In my case the goto libs were compiled in situ, without the patches (1.01 and 1.02) and <br>
compilation options dictated by the configuration script.<br>
<br>
<br>
The W2K version tested is the last release in a 1.5Ghz IA2 Altix. machine.<br>
<br>
Note that this is a 1.5 vs. 1.6 GHz.<br>
<br>
Emilio Orgaz<br>
<br>
<br><div><span class="gmail_quote">On 6/15/06, <b class="gmail_sendername">Peter Blaha</b> <<a href="mailto:pblaha@theochem.tuwien.ac.at">pblaha@theochem.tuwien.ac.at</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Just noticed, that your timing does not seem to be cortrect. Check<br>cputim.c; Some machines report the cputime in different units and you<br>should divide by 1024 instead of 100.<br><br>I've also heared from an Bull Itanium system and they gave very similar
<br>timings as the Altix. So I guess these values should be reachable. Since<br>your "Hamilt-time" is ok (besides the factor 10), it is mainly the<br>gotoblas, which makes the difference. Also your report that you get the
<br>same timings with/without goto tells me, that you are NOT using the<br>gotolib (or not a good version). I guess, the goto-libraries were changed<br>with version 1.0 and one has to compile them himself, so maybe older<br>
versions (precompiled) were more efficient .... ???<br><br>Always remember: The hamilt time is purely f90 + svml-lib code, hns is<br>mixed f90 + blas, and diag is purely blas-code. So compare the partial<br>times.<br><br>
>
Thank you for your confirmation. I have tried really hard to reproduce
the 122 sec posted in the Wien2K benchmark page, unfortunately without
success.<br>><br>> I get exactly the same as you. After x lapw1 -c in the test_case:<br>><br>> TIME HAMILT (CPU) = 133.6, HNS = 222.4, DIAG = 1331.3<br>> NUMBER
OF
K-POINTS: 1<br>> ===>
TOTAL CPU TIME: 1710.1
(INIT = 21.3 + K-POINTS = 1688.8)<br>> >
SUM OF WALL CLOCK TIMES: 167.4 (INIT
= 2.2 + K-POINTS
= 165.2)<br>> Maximum WALL clock time: 167.726181983948<br>>
Maximum CPU
time:
1710.76000000000<br>><br>> Interestingly, I get the same with and
without GOTO libraries. Here are my compilation options corresponding
to the times listed above:<br>><br>> O
Compiler options: -O3
-tpp2 -ftz -ip -fno-alias -cm -w -FR -DALTIX -DINTEL_VML<br>> L
Linker
Flags: -L../SRC_lib
-Vaxlib -L/opt/intel/mkl/8.0.2/lib/64<br>> P Preprocessor flags '-DParallel'<br>> R R_LIB (LAPACK+BLAS): -lmkl_lapack64 -lmkl -lvml<br>><br>> An here are the times of Martin Hilgeman on a similar machine (as posted in a recent email):
<br>><br>><br>> TIME
HAMILT (CPU) = 11.2, HNS
= 15.8, DIAG = 94.0<br>> ===>
TOTAL CPU
TIME: 123.2 (INIT
= 2.1 + K-POINTS =<br>> 121.1)<br>> >
SUM OF WALL CLOCK TIMES: 123.6 (INIT
= 2.2 + K-POINTS =<br>> 121.4)<br>> Maximum WALL clock time: 123.825130939484<br>>
Maximum CPU
time:
123.211914062500<br>><br>> I carried out the test calculations
with Wien2K version VERSION_05.6 while Martin used version WIEN2k_06.1.
I do not believe that the different versions are the cause of such
difference in running times (aprox. 45 sec.); however, I will check it.
Maybe the problem has to to with the differences in CPU times. I will
investigate more about this. Please, let me know if you manage to
reduce the running time of the test_case.<br>><br>> Regards<br>><br>><br>> Cesar Lazo<br>> Technische Universität Hamburg-Harburg<br>><br>><br>><br>><br>><br>><br>> > -----Ursprüngliche Nachricht-----
<br>> > Von: A Mailing list for WIEN2k users <<a href="mailto:wien@zeus.theochem.tuwien.ac.at">wien@zeus.theochem.tuwien.ac.at</a>><br>> > Gesendet: 14.06.06 18:44:28<br>> > An: "A Mailing list for WIEN2k users" <
<a href="mailto:wien@zeus.theochem.tuwien.ac.at">wien@zeus.theochem.tuwien.ac.at</a>><br>> > Betreff: Re: [Wien] Benchmark-Compilation options<br>><br>><br>> > For information;<br>> ><br>> > I compiled W2K as Cesar Clazo did'it (
ifort9.0 +mkl8.0, libgoto_itanium2_64p-r1.00)<br>> ><br>> > Results for the benchmark in a ALTIX Itanium2 @ 1.5 GHz with 4 MB in L3 are ;<br>> ><br>> > ********************************************************
<br>> ><br>>
> NUMBER OF
K-POINTS: 1<br>> > ===> TOTAL CPU<br>> > TIME: 1734.1 (INIT = 21.3 +<br>> > K-POINTS = 1712.9)<br>> > > SUM OF WALL CLOCK TIMES: 169.8<br>> > (INIT = 2.2
+ K-POINTS<br>> > = 167.6)<br>> > Maximum WALL clock time: 169.997846841812<br>>
> Maximum CPU
time:
1734.78000000000<br>> ><br>> ><br>> > best regards<br>> ><br>> > Dr. Emilio Orgaz<br>> > Departamento de Física y Química Teórica<br>> > Facultad de Química, UNAM<br>> > <a href="mailto:Emilio.Orgaz@gmail.com">
Emilio.Orgaz@gmail.com</a><br>> ><br>> ><br>> ><br>> > On 6/9/06, Clazo Clazo <<a href="mailto:clazo36@web.de">clazo36@web.de</a>> wrote:<br>> > Dear all,<br>> ><br>> > I<br>
