Of course it is using 600 or 800% of the cpu, BUT it does not run faster
at all !!!
Compare the timings in case.dayfile. You really can see at what time a
step starts and how long it really took.
On 04/04/2014 11:23 AM, shamik chakrabarti wrote:
Dear Prof. Blaha Sir,
Thank you for your response. We have started the calculation by
following your suggestions. But as we are using MSR1a instead of
min_lapw...it may now take more time than earlier.
We have actually 16 cpus in our system & I have set OMP_NUM_THREADS=8 &
for a single calculation we are getting up to 800% usage. However, if we
start another calculation it then takes rest of the cpus and we can get
all the cpus running with two simultaneous calculation divided in 16
cpus with 800% for each calculation. I have changed OMP_NUM_THREADS
value starting from 8 to 16 with gradual increase, but the cpu usage
remain same for all the cases.
Yes, we should look for k-parallel run for optimum use of our
resources....but at the moment I don't have any expertise over that kind
of installation. We will try to installed k-parallelization with the
help of your suggestions and surely will try to improve calculation speed.
Thank you Sir....thank your for all your suggestions.
with regards,
On Fri, Apr 4, 2014 at 12:48 PM, Peter Blaha
<[email protected] <mailto:[email protected]>> wrote:
The message does not say anything about the reason. You have to
search for other messages with more information.
My suggestion: MSR1a is MUCH more stable for an inexperienced user
then min_lapw, which has problems when its "history" contains a "bad
point".
I suggest:
cp case.inm case.inm_msr1a
edit case.inm_msr1a and put MSR1a (instead of MSR1)
in your optimize.job script:
comment out the line min_lapw .....
instead, activate again runsp -orb -fc 1 -cc 0.001
and insert the following line just BEFORE the runsp line:
cp case.inm_msr1a case.inm
PS: from your dayfile I can see:
> lapw1 -dn -orb -c (20:18:44)
> 1603.746u 24.052s 4:26.95 609.7%0+0k 0+382600io 0pf+0w
the 600% tells me that you are using OMP_NUM_THREADS=6
which is completely useless. Set it to 2 and do a
k-parallel run (study the UG about parallelization) and test it on a
small testcase.
I suppose you installed wien2k using "shared memory" machine ?
Then you basically need to create .machines file with 3 lines:
1:localhost
1:localhost
1:localhost
(and together with OMP_NUM_THREADS=2 you will use all 6 cores and
your time will go down from 4:30 to 1:30 for an lapw1 step.
On 04/04/2014 05:55 AM, shamik chakrabarti wrote:
Dear Prof. Blaha Sir,
Sorry to bother you again. But our
calculation again
get stopped after 4th iteration with the following message.
stop error
error: command /usr/local/WIEN2k/mini mini.def failed
> mini(20:24:31) 0.004u 0.003s 0:00.03 0.0%0+0k 2472+48io
11pf+0w
> stop
:CHARGE convergence: 0 0.0000 .0002465
:ENERGY convergence: 0 0 .0000073950000000
> mixer -orb(20:24:29) 7.820u 0.390s 0:01.65 497.5%0+0k
168+40568io
1pf+0w
> lcore -dn(20:24:29) 0.046u 0.003s 0:00.05 80.0%0+0k
0+656io 0pf+0w
> lcore -up(20:24:29) 0.045u 0.006s 0:00.05 80.0%0+0k
0+656io 0pf+0w
> lapwdm -dn -c (20:24:27) 8.644u 0.378s 0:01.64 549.3%0+0k
0+48io
0pf+0w
> lapwdm -up -c (20:24:25) 9.470u 0.403s 0:01.64 601.8%0+0k
0+48io
0pf+0w
> lapw2 -dn -c -orb (20:23:35) 282.734u 8.097s 0:49.73
584.7%0+0k
0+9816io 0pf+0w
> lapw2 -up -c -orb (20:22:53) 226.301u 6.095s 0:42.32
549.1%0+0k
0+9824io 0pf+0w
1524.395u 21.656s 4:09.36 620.0%0+0k 0+385000io 0pf+0w
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
> lapw1 -dn -orb -c (20:18:44) _nb in zhcgst.F 640
128
1603.746u 24.052s 4:26.95 609.7%0+0k 0+382600io 0pf+0w
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
_nb in zhcgst.F 640 128
> lapw1 -up -orb -c (20:14:17) _nb in zhcgst.F 640
128
> orb -dn (20:14:17) 0.002u 0.002s 0:00.00 0.0%0+0k 0+32io
0pf+0w
> orb -up (20:14:17) 0.002u 0.000s 0:00.00 0.0%0+0k 0+32io
0pf+0w
:FORCE convergence: 1 1.0 .010 XCO .009 XCO .037 XCO .013 XCO
.031 XCO
.014 XCO
> lapw0 (20:14:08) 8.316u 0.047s 0:08.36 99.8%0+0k 0+16424io
0pf+0w
cycle 4 (Thu Apr 3 20:14:08 IST 2014) (37/96 to go)
*We have checked case.scf_mini and the forces in the structure
are as
given below:*
TOTAL FORCE IN mRy/a.u. = |F| Fx Fy
Fz
with/without FOR in
:FOR001: 1.ATOM 16.382428 -2.026946 -1.179494
16.213705 total forces
:FOR002: 2.ATOM 13.581345 13.580984 0.098977
0.000000 total forces
:FOR003: 3.ATOM 9.631644 6.298244 7.287022
0.000000 total forces
:FOR004: 4.ATOM 9.227031 -1.691938 -2.752735
-8.642795 total forces
:FOR005: 5.ATOM 3.096054 -1.485397 2.716458
0.000000 total forces
:FOR006: 6.ATOM 27.808540 -10.866967 -25.597343
0.000000 total forces
Sir now what could be the reason for this failure. Looking
forward to
your reply eagerly
Thanks in advance,
with regards.
