Re: [gmx-users] Gromacs compilation on AMD multicore

[Alexey Shvetsov]

Also for additional performance benefit you can use atlas math library
and compile gromacs and fftw with additional cflags like '-mfpmath=sse -
m' This will enable sse math (by default gcc generates i386 and
sse).

PS Also you can use CFLAGS="-O2 -pipe -march=native -mfpmath=sse -
m'

PPS also on large numa systems cpuaffinity gains performance benefit. I made
simple test with modifyed d.ddpc [1] params from gmxbench. Test platform was
AMD 4xOpteron 6174 (12 core each) with 128G ram (32G per socket) so it was 48
cores total.

All tests were with double precision version of gromacs and openmpi 1.5.3

mpi  without cpu affinity

 R E A L   C Y C L E   A N D   T I M E   A C C O U N T I N G

 Computing: Nodes Number G-CyclesSeconds %
---
 Domain decomp.40   5001 4782.871 2173.6 0.7
 DD comm. load 40   5000  262.645  119.4 0.0
 DD comm. bounds   40   5000 1110.348  504.6 0.2
 Send X to PME 40  50001 3375.583 1534.1 0.5
 Comm. coord.  40  5000125220.57911461.8 3.8
 Neighbor search   40   500118235.819 8287.5 2.8
 Force 40  50001   274274.633   124647.741.4
 Wait + Comm. F40  5000186122.98439139.713.0
 PME mesh   8  5000176654.25134836.511.6
 Wait + Comm. X/F   8   33769.34015346.9 5.1
 Wait + Recv. PME F40  5000164206.34029179.4 9.7
 Write traj.   40  7   21.8499.9 0.0
 Update40  50001  596.730  271.2 0.1
 Constraints   40  5000171372.09932436.010.8
 Comm. energies40   5002 1065.027  484.0 0.2
 Rest  401471.920  668.9 0.2
---
 Total 48  662543.017   301101.4   100.0
---
---
 PME redist. X/F8 1224791.28211266.7 3.7
 PME spread/gather  8 1219021.991 8644.8 2.9
 PME 3D-FFT 8 1227949.86112702.2 4.2
 PME solve  8  50001 4885.045 2220.1 0.7
---

Parallel run - timing based on wallclock.

   NODE (s)   Real (s)  (%)
   Time:   6272.946   6272.946100.0
   1h44:32
   (Mnbf/s)   (GFlops)   (ns/day)  (hour/ns)
Performance:750.030 37.152  1.377 17.425


and mpi version with cpuaffinity via hwloc library
 R E A L   C Y C L E   A N D   T I M E   A C C O U N T I N G

 Computing: Nodes Number G-CyclesSeconds %
---
 Domain decomp.40   5001 2731.213 1241.3 0.7
 DD comm. load 40   5000   13.2006.0 0.0
 DD comm. bounds   40   5000  185.423   84.3 0.1
 Send X to PME 40  50001 1034.850  470.3 0.3
 Comm. coord.  40  50001 3810.865 1732.0 1.0
 Neighbor search   40   500117518.149 7961.9 4.8
 Force 40  50001   256996.831   116803.770.3
 Wait + Comm. F40  50001 9129.265 4149.2 2.5
 PME mesh   8  5000133662.70515299.5 9.2
 Wait + Comm. X/F   8   27242.61712381.6 7.5
 Wait + Recv. PME F40  50001  526.714  239.4 0.1
 Write traj.   40  47.4253.4 0.0
 Update40  50001  486.077  220.9 0.1
 Constraints   40  5000110789.300 4903.7 3.0
 Comm. energies40   5002  199.327   90.6 0.1
 Rest  401098.793  499.4 0.3
---
 Total 48  365432.754   166087.3   100.0
---
---
 PME redist. X/F8 12 2379.702 1081.6 0.7
 PME spread/gather  8 1211659.625 5299.2 3.2
 PME 3D-FFT 8 1214864.210 6755.7 4.1
 PME solve  8  50001 4753.059 2160.2 1.3
---

Parallel run - timing based on wallclock.

   NODE (s)   Real (s)  (%)
   Time:   3460.152   3460.152100.0
   

Re: [gmx-users] Gromacs compilation on AMD multicore

[Szilárd Páll]

Additionally, if you care about a few percent extra performance, you
should use gcc 4.5 or 4.6 for compiling Gromacs as well as FFTW
(unless you have a bleeding-edge OS which was built with any of these
latest gcc versions). While you might not see a lot of improvement in
mdrun performance (wrt gcc >v4.5), as far as I remember, FFTW gets
slightly more boost form the new gcc versions.

I can't comment on other compilers, I haven't tried to run binaries
compiled with Intel Compiler on AMD lately.