> > compiled the wien code VERSION_05.6:<br>> > 20.10.2005 in a SGI Altix 4700 system with Itanium2 Madison<br>> > 9M Prozessoren (1.66GHz/6MB L3 Cache) , ifort version 9.0 + mkl8.0 +<br>> > libgoto_ITANIUM2-
r1.00.a, and I carried out the standard<br>> > test_case provided in the wien2k web site in order to see how well the<br>> > tunning of the compiler and libraries was. I obtained a running time of<br>> > 160 Sec. The running time for a similar setup reported in the wien2k
<br>> > Benchmarks site is:<br>> ><br>> > Itanium2(1.6GHz,SGI Altix 3700) 122 sec ifort9.0 +mkl8.0, libgoto_itanium2_64p-r1.00<br>> ><br>> > Thus,<br>> > faster than the time I got. The compiling options (most of
<br>> > them are the default options recommended by the siteconfig script) are<br>> > the following:<br>> ><br>> > O Compiler<br>> > options: -O3 -tpp2 -ftz<br>> > -ip -fno-alias -cm -w -FR -DALTIX -DINTEL_VML
<br>> ><br>> > L Linker<br>> > Flags: -L../SRC_lib<br>> > -Vaxlib<br>> > P Preprocessor flags '-DParallel'<br>>
> R R_LIB
(LAPACK+BLAS): -L/fibus/fs1/16/vt4cl/LIB/GotoBLAS/libgoto_ITANIUM2-r1.00.a -lscs -L/opt/intel/mkl/8.0.2/lib/64/ -lvml<br>> ><br>> > I downloaded and compiled the last version of GOTO library from the web.
<br>> ><br>> ><br>> > Has<br>> > somebody having a similar system been able to reproduce the running<br>> > time posted in the wien2k Benchmark page (122 sec)? If so, could you<br>> > post the options used to compile the code?
<br>> ><br>> > The running time I<br>> > get (160 sec) is already good, but I think, if the code is able to run<br>> > faster in the computational setup available, I should try to tune the<br>> > compiler and libraries in order to get the best possible performance.
<br>> ><br>> > Thank you in advance,<br>> ><br>> ><br>> > Cesar Lazo<br>> > Technische Universität Hamburg-Harburg<br>> ><br>> ><br>> ><br>> ><br>> ><br>> >
<br>> > _____________________________________________________________________<br>> ><br>> > Der <a href="http://WEB.DE">WEB.DE</a> SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen!<br>> >
<a href="http://smartsurfer.web.de/?mc=100071&distributionid=000000000071">http://smartsurfer.web.de/?mc=100071&distributionid=000000000071</a><br>> ><br>> ><br>> > _______________________________________________
<br>> > Wien mailing list<br>> > <a href="mailto:Wien@zeus.theochem.tuwien.ac.at">Wien@zeus.theochem.tuwien.ac.at</a><br>> ><br>> > <a href="http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien">
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien</a><br>> ><br>> ><br>> ><br>> > --<br>> > Dr. Emilio Orgaz<br>> > Departamento de Física y Química Teórica<br>> > Facultad de Química, UNAM
<br>> > <a href="mailto:Emilio.Orgaz@gmail.com">Emilio.Orgaz@gmail.com</a><br>> > Tel. 5622-3776<br>> > Fax. 5622-3521<br>> ><br>> > -----------------------------------------------------------------
<br>> > _______________________________________________<br>> > Wien mailing list<br>> > <a href="mailto:Wien@zeus.theochem.tuwien.ac.at">Wien@zeus.theochem.tuwien.ac.at</a><br>> > <a href="http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien">
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien</a><br>> ><br>><br>><br>> ____________________________________________________________________________<br>> ____________________________________________________________________________
<br>> WM Spa&szlig; mit Style! Coole Spr&uuml;che, Outfits & Co. rund um Deine<br>> Lieblingsspieler als kostenlose eCard versenden! <a href="http://www.sendaplayer.net">http://www.sendaplayer.net</a><br>
><br>> _______________________________________________<br>> Wien mailing list<br>> <a href="mailto:Wien@zeus.theochem.tuwien.ac.at">Wien@zeus.theochem.tuwien.ac.at</a><br>> <a href="http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien">
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien</a><br>><br><br><br> P.Blaha<br>--------------------------------------------------------------------------<br>Peter BLAHA, Inst.f
. Materials Chemistry, TU Vienna, A-1060 Vienna<br>Phone: +43-1-58801-15671 FAX: +43-1-58801-15698<br>Email: <a href="mailto:blaha@theochem.tuwien.ac.at">blaha@theochem.tuwien.ac.at</a> WWW: <a href="http://info.tuwien.ac.at/theochem/">
http://info.tuwien.ac.at/theochem/</a><br>--------------------------------------------------------------------------<br>_______________________________________________<br>Wien mailing list<br><a href="mailto:Wien@zeus.theochem.tuwien.ac.at">
Wien@zeus.theochem.tuwien.ac.at</a><br><a href="http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien">http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien</a><br></blockquote></div><br><br clear="all"><br>-- <br>Dr. Emilio Orgaz
<br>Departamento de Física y Química Teórica<br>Facultad de Química, UNAM<br><a href="mailto:Emilio.Orgaz@gmail.com">Emilio.Orgaz@gmail.com</a><br>Tel. 5622-3776 <br>Fax. 5622-3521