On Thu, Apr 3, 2014 at 8:19 PM, shamik chakrabarti
<[email protected] <mailto:[email protected]>
<mailto:[email protected] <mailto:[email protected]>>> wrote:
Dear Prof. Peter Blaha & Gavin Abo Sir,
Thanks for your responses. Yes Prof. Blaha is
right. It
was indeed the problem. We have removed case.scf and rerun the
calculation & now it is running smoothly. Thank you Sir.
with regards,
On Thu, Apr 3, 2014 at 7:34 PM, Peter Blaha
<[email protected]
<mailto:[email protected]>
<mailto:pblaha@theochem.__tuwien.ac.at
<mailto:[email protected]>>>
wrote:
You are using min_lapw ?
I suggest:
When the scf cycle crashes after a few steps and you
restart,
you have to remove case.scf before resubmission.
Otherwise it
attempts to start immediately the min program and
generates a
new structure (with non-converged forces).
On 04/03/2014 02:34 PM, shamik chakrabarti wrote:
Dear wien2k users,
We are working on a Li based silicate
materials. We
are trying to
do simultaneous optimization of lattice parameters
and atomic
coordinates. For that we are using "Option 6" in volume
optimize program
while edited optimized.job to perform simultaneous
force
minimization.
The calculation was run smoothly up to 2 structures
i.e,
case_abc_1.0
and case_abc_2.0. Then due to power failure the
calculation
was remain
stopped for several hours.. We restarted the
calculation by
putting # to
the two lines corresponding to 1st and 2nd structure in
optimize.job...such that the calculation starts
from the
structure
case_abc_3.0....However, while running for 3rd
structure the
following
display came in "show dayfile" option.
ERROR status in case_abc___3.0
mini 00000000004035D9 Unknown
Unknown Unknown
libc.so.6 000000349521ECDD Unknown
Unknown Unknown
mini 00000000004036E6 Unknown
Unknown Unknown
mini 0000000000412AA6 MAIN__
25 mini.f
mini 000000000040C6AB haupt_
593 haupt.f
mini 000000000041A39F wrtscf_
23 wrtscf.f
mini 0000000000451F5A Unknown
Unknown Unknown
mini 0000000000453C1E Unknown
Unknown Unknown
Image PC Routine
Line
Source
forrtl: severe (64): input conversion error, unit
-5, file
Internal
Formatted Read
3.887u 0.013s 0:03.90 99.7%0+0k 0+11912io 0pf+0w
DSTART ENDS
3.879u 0.008s 0:03.88 99.7%0+0k 0+11904io 0pf+0w
DSTART ENDS
3.879u 0.011s 0:03.89 99.7%0+0k 0+11904io 0pf+0w
DSTART ENDS
clmextrapol_lapw has generated a new case.clmdn
0.196u 0.007s 0:00.20 95.0%0+0k 0+8032io 0pf+0w
3.888u 0.015s 0:03.90 99.7%0+0k 0+13528io 0pf+0w
DSTART ENDS
running dstart in single mode
clmextrapol_lapw has generated a new case.clmup
0.196u 0.010s 0:00.20 100.0%0+0k 0+8032io 0pf+0w
3.929u 0.020s 0:03.94 100.0%0+0k 0+13528io 0pf+0w
DSTART ENDS
running dstart in single mode
clmextrapol_lapw has generated a new case.clmsum
0.195u 0.004s 0:00.20 95.0%0+0k 0+8032io 0pf+0w
3.923u 0.017s 0:03.94 99.7%0+0k 0+13528io 0pf+0w
DSTART ENDS
running dstart in single mode
3.888u 0.003s 0:03.89 99.7%0+0k 0+11904io 0pf+0w
DSTART ENDS
3.895u 0.010s 0:03.91 99.7%0+0k 0+11904io 0pf+0w
DSTART ENDS
3.934u 0.010s 0:03.94 100.0%0+0k 0+11912io 0pf+0w
DSTART ENDS
we are using wien2k 13.1
What could be the possible reasons for this error?.any
response in this
regard will be fruitful for us. Thanks in advance.