--
Szilárd



On Wed, Jul 6, 2011 at 3:08 AM, Mark Abraham  wrote:
> On 6/07/2011 1:36 AM, Anthony Cruz Balberdi wrote:
>
> Thank you very much Mark. Another question, will be better if I compile FFTW
> than use the one included in my linux distro?
>
> Per the GROMACS installation instructions, you want the precision of FFTW to
> match the precision of GROMACS. You'll have to find out what precision of
> FFTW is available for your distro and then decide. Note that
> single-precision will often be labelled fftwf or fftws, and double-precision
> fftw or fftwd.
>
> Mark
>
> Anthony
>
> On Tue, Jul 5, 2011 at 11:08 AM, Mark Abraham 
> wrote:
>>
>> On 6/07/2011 12:28 AM, Matthew Zwier wrote:
>>>
>>> Sorry about that.
>>>
>>> The default options are nearly optimal, and the difference between a
>>> modern (4.4 or 4.5) series GCC and the Intel compilers are only a
>>> couple of percent.  Just be sure to have FFTW available.
>>
>> ... if planning to use PME. Also, Intel MKL is only a percent or two
>> faster than FFTW3, last I tested, but much more fiddly to link.
>>
>> Mark
>>
>>> On Tue, Jul 5, 2011 at 10:26 AM, Matthew Zwier  wrote:

 Hi Anthony,

 The default options are nearly optimal, and the difference between a
 modern (4.4 or 4.5) GCC

 On Tue, Jul 5, 2011 at 10:01 AM, Anthony Cruz Balberdi
   wrote:
>
> Dear Users:
>
> We recently received our new computer.  This computer have 4 multicore
> Opteron AMD cpus and I am planning to install GROMAS. Which options I
> will
> need to use in order to achieve the best possible performance?
> Which compiler GNU or Intel?
>
> Thank you for all you help.
>
> Anthony
>
> --
> gmx-users mailing list    gmx-users@gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-requ...@gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
>>
>> --
>> gmx-users mailing list    gmx-users@gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> Please don't post (un)subscribe requests to the list. Use the www
>> interface or send it to gmx-users-requ...@gromacs.org.
>> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
>
>
> --
> gmx-users mailing list    gmx-users@gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-requ...@gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] Gromacs compilation on AMD multicore

[Mark Abraham]

On 6/07/2011 1:36 AM, Anthony Cruz Balberdi wrote:
Thank you very much Mark. Another question, will be better if I 
compile FFTW than use the one included in my linux distro?


Per the GROMACS installation instructions, you want the precision of 
FFTW to match the precision of GROMACS. You'll have to find out what 
precision of FFTW is available for your distro and then decide. Note 
that single-precision will often be labelled fftwf or fftws, and 
double-precision fftw or fftwd.


Mark


Anthony

On Tue, Jul 5, 2011 at 11:08 AM, Mark Abraham > wrote:


On 6/07/2011 12:28 AM, Matthew Zwier wrote:

Sorry about that.

The default options are nearly optimal, and the difference
between a
modern (4.4 or 4.5) series GCC and the Intel compilers are only a
couple of percent.  Just be sure to have FFTW available.


... if planning to use PME. Also, Intel MKL is only a percent or
two faster than FFTW3, last I tested, but much more fiddly to link.

Mark


On Tue, Jul 5, 2011 at 10:26 AM, Matthew
Zwiermailto:mczw...@gmail.com>>  wrote:

Hi Anthony,

The default options are nearly optimal, and the difference
between a
modern (4.4 or 4.5) GCC

On Tue, Jul 5, 2011 at 10:01 AM, Anthony Cruz Balberdi
mailto:anthony.cr...@upr.edu>>  wrote:

Dear Users:

We recently received our new computer.  This computer
have 4 multicore
Opteron AMD cpus and I am planning to install GROMAS.
Which options I will
need to use in order to achieve the best possible
performance?
Which compiler GNU or Intel?

Thank you for all you help.

Anthony

--
gmx-users mailing list gmx-users@gromacs.org

http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at
http://www.gromacs.org/Support/Mailing_Lists/Search
before posting!
Please don't post (un)subscribe requests to the list.
Use the
www interface or send it to
gmx-users-requ...@gromacs.org
.
Can't post? Read
http://www.gromacs.org/Support/Mailing_Lists


-- 
gmx-users mailing list gmx-users@gromacs.org


http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the www
interface or send it to gmx-users-requ...@gromacs.org
.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists




-- 
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] Gromacs compilation on AMD multicore

[Anthony Cruz Balberdi]

Thank you very much Mark. Another question, will be better if I compile FFTW
than use the one included in my linux distro?

