Welcome to MilkyWay@home

Posts by khryl

1) Message boards : News : Scheduled Maintenance Concluded (Message 65998)
Posted 10 Dec 2016 by Profile khryl
Post:
deleted
2) Message boards : Number crunching : not getting any work - boinc not recognizing gpu anymore (Message 65997)
Posted 10 Dec 2016 by Profile khryl
Post:
thank you wb8ili.

i had tried reinstalling drivers before your reply, but i reinstalled the same version cause i wanted to keep it. it still hadnt worked, so i thought id give the new driver a try (release dec 8) and now it works like a charm.

edit: i'm getting a ridiculous amount of validation inconclusive WUs. is that normal with these new ones?

http://milkyway.cs.rpi.edu/milkyway/results.php?userid=150155&offset=0&show_names=0&state=4&appid=
3) Message boards : Number crunching : not getting any work - boinc not recognizing gpu anymore (Message 65993)
Posted 8 Dec 2016 by Profile khryl
Post:
hello guys,

i wanted to get back into doing some crunching, but for some reason, boinc doesn't properly recognize my gpu anymore (i havent changed anything since last time i had mw on).

12/08/16 11:40:50 PM | | CAL: ATI GPU 0: AMD Radeon HD 7870/7950/7970/R9 280/R9 280X series (Tahiti) (CAL version 1.4.1848, 3072MB, 3032MB available, 5530 GFLOPS peak)
12/08/16 11:40:50 PM | Milkyway@Home | Application uses missing ATI GPU

i use boinc 7.6.33, 1 or 2 r9 280x with catalyst 15.7 (it worked just fine with this driver before) on windows 10.

since i really did not change a thing, i am really curious as to why my gpu[s] arent recognized anymore.
anyone got a heads up?

thanks in advance

edit: this applies to all projects, not only milkyway.
4) Message boards : Number crunching : threads per gpu? (Message 64892)
Posted 15 Jul 2016 by Profile khryl
Post:
with the new short workunits having arrived, could you guys (preferably someone having 280x or 7970 cards) give some info as to how many workunits you have run simultaneously?

i ve tried 1, 2, 4, 6 and i dont know whats best.
anyone would like to share their experience?

1 simultaneously is just not enough to keep the gpu at even 50% load. 2 simultaneously shows a load of ~60% with 1 card crunching 2 workunits every 10 seconds. 4 works best in terms of full load, but the workunits have those annoying patterns that even when you manually suspend some of them to keep the card loaded at all times, they return to their initial state of 4 units with all the same progress.

thanks in advance and happy crunching
5) Message boards : Number crunching : What am I missing in order to get more GPU tasks? (Message 64853)
Posted 8 Jul 2016 by Profile khryl
Post:
i just got 50 with one fetch, twice the amount i got max. at any time. thanks for the change!
6) Message boards : Number crunching : New AMD 480 GPU (Message 64815)
Posted 5 Jul 2016 by Profile khryl
Post:
someone on another forum claimed that SP/DP ratio of the rx480 chip is 1/16th, so with a listed SP of 5161 (wikipedia) it d be far worse than the top performers (7970/280x), which seem to remain the best cards for milkyway - in case the statement about 1/16th of SP performance turns out to be correct

update:

from the seti-germany forum:
Tja, keine Ahnung ob's bei Milkyway ein Problem gibt, aber ich bekomme keine AMD/ATI WUs...

Edit:
Irgendwie scheint Milkyway meine Graka als "R6xx" zu erkennen. Und die wäre nicht OpenCL kompatibel.

Da muss wohl noch nachgebessert werden, eh wir Milky auf der RX480 crunchen können...


'well... no idea if there s a problem at milkyway's side, but i cannot get any AMD/ATI workunits.
edit: it would appear as if milkyway recognized my card as "r6xx" - which is not compatible to OpenCL.
there is some work to be done before we can crunch milkyway with the rx480, it seems.'
7) Message boards : News : Running Modfit on MilkyWay@home (Message 64794)
Posted 2 Jul 2016 by Profile khryl
Post:
why is it that after every failed request (failed as in zero work received) milkyway stops fetching work and I have to manually request to receive workunits again?

my pc keeps running out of work every 15 or so minutes and just stops fetching work thereafter ..
8) Message boards : Number crunching : AMD GPU setup (Message 64775)
Posted 29 Jun 2016 by Profile khryl
Post:
<app>
<name>milkyway_separation__modified_fit</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.01</cpu_usage>
</gpu_versions>
</app>

</app_config>

complete, it d look about like this:

<app_config>

<app>
<name>milkyway</name>
<max_concurrent>4</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.01</cpu_usage>
</gpu_versions>
</app>

