3:36 p.m.
Hi,
sorry if this is just my ignorance but can I have access to values set
by "set"?
I know how to access current workdir value: it's just @workdir (as
string), but can I retrive value of omp_num_threads (as integer or
string, does not matter actually)?
Or can I do something like "eval set omp_num_threads" [of course not in
the current sence of eval] to get string "omp_num_threads: positive
integer, currently 4" which I could parse next?
Marcin
--
dr Marcin Błażejowski
Katedra Zarządzania oraz Instytut Rozwoju Badań Naukowych
Wyższa Szkoła Bankowa w Toruniu
ul. Młodzieżowa 31a
87-100 Toruń
pok. 116
--
Marcin Błażejowski
4:41 p.m.
On Sat, 28 Jan 2017, Marcin Błażejowski wrote:
Or can I do something like "eval set omp_num_threads" [of
course not in
the current sence of eval] to get string "omp_num_threads: positive
integer, currently 4" which I could parse next?
This is a bit baroque, but seems to work:
<hansl>
set echo off
set messages off
function scalar ompthds(void)
scalar ret
outfile s --buffer --quiet
set omp_num_threads
outfile --close
l = strlen(s)
c = ""
loop while c != " " --quiet
c = substr(s, l, l)
l--
endloop
sscanf(s+l, " %d", ret)
return ret
end function
n = ompthds()
printf "As of now, %d\n", n
</hansl>
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
5:32 p.m.
On Sat, 28 Jan 2017, Riccardo (Jack) Lucchetti wrote:
On Sat, 28 Jan 2017, Marcin Błażejowski wrote:
> Or can I do something like "eval set omp_num_threads" [of course not in
> the current sence of eval] to get string "omp_num_threads: positive
> integer, currently 4" which I could parse next?
This is a bit baroque, but seems to work:
<hansl>
set echo off
set messages off
function scalar ompthds(void)
scalar ret
outfile s --buffer --quiet
set omp_num_threads
outfile --close
l = strlen(s)
c = ""
loop while c != " " --quiet
c = substr(s, l, l)
l--
endloop
sscanf(s+l, " %d", ret)
return ret
end function
n = ompthds()
printf "As of now, %d\n", n
</hansl>
Slightly less baroque:
<hansl>
function scalar ompthds(void)
scalar ret
outfile s --buffer --quiet
set omp_num_threads
outfile --close
sscanf(strstr(s, ",") + 1, "%*s %d", ret)
return ret
end function
</hansl>
Allin
6:29 p.m.
On Sat, 28 Jan 2017, Allin Cottrell wrote:
On Sat, 28 Jan 2017, Riccardo (Jack) Lucchetti wrote:
> On Sat, 28 Jan 2017, Marcin Błażejowski wrote:
>
>> Or can I do something like "eval set omp_num_threads" [of course not
in
>> the current sence of eval] to get string "omp_num_threads: positive
>> integer, currently 4" which I could parse next?
>
> This is a bit baroque, but seems to work:
>
> <hansl>
> set echo off
> set messages off
>
> function scalar ompthds(void)
> scalar ret
> outfile s --buffer --quiet
> set omp_num_threads
> outfile --close
> l = strlen(s)
> c = ""
> loop while c != " " --quiet
> c = substr(s, l, l)
> l--
> endloop
> sscanf(s+l, " %d", ret)
> return ret
> end function
>
> n = ompthds()
> printf "As of now, %d\n", n
> </hansl>
Slightly less baroque:
<hansl>
function scalar ompthds(void)
scalar ret
outfile s --buffer --quiet
set omp_num_threads
outfile --close
sscanf(strstr(s, ",") + 1, "%*s %d", ret)
return ret
end function
</hansl>
Well, I had thought of something like that too, but couldn't be bothered
to check all the po files if that would be robust to translations, so I
just assumed the number would come last.
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
6:36 p.m.
Am 28.01.2017 um 19:29 schrieb Riccardo (Jack) Lucchetti:
On Sat, 28 Jan 2017, Allin Cottrell wrote:
>
> Slightly less baroque:
>
> <hansl>
> function scalar ompthds(void)
> scalar ret
> outfile s --buffer --quiet
> set omp_num_threads
> outfile --close
> sscanf(strstr(s, ",") + 1, "%*s %d", ret)
> return ret
> end function
> </hansl>
Well, I had thought of something like that too, but couldn't be bothered
to check all the po files if that would be robust to translations, so I
just assumed the number would come last.
