Formula Boinc 2017

Yes, well done all and congrats to EVGA.

Thankfully TAATs big bunker fell short of the mark, giving us a well deserved podium place.

Just two sprints left now this year.
 
@MGP

Yeah I had f@h already running before I saw a thread in GPU section asking for members to chip in. Cheers for info :) .

@andy_taximan

Cheers, it's a recent build :) , mostly overkill by a long yardstick!

Yep had joined OCuk team :) . I hadn't fathomed last night how to setup GPU to be used, so it only ran "Generalized..." on CPU :o . I have enabled today and still no go, understand "Sprint" is over, but will aim to read up/play with this so am ready for next time :) .
 
Welcom gupsterg, your Rig 1 is similar to mine, excepts mines got 32GB and a 1080Ti, but I do have a RX Vega 64 in my other rig :)
 
gupsterg said:
@MGP

Yeah I had f@h already running before I saw a thread in GPU section asking for members to chip in. Cheers for info :) .

@andy_taximan

Cheers, it's a recent build :) , mostly overkill by a long yardstick!

Yep had joined OCuk team :) . I hadn't fathomed last night how to setup GPU to be used, so it only ran "Generalized..." on CPU :o . I have enabled today and still no go, understand "Sprint" is over, but will aim to read up/play with this so am ready for next time :) .
Click to expand...

We like overkill rigs in here, puts them to good use and see if they stable lol, if you have any problem setting it up would take any 1 of us seconds to do if you use team viewer or similar? or we can just answer questions the hard way :)
Feel free to carry on with Primegrid or finish those and help out on milkyway or einstein whatever you fancy
http://www.primegrid.com/results.php?hostid=912090&offset=0&show_names=0&state=4&appid=
http://www.primegrid.com/team_members.php?teamid=198&sort_by=expavg_credit&offset=0

these show how you did/are doing
 
EdwardTobia said:
how do i help?!
Click to expand...
gupsterg said:
@MGP

Yeah I had f@h already running before I saw a thread in GPU section asking for members to chip in. Cheers for info :) .

@andy_taximan

Cheers, it's a recent build :) , mostly overkill by a long yardstick!

Yep had joined OCuk team :) . I hadn't fathomed last night how to setup GPU to be used, so it only ran "Generalized..." on CPU :o . I have enabled today and still no go, understand "Sprint" is over, but will aim to read up/play with this so am ready for next time :) .
Click to expand...
although the sprint is over, EVGA are about to overtake us in the overall boinc league , so if you fancy it then stick around and keep on running primegrid :D

also theres many cpu projects that you could take advantage of , you have some nice cpus and could be a nice addition to team OcUK :D:D
 
@Biffa

Cheers for the welcome :) . Sweet rigs :) .

@andy_taximan @ozaudio

No problem will do more ASAP :) . I plan today to swot up on configuring client to use GPU ;) . I don't use teamviewer, etc and prefer to get to know how to use the client. If I do get stuck will let you guys know for sure :) .

Yesterday some testing on VEGA with differing apps revealed pushing HBM clocks too far yielded performance loss. So once I have assessed a decent tweaked setup I wish to throw Bionic at it. In the past f@h (plus 3D loads) helped gain nice tweaked setup on GPUs, now Bionic will also be used in process :) .
 
gupsterg said:
@Biffa

Cheers for the welcome :) . Sweet rigs :) .

@andy_taximan @ozaudio

No problem will do more ASAP :) . I plan today to swot up on configuring client to use GPU ;) . I don't use teamviewer, etc and prefer to get to know how to use the client. If I do get stuck will let you guys know for sure :) .

Yesterday some testing on VEGA with differing apps revealed pushing HBM clocks too far yielded performance loss. So once I have assessed a decent tweaked setup I wish to throw Bionic at it. In the past f@h (plus 3D loads) helped gain nice tweaked setup on GPUs, now Bionic will also be used in process :) .
Click to expand...

app_config.xml is the way to go for gpu, heres the 1 i use for milkyway
<app_config>
<app>
<name>milkyway</name>
<gpu_versions>
<gpu_usage>0.25</gpu_usage>
<cpu_usage>0.05</cpu_usage>
</gpu_versions>
</app>
</app_config>

gives 4 wu on each gpu
 
gupsterg said:
@Biffa

Cheers for the welcome :) . Sweet rigs :) .

@andy_taximan @ozaudio

No problem will do more ASAP :) . I plan today to swot up on configuring client to use GPU ;) . I don't use teamviewer, etc and prefer to get to know how to use the client. If I do get stuck will let you guys know for sure :) .

Yesterday some testing on VEGA with differing apps revealed pushing HBM clocks too far yielded performance loss. So once I have assessed a decent tweaked setup I wish to throw Bionic at it. In the past f@h (plus 3D loads) helped gain nice tweaked setup on GPUs, now Bionic will also be used in process :) .
Click to expand...
if you want something to test you gpu and give it a work out then a CFG file to run more than one work unit at a time like andy said above will do that no problem :D

for example on primegrid I only run the work units that give max points so this is what I do.... go to your primegrid account below and untick the cpu box at the top so it uses GPU only, then scroll down and make sure that the box Proth Prime Search Sieve (PPS-Sieve) is ticked for cude / openCL . then scroll to the bottom and click update pref
http://www.primegrid.com/prefs_edit...9563463&ttok=0b1a10cde7fdb3ec7311b8b861df674c

you can also edit / make a cfg file aswell by doing the following

C:\ProgramData\BOINC\projects\www.primegrid.com

once in that folder make an empty text doc file and inside it you paste the following below

<app_config>
<app>
<name>pps_sr2sieve</name>
<gpu_versions>
<gpu_usage>0.33</gpu_usage>
<cpu_usage>0.2</cpu_usage>
</gpu_versions>
<fraction_done_exact/>
</app>
</app_config>


the 0.33 in gpu changes how many units at once you run, so 0.25 for 4 at once or 0.5 for 2

you then need to save the file with the name as app_config.xml but also you need to change the file type as " all files " then click save. go into your boinc client and click view > advanced view, i noticed you only have yours in simple view so this opens upo the client to a more advanced view, then go to options and click read config files. primegrid will then run 3 work units at once if you was to follow the above cfg i pasted from my own folder
 
Last edited:
i have to, it must be PCI lane related on the 32 core opteron system, otherwise the GPU wont load to 100% its the same for Einstein / milkyway aswell
 
MDPlatts said:
I'm baaaaack !!

Resurrected accounts for MW@Home (WU :D) and Leiden Classical (no WU :() on Ryzen 1800X @ 4GHz - might need to look into something for Vega to chew on too.

TJC
Click to expand...
gupsterg said:
@MGP

Yeah I had f@h already running before I saw a thread in GPU section asking for members to chip in. Cheers for info :) .

@andy_taximan

Cheers, it's a recent build :) , mostly overkill by a long yardstick!

Yep had joined OCuk team :) . I hadn't fathomed last night how to setup GPU to be used, so it only ran "Generalized..." on CPU :o . I have enabled today and still no go, understand "Sprint" is over, but will aim to read up/play with this so am ready for next time :) .
Click to expand...
our next challenge starts tomorrow, its CPU only and you would do well at it, universe@home, same as primegrid youll