Anthony

On Tue, Jul 5, 2011 at 11:08 AM, Mark Abraham wrote:

> On 6/07/2011 12:28 AM, Matthew Zwier wrote:
>
>> Sorry about that.
>>
>> The default options are nearly optimal, and the difference between a
>> modern (4.4 or 4.5) series GCC and the Intel compilers are only a
>> couple of percent.  Just be sure to have FFTW available.
>>
>
> ... if planning to use PME. Also, Intel MKL is only a percent or two faster
> than FFTW3, last I tested, but much more fiddly to link.
>
> Mark
>
>
>  On Tue, Jul 5, 2011 at 10:26 AM, Matthew Zwier  wrote:
>>
>>> Hi Anthony,
>>>
>>> The default options are nearly optimal, and the difference between a
>>> modern (4.4 or 4.5) GCC
>>>
>>> On Tue, Jul 5, 2011 at 10:01 AM, Anthony Cruz Balberdi
>>>   wrote:
>>>
 Dear Users:

 We recently received our new computer.  This computer have 4 multicore
 Opteron AMD cpus and I am planning to install GROMAS. Which options I
 will
 need to use in order to achieve the best possible performance?
 Which compiler GNU or Intel?

 Thank you for all you help.

 Anthony

 --
 gmx-users mailing listgmx-users@gromacs.org
 http://lists.gromacs.org/**mailman/listinfo/gmx-users
 Please search the archive at
 http://www.gromacs.org/**Support/Mailing_Lists/Searchbefore
  posting!
 Please don't post (un)subscribe requests to the list. Use the
 www interface or send it to gmx-users-requ...@gromacs.org.
 Can't post? Read 
 http://www.gromacs.org/**Support/Mailing_Lists


> --
> gmx-users mailing listgmx-users@gromacs.org
> http://lists.gromacs.org/**mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/**
> Support/Mailing_Lists/Searchbefore
>  posting!
> Please don't post (un)subscribe requests to the list. Use the www interface
> or send it to gmx-users-requ...@gromacs.org.
> Can't post? Read 
> http://www.gromacs.org/**Support/Mailing_Lists
>
-- 
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] Gromacs compilation on AMD multicore

[Mark Abraham]

On 6/07/2011 12:28 AM, Matthew Zwier wrote:

Sorry about that.

The default options are nearly optimal, and the difference between a
modern (4.4 or 4.5) series GCC and the Intel compilers are only a
couple of percent.  Just be sure to have FFTW available.


... if planning to use PME. Also, Intel MKL is only a percent or two 
faster than FFTW3, last I tested, but much more fiddly to link.


Mark


On Tue, Jul 5, 2011 at 10:26 AM, Matthew Zwier  wrote:

Hi Anthony,

The default options are nearly optimal, and the difference between a
modern (4.4 or 4.5) GCC

On Tue, Jul 5, 2011 at 10:01 AM, Anthony Cruz Balberdi
  wrote:

Dear Users:

We recently received our new computer.  This computer have 4 multicore
Opteron AMD cpus and I am planning to install GROMAS. Which options I will
need to use in order to achieve the best possible performance?
Which compiler GNU or Intel?

Thank you for all you help.

Anthony

--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists



--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] Gromacs compilation on AMD multicore

[Matthew Zwier]

Sorry about that.

The default options are nearly optimal, and the difference between a
modern (4.4 or 4.5) series GCC and the Intel compilers are only a
couple of percent.  Just be sure to have FFTW available.

On Tue, Jul 5, 2011 at 10:26 AM, Matthew Zwier  wrote:
> Hi Anthony,
>
> The default options are nearly optimal, and the difference between a
> modern (4.4 or 4.5) GCC
>
> On Tue, Jul 5, 2011 at 10:01 AM, Anthony Cruz Balberdi
>  wrote:
>> Dear Users:
>>
>> We recently received our new computer.  This computer have 4 multicore
>> Opteron AMD cpus and I am planning to install GROMAS. Which options I will
>> need to use in order to achieve the best possible performance?
>> Which compiler GNU or Intel?
>>
>> Thank you for all you help.
>>
>> Anthony
>>
>> --
>> gmx-users mailing list    gmx-users@gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-requ...@gromacs.org.
>> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>
--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] Gromacs compilation on AMD multicore

[Matthew Zwier]

Hi Anthony,

The default options are nearly optimal, and the difference between a
modern (4.4 or 4.5) GCC

On Tue, Jul 5, 2011 at 10:01 AM, Anthony Cruz Balberdi
 wrote:
> Dear Users:
>
> We recently received our new computer.  This computer have 4 multicore
> Opteron AMD cpus and I am planning to install GROMAS. Which options I will
> need to use in order to achieve the best possible performance?
> Which compiler GNU or Intel?
>
> Thank you for all you help.
>
> Anthony
>
> --
> gmx-users mailing list    gmx-users@gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-requ...@gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

[gmx-users] Gromacs compilation on AMD multicore

[Anthony Cruz Balberdi]

Dear Users:

We recently received our new computer.  This computer have 4 multicore
Opteron AMD cpus and I am planning to install GROMAS. Which options I will
need to use in order to achieve the best possible performance?
Which compiler GNU or Intel?

Thank you for all you help.

Anthony
-- 
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists