Peng, Since you're on the IO-500 mailing list, I would think you're aware of the IO500...? But, the IO500 is still probably the best answer to your question. It specifies the file system used as part of each entry. https://www.vi4io.org/io500/start?fields=information__system,information_... Regards, Patrick ________________________________ From: IO-500 <io-500-bounces(a)vi4io.org&gt; on behalf of Peng Yu via IO-500 <io-500(a)vi4io.org&gt; Sent: Tuesday, September 17, 2019 9:31 AM To: io-500(a)vi4io.org <io-500(a)vi4io.org&gt; Subject: [IO-500] A list file systems used in top I/O system Hi, I'd like to see a list of file system used in top I/O systems. But top500.org is mainly for computation not for I/O. Is there a website ranking by I/O? Thanks. -- Regards, Peng _______________________________________________ IO-500 mailing list IO-500(a)vi4io.org https://www.vi4io.org/mailman/listinfo/io-500

Also, why there is only 100 entries? Should there be 500 entries? Thanks. On Tue, Sep 17, 2019 at 10:50 AM Peng Yu <pengyu.ut(a)gmail.com&gt; wrote: -- Regards, Peng

Hey Peng, Thanks for your interest in IO500 and we're excited to have you be part of the community. If you do any analysis of the list, please do share it back with us as we're always interested in things like that. Thanks, John On Tue, Sep 17, 2019 at 10:16 AM Julian Kunkel via IO-500 <io-500(a)vi4io.org&gt; wrote:

Peng, You might find the graph on slide 7 here interesting: https://hps.vi4io.org/_media/events/2019/hpc-iodc-lustre_next_20_years-di... This gives a very rough overview, from public sources, of the scaling of large supercomputer file systems over time. It is not very detailed, by necessity, but it is interesting as an overview. Regards, Patrick ________________________________ From: IO-500 <io-500-bounces(a)vi4io.org&gt; on behalf of Patrick Farrell via IO-500 <io-500(a)vi4io.org&gt; Sent: Tuesday, September 17, 2019 11:22 AM To: Julian Kunkel <juliankunkel(a)googlemail.com&gt;; Peng Yu <pengyu.ut(a)gmail.com&gt; Cc: io-500(a)vi4io.org <io-500(a)vi4io.org&gt; Subject: Re: [IO-500] A list file systems used in top I/O system It's also worth noting that the TOP500 list itself is far from a comprehensive compendium of the fastest machines. Many large government and industry owned machines do not participate in the list, for various reasons. Regards, Patrick ________________________________ From: Julian Kunkel <juliankunkel(a)googlemail.com&gt; Sent: Tuesday, September 17, 2019 11:15 AM To: Peng Yu <pengyu.ut(a)gmail.com&gt; Cc: Patrick Farrell <pfarrell(a)whamcloud.com&gt;; io-500(a)vi4io.org <io-500(a)vi4io.org&gt; Subject: Re: [IO-500] A list file systems used in top I/O system Hi Peng, welcome. The list is considerably new, that is why we hadn't have 500 entries... Look at the beginning of the Green500, for instance. Everyone that submits is there, not every user/data center in the world submits. Please consider to join the Slack: https://join.slack.com/t/vi4io/shared_invite/enQtMjMyOTgxMDg0OTQ1LTcyYWJk... That makes answering easy questions easier. Best, Julian On Tue, Sep 17, 2019 at 5:03 PM Peng Yu via IO-500 <io-500(a)vi4io.org&gt; wrote: -- Dr. Julian Kunkel Lecturer, Department of Computer Science +44 (0) 118 378 8218 http://www.cs.reading.ac.uk/ https://hps.vi4io.org/ PGP Fingerprint: 1468 1A86 A908 D77E B40F 45D6 2B15 73A5 9D39 A28E