Why not have a $set accessor bundle with all the current settings?
thanks,
sven
7:07 p.m.
On Sat, 28 Jan 2017, Sven Schreiber wrote:
Am 28.01.2017 um 19:29 schrieb Riccardo (Jack) Lucchetti:
> On Sat, 28 Jan 2017, Allin Cottrell wrote:
>>
>> Slightly less baroque:
>>
>> <hansl>
>> function scalar ompthds(void)
>> scalar ret
>> outfile s --buffer --quiet
>> set omp_num_threads
>> outfile --close
>> sscanf(strstr(s, ",") + 1, "%*s %d", ret)
>> return ret
>> end function
>> </hansl>
>
> Well, I had thought of something like that too, but couldn't be bothered
> to check all the po files if that would be robust to translations, so I
> just assumed the number would come last.
>
Why not have a $set accessor bundle with all the current settings?
Yes, that would be the "right" way. In fact we could revitalise the
$sysinfo bundle to that purpose. Now that I check, there is a much saner
way to do what Marcin needs:
<hansl>
scalar n = $sysinfo.omp_num_threads
</hansl>
:)
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
7:42 p.m.
On Sat, 28 Jan 2017, Riccardo (Jack) Lucchetti wrote:
On Sat, 28 Jan 2017, Sven Schreiber wrote:
> Am 28.01.2017 um 19:29 schrieb Riccardo (Jack) Lucchetti:
>> On Sat, 28 Jan 2017, Allin Cottrell wrote:
>
>>>
>>> Slightly less baroque:
>>>
>>> <hansl>
>>> function scalar ompthds(void)
>>> scalar ret
>>> outfile s --buffer --quiet
>>> set omp_num_threads
>>> outfile --close
>>> sscanf(strstr(s, ",") + 1, "%*s %d", ret)
>>> return ret
>>> end function
>>> </hansl>
>>
>> Well, I had thought of something like that too, but couldn't be bothered
>> to check all the po files if that would be robust to translations, so I
>> just assumed the number would come last.
>>
>
> Why not have a $set accessor bundle with all the current settings?
Yes, that would be the "right" way. In fact we could revitalise the $sysinfo
bundle to that purpose. Now that I check, there is a much saner way to do
what Marcin needs:
<hansl>
scalar n = $sysinfo.omp_num_threads
</hansl>
Urgh, I'd forgotten about that! Recall also:
print $sysinfo
for a look at what's available.
(I'd say if we reckon a "set" bundle is a worthwhile thing to have
it should be separate from $sysinfo, which is quite focused as it
stands.)
Allin
8:08 p.m.
W dniu 28.01.2017 o 20:42, Allin Cottrell pisze:
(I'd say if we reckon a "set" bundle is a worthwhile thing to have it
should be separate from $sysinfo, which is quite focused as it stands.)
IMHO such
bundle would be great think.
Nevertheless $sysinfo gives me info which I need - thank you, Jack!
Marcin
--
Marcin Błażejowski
8:02 p.m.
W dniu 28.01.2017 o 17:41, Riccardo (Jack) Lucchetti pisze:
On Sat, 28 Jan 2017, Marcin Błażejowski wrote:
> Or can I do something like "eval set omp_num_threads" [of course not in
> the current sence of eval] to get string "omp_num_threads: positive
> integer, currently 4" which I could parse next?
This is a bit baroque, but seems to work:
<hansl>
set echo off
set messages off
function scalar ompthds(void)
scalar ret
outfile s --buffer --quiet
set omp_num_threads
outfile --close
l = strlen(s)
c = ""
loop while c != " " --quiet
c = substr(s, l, l)
l--
endloop
sscanf(s+l, " %d", ret)
return ret
end function
n = ompthds()
printf "As of now, %d\n", n
</hansl>
I was thinking about something like that but not serious... ;)
It remains be a way for having a dynamic matrices in Turbo Pascal.
Marcin
--
Marcin Błażejowski
2930
days inactive
2930
days old
8 comments
4 participants
participants (4)
-
Allin Cottrell -
Marcin Błażejowski -
Riccardo (Jack) Lucchetti -
Sven Schreiber