[gmx-users] Details of 5.0 benchmark report?

Szilárd Páll pall.szilard at gmail.com
Mon Apr 18 22:16:59 CEST 2016


On Mon, Apr 18, 2016 at 9:25 PM, Adam Huffman <adam.huffman at gmail.com> wrote:
> Hi Szilárd,
>
> Yes, it's certainly not worth a large effort to try and recover the information.
>
> Thanks for the link, which should be very useful.
>
> The next time you're looking at something like this, it might be worth
> trying to capture what might be useful for others to run tests
> themselves.

Certainly, more information could be useful, but let me emphasize:
simple raw details of benchmark setups will be useful for reference
(especially for the experienced), but less useful than a thorough
study/analysis that explains the how and why as well.

Feel free to ask if you'll have further general questions or even
specific ones regarding your machine or simulation systems.

Cheers,
--
Szilárd


>
> Thanks again,
> Adam
>
>
> On Tue, Apr 12, 2016 at 5:36 PM, Szilárd Páll <pall.szilard at gmail.com> wrote:
>> Adam,
>>
>> The job scripts used in those benchmarks are all specific to the
>> machines, and while admittedly the launch configuration information
>> could be useful to have next to each data point, I'm afraid we don't
>> have such data readily extracted from the run logs, I believe. I did
>> not compile the document myself, but unless such data has been parsed
>> and it's just not presented, it would be a quite large effort to go
>> back and extract such information for quite little benefit (as the
>> exact ranks/thread configurations will change with GROMACS versions,
>> compilers, libraries, machine load, etc.)
>>
>> I recommend you to have a look at recent benchmark papers that contain
>> detailed information on what to tune and how to run parallel jobs,
>> e.g.
>> http://onlinelibrary.wiley.com/doi/10.1002/jcc.24030/full
>> Scaling of the GROMACS 4.6 molecular dynamics code on SuperMUC.
>>
>> Cheers,
>> --
>> Szilárd
>>
>>
>> On Mon, Apr 11, 2016 at 2:00 PM, Adam Huffman <adam.huffman at gmail.com> wrote:
>>> Hello
>>>
>>> Ideal would be to have the command-lines that were used, for those of
>>> us that aren't GROMACS experts.
>>>
>>> Cheers,
>>> Adam
>>>
>>> On Tue, Mar 8, 2016 at 5:08 PM, Szilárd Páll <pall.szilard at gmail.com> wrote:
>>>> Hi,
>>>>
>>>> What details are you interested in? IIRC the inputs are clearly specified
>>>> (although perhaps not provided there).
>>>>
>>>> --
>>>> Szilárd
>>>>
>>>> On Fri, Mar 4, 2016 at 5:15 PM, Adam Huffman <adam.huffman at gmail.com> wrote:
>>>>
>>>>> Hello
>>>>>
>>>>> Are details of the running of the benchmarks reported for version 5.0
>>>>> available anywhere?
>>>>>
>>>>> I'm referring to the report published at:
>>>>>
>>>>> http://www.gromacs.org/@api/deki/files/240/=gromacs-5.0-benchmarks.pdf
>>>>>
>>>>> I would like to be able to run these myself, on three different systems.
>>>>>
>>>>>
>>>>> Cheers,
>>>>> Adam
>>>>> --
>>>>> Gromacs Users mailing list
>>>>>
>>>>> * Please search the archive at
>>>>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>>>>> posting!
>>>>>
>>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>>
>>>>> * For (un)subscribe requests visit
>>>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>>>>> send a mail to gmx-users-request at gromacs.org.
>>>>>
>>>> --
>>>> Gromacs Users mailing list
>>>>
>>>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
>>>>
>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>
>>>> * For (un)subscribe requests visit
>>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-request at gromacs.org.
>>> --
>>> Gromacs Users mailing list
>>>
>>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
>>>
>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>
>>> * For (un)subscribe requests visit
>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-request at gromacs.org.
>> --
>> Gromacs Users mailing list
>>
>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
>>
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>> * For (un)subscribe requests visit
>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-request at gromacs.org.
> --
> Gromacs Users mailing list
>
> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-request at gromacs.org.


More information about the gromacs.org_gmx-users mailing list