I assume the top500.org systems are for applications different from those of io500. After all, top500.org, in my opinion, are more related to scientific computation, in which the scientific problems to solve is not embarrassingly parallelizable. For io500, I'd expect more embarrassingly parallelizable applications. In such cases, the storage size is more important, and the write/read pattern should be more serial. This probably explains the big difference between the two. Maybe others can also share their opinions on the difference. On Tue, Sep 17, 2019 at 1:57 PM John Bent <johnbent(a)gmail.com&gt; wrote: -- Regards, Peng

Hi Peng, IO500 is intended to be a companion to top500 in that the machines in that the benchmarks target similar workloads. Also note that Blue Waters at NCSA/UIUC is one of the most high profile machines to not submit to top500. There are certainly many others as has been pointed out. If you have a particular goal in mind in looking over the list, we may be able to help guide your search. Let us know if we can help. If there are particular benchmark goals you have, we can probably help with that as well. The community guided the development of the IO500 suite. There are a few very specific tests we don't have, such as a massively parallel load of python or java libraries. We haven't found a generally acceptable benchmark to represent these specific operations. Otherwise, general IO tasks should be well represented for both easy and hard test cases. Best, Jay From: IO-500 <io-500-bounces(a)vi4io.org&gt; on behalf of George Markomanolis via IO-500 <io-500(a)vi4io.org&gt; Reply-To: Georgios Markomanolis <george(a)markomanolis.com&gt; Date: Tuesday, September 17, 2019 at 1:52 PM To: Peng Yu <pengyu.ut(a)gmail.com&gt; Cc: &quot;io-500(a)vi4io.org&quot; <io-500(a)vi4io.org&gt;, Julian Kunkel <juliankunkel(a)googlemail.com&gt; Subject: [EXTERNAL] Re: [IO-500] A list file systems used in top I/O system Hi, Top500 is based on Linpack ( https://www.top500.org/project/linpack/ ) which many people believe that doesn't correspond to the scientific applications that we use today. So, I wouldn't say is more related to scientific computation. IO500 is about I/O, if you do science or not, this is another topic, we hope that the people who use storage, they do science. It is not embarking parallel, we have many benchmarks included in the IO500 suite that you can write one file per MPI process or shared file across MPI processes. Top500 and IO500 don't have the same purpose, except of evaluating different systems. A parallel application maybe writes dats. IO500 evaluates how efficient you will read/write data, accessing metadata operations etc. You can find some talks about IO500 here: https://www.vi4io.org/io500/news regards, George On Tue, Sep 17, 2019 at 3:41 PM Peng Yu via IO-500 <io-500@vi4io.org<mailto:io-500@vi4io.org>> wrote: I assume the top500.org<http://top500.org> systems are for applications different from those of io500. After all, top500.org<http://top500.org>, in my opinion, are more related to scientific computation, in which the scientific problems to solve is not embarrassingly parallelizable. For io500, I'd expect more embarrassingly parallelizable applications. In such cases, the storage size is more important, and the write/read pattern should be more serial. This probably explains the big difference between the two. Maybe others can also share their opinions on the difference. On Tue, Sep 17, 2019 at 1:57 PM John Bent <johnbent@gmail.com<mailto:johnbent@gmail.com>> wrote: Thanks Patrick. Peng, here's something else from Cray showing that Lustre was the file system on 77% of the top 100 systems from the June 2018 top500.org<http://top500.org> list: [cid:image001.png@01D571F9.6C02B5D0] https://www.cray.com/blog/business-cards-change-passion-open-source-lustr... Which is not nearly the same ratio that we have seen thus far in IO500: [cid:image002.png@01D571F9.6C02B5D0] Which I just created using https://www.vi4io.org/assets/io500/2019-06/data.csv ("The list shows the best result for a given combination of system/institution/filesystem"). Note that I did combine a few to reduce the number of slices (e.g. s/GPFS/Spectrum Scale/). -- Regards, Peng _______________________________________________ IO-500 mailing list IO-500@vi4io.org<mailto:IO-500@vi4io.org> https://www.vi4io.org/mailman/listinfo/io-500