<app>
<name>milkyway_separation__modified_fit</name>
<max_concurrent>4</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.01</cpu_usage>
</gpu_versions>
</app>

</app_config>


just adjust numbers to your liking and you are good to go
9) Message boards : News : Running Modfit on MilkyWay@home (Message 64753)
Posted 26 Jun 2016 by Profile khryl
Post:
just would like to state that i m having the problem of no work as well,

my 2x280x are running out of work every 4th - give or take - scheduler request. is there possibly a way to increase the maximum tasks users can receive per 1 request?
10) Message boards : Number crunching : What am I missing in order to get more GPU tasks? (Message 64559)
Posted 20 May 2016 by Profile khryl
Post:
the amount of tasks received from milkyway@home is restricted by your cpu core count. i think it's something like 10 or 12 tasks/core
11) Message boards : Number crunching : got 0 new tasks ATI GPU (Message 64343)
Posted 24 Feb 2016 by Profile khryl
Post:
last line explains why: 24.02.2016 16:08:11 | Milkyway@Home | Not sending work - last request too recent: 21 sec
12) Message boards : Number crunching : I only get a util. of about 30% with 2 280x's -- how to fix??? Help please (Message 64322)
Posted 7 Feb 2016 by Profile khryl
Post:
nevermind the above, found out myself :P
13) Message boards : Number crunching : I only get a util. of about 30% with 2 280x's -- how to fix??? Help please (Message 64321)
Posted 7 Feb 2016 by Profile khryl
Post:
how did you get boinc to run with 2 gpus?

my boinc (newest version) recognizes 2 gpus ( i have 2x r9 280x), but only uses 1. my app info is:

<app_config>

<app>
<name>milkyway</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.01</cpu_usage>
</gpu_versions>
</app>

<app>
<name>milkyway_separation__modified_fit</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.01</cpu_usage>
</gpu_versions>
</app>

</app_config>
14) Message boards : Number crunching : Graphics card performances (Message 64149)
Posted 3 Dec 2015 by Profile khryl
Post:
if you want to upgrade your graphic card based on milkyway performance you may want to consider taking an amd card as those are far superior for this project
15) Message boards : Number crunching : Problem with running two milkyway work units (Message 64135)
Posted 29 Nov 2015 by Profile khryl
Post:
then he will have the problem that 2 workunits at once (with separation modfits) wont be enough to keep the gpu on full load, since he has the same card as i do. i need at least 4 simultaneously with modfits running to have my gpu at 99% (permanently)
16) Message boards : Number crunching : Problem with running two milkyway work units (Message 64133)
Posted 29 Nov 2015 by Profile khryl
Post:
i have the same problem on my 280x, they start finishing after 37-38 seconds, and after a while (sooner or later), they take 42 seconds because they somewhat synchronized again.

you could fix that by having run 3 simultaneously (less likelihood of having that happen), but if you want to stick to the 2 wu at the same time solution, i have no clue how to fix that
17) Message boards : Number crunching : MilkyWay@Home 1.02 vs MilkyWAy@home Separetion (modified fit) 1.36 (flops/award) (Message 64063)
Posted 5 Nov 2015 by Profile khryl
Post:
i m pretty positive that they do reward the same output as the normal ones.

as to the latter, i think you have to uncheck one more box a bit lower on the site that says If no work for selected applications is available, accept work from other applications?
if you uncheck that, you shouldnt be receiving them if you dont want to ;)
18) Message boards : Number crunching : MilkyWay@Home 1.02 vs MilkyWAy@home Separetion (modified fit) 1.36 (flops/award) (Message 64057)
Posted 4 Nov 2015 by Profile khryl
Post:
another approach: one single crunched modified fit takes 11 seconds on my r9 280x.

_5_ full seconds of those are only cpu work.
meaning that my gpu crunches only 6 seconds long. additionally, with only crunching one of those, my gpu is just barely above 30% load. i have to run 4 simultaneously so my card is on full load, while the times still stay between 12 - 17 seconds per wu.

on average, they pretty much give you the same output as the normal gpu workunits
19) Message boards : Number crunching : Run Multiple WU's on Your GPU (Message 63986)
Posted 11 Oct 2015 by Profile khryl
Post:
yea i know that.

i need 100% whenever i have milkyway running anyway so i just have 2 profiles,

1 with 100% fan speed for milkyway and the 2nd for other applications (games or the like) with 60-70%

but they only go with a cap of 3000 rpm, the ones from my 6850 went up to 5800rpm at 100%
20) Message boards : Number crunching : Run Multiple WU's on Your GPU (Message 63984)
Posted 10 Oct 2015 by Profile khryl
Post:
i use afterburner for overclocking and gpu z and hwinfo64 for temperature monitoring


Next 20

©2024 Astroinformatics Group