, I'm considering BranchProbabilityInfo and InlineCost. Does anyone have any other suggestions? Thanks, Robert

..., Oct 14, 2017 at 10:58:17PM -0500, Kavon Farvardin via llvm-dev wrote: > > something simpler will do, IMHO. Happy to discuss this further if > > folks are in California next week :) > > Yes, I'll be in California next week, let's chat! > > We could make use of the autotuner I'm currently building: > > https://github.com/kavon/autotune > > It tries to find an optimal sequence of IR passes, and I've run into a bunch of different bugs with it so far (e.g., it seems structurizecfg interacts poorly with 'invoke'). I'll start submitting bu...

On Sat, Oct 14, 2017 at 11:05 AM, John Regehr via llvm-dev <llvm-dev at lists.llvm.org> wrote: > These are definitely LLVM bugs. It would be best to report reduced test > cases against top of tree. > > We should have some automated infrastructure for finding these too... > > John > Zhendong & friends generally do that (and reported many bugs :) I tried that myself,

Hi, I'm trying to autotune a good sequence of IR optimization passes and I seem to run into segfaults in opt (in LLVM5) with certain pass orderings. Is this expected behavior? If so, what would be the recommended way of determining pass dependencies so that I can encode them into the tuner? The test program can be found...

http://bugzilla.mindrot.org/show_bug.cgi?id=1332 Summary: Add ability to set TCP rcvbuf size Product: Portable OpenSSH Version: 4.6p1 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: ssh AssignedTo: bitbucket at mindrot.org ReportedBy: imorgan at

Hello list, right now SSH Tunnel / scp is reaches just around 76Mb/s on my E5 Xeon using AES-NI but openssl reaches around 600-700Mb/s using 128aes-cbc cipher. As far as i understand http://www.psc.edu/index.php/hpn-ssh this is due to very small buffers in ssh / scp. Is there any work on this? Like autotuning the buffer size? Are there plans to integrate the hpn patches? Greets, Stefan

Hello again! Previously, I introduced myself as a prospective GSoC student interested in LLVM. Per Mr. Doerfert suggestion I took a look at two other projects that he said could interest me due to my ongoing research in autotuning. The projects are: Advanced Heuristics for Ordering Compiler Optimization Passes Machine learning and compiler optimizations: using inter-procedural analysis to select

...loreenhancementsto LLVM enabling better modeling and exploration of the complex space of potential optimizations in addition toenhancingLLVM's optimization capabilities. One focus of this work will be to enable more-effective autotuning by strengthening the two-way communication between the autotuner and the optimizer. A passion forimprovingcompiler technology, a keen interest in high-performance computing hardware, anda desireto advance scientific computing, will be rewarded in this position. Position Requirements We expect you to have: * Minimum education and experience or equivale...

Le 07/01/2016 12:22, Reindl Harald a écrit : > > /usr/sbin/ifconfig eth0 txqueuelen 100 > ______________________________________________ > > ifcfg-eth0: > > ETHTOOL_OPTS="-K ${DEVICE} tso on lro off; -G ${DEVICE} rx 128 tx 128" > ______________________________________________ > > sysctl.conf: > > net.core.rmem_max = 65536 > net.core.wmem_max = 65536

From: Eric Dumazet <eric.dumazet at gmail.com> Date: Mon, 28 Oct 2013 16:19:49 -0700 > On Mon, 2013-10-28 at 15:44 -0700, Michael Dalton wrote: >> The virtio_net driver's mergeable receive buffer allocator >> uses 4KB packet buffers. For MTU-sized traffic, SKB truesize >> is > 4KB but only ~1500 bytes of the buffer is used to store >> packet data, reducing

From: Eric Dumazet <eric.dumazet at gmail.com> Date: Mon, 28 Oct 2013 16:19:49 -0700 > On Mon, 2013-10-28 at 15:44 -0700, Michael Dalton wrote: >> The virtio_net driver's mergeable receive buffer allocator >> uses 4KB packet buffers. For MTU-sized traffic, SKB truesize >> is > 4KB but only ~1500 bytes of the buffer is used to store >> packet data, reducing

I have been experimenting with dirhash settings, and have scoured the internet for other peoples' experience with it. (I found the performance improvement in compiling has forestalled the need to add an SSD drive. ;-) I believe that increasing the following values by 10 would benefit most FreeBSD users without disadvantage. vfs.ufs.dirhash_maxmem: 2097152 to 20971520

Dear rsync folks, I'd like to request/suggest that cli options to set TCP send/receive buffers be added to rsync client-side. Summary: I'm aware that a daemon's config-file can set socket options for the server side (e.g. SO_SNDBUF, SO_RCVBUF). That is useful. But when trying to get high-throughput rsync over long paths (i.e. large bandwidth*delay product), since

...7 gives this problem. Opening e.g. jpegs and regular plain text files with Irfanview and Notepad++ respectively is just fine. Pre-CentOS 6.6 opening any file on the samba share from Windows was lightning-fast. I'm not too sure where to start trouble-shooting. I've tried disabling the TCP autotune on Windows for starters, which a small improvement, but what next? Hints would be appreciated. Thanks. -- BW, Sorin ----------------------------------------------------------- # Sorin Srbu, Sysadmin # Uppsala University # Dept of Medicinal Chemistry # Div of Org Pharm Chem # Box 574 # SE-75123...