with regards,
--
Shamik Chakrabarti
Senior Research Fellow
Dept. of Physics & Meteorology
Material Processing & Solid State Ionics Lab
IIT Kharagpur
Kharagpur 721302
INDIA
___________________________________________________
Wien mailing list
[email protected].____at
<mailto:[email protected].__tuwien.ac.at
<mailto:[email protected]>>
http://zeus.theochem.tuwien.____ac.at/mailman/listinfo/wien
<http://ac.at/mailman/listinfo/wien>
<http://zeus.theochem.tuwien.__ac.at/mailman/listinfo/wien
<http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien>>
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/[email protected].____at/index.html
<http://www.mail-archive.com/[email protected].__at/index.html>
<http://www.mail-archive.com/[email protected].__at/index.html
<http://www.mail-archive.com/[email protected]/index.html>>
--
P.Blaha
------------------------------____----------------------------__--__--------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna,
A-1060 Vienna
Phone: +43-1-58801-165300 FAX:
+43-1-58801-165982
Email: [email protected]
<mailto:[email protected]>
<mailto:[email protected].__ac.at
<mailto:[email protected]>> WWW:
http://info.tuwien.ac.at/____theochem/
<http://info.tuwien.ac.at/__theochem/>
<http://info.tuwien.ac.at/__theochem/
<http://info.tuwien.ac.at/theochem/>>
------------------------------____----------------------------__--__--------------
___________________________________________________
Wien mailing list
[email protected].____at
<mailto:[email protected].__tuwien.ac.at
<mailto:[email protected]>>
http://zeus.theochem.tuwien.____ac.at/mailman/listinfo/wien
<http://ac.at/mailman/listinfo/wien>
<http://zeus.theochem.tuwien.__ac.at/mailman/listinfo/wien
<http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien>>
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/[email protected].____at/index.html
<http://www.mail-archive.com/[email protected].__at/index.html>
<http://www.mail-archive.com/[email protected].__at/index.html
<http://www.mail-archive.com/[email protected]/index.html>>
--
Shamik Chakrabarti
Senior Research Fellow
Dept. of Physics & Meteorology
Material Processing & Solid State Ionics Lab
IIT Kharagpur
Kharagpur 721302
INDIA
--
Shamik Chakrabarti
Senior Research Fellow
Dept. of Physics & Meteorology
Material Processing & Solid State Ionics Lab
IIT Kharagpur
Kharagpur 721302
INDIA
_________________________________________________
Wien mailing list
[email protected].__at
<mailto:[email protected]>
http://zeus.theochem.tuwien.__ac.at/mailman/listinfo/wien
<http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien>
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/[email protected].__at/index.html
<http://www.mail-archive.com/[email protected]/index.html>
--
P.Blaha
------------------------------__------------------------------__--------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300 FAX: +43-1-58801-165982
Email: [email protected]
<mailto:[email protected]> WWW:
http://info.tuwien.ac.at/__theochem/
<http://info.tuwien.ac.at/theochem/>
------------------------------__------------------------------__--------------
_________________________________________________
Wien mailing list
[email protected].__at
<mailto:[email protected]>
http://zeus.theochem.tuwien.__ac.at/mailman/listinfo/wien
<http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien>
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/[email protected].__at/index.html
<http://www.mail-archive.com/[email protected]/index.html>
--
Shamik Chakrabarti
Senior Research Fellow
Dept. of Physics & Meteorology
Material Processing & Solid State Ionics Lab
IIT Kharagpur
Kharagpur 721302
INDIA
_______________________________________________
Wien mailing list
[email protected]
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/[email protected]/index.html
--
P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300 FAX: +43-1-58801-165982
Email: [email protected] WWW:
http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------
_______________________________________________
Wien mailing list
[email protected]
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
SEARCH the MAILING-LIST at:
http://www.mail-archive.com/[email protected]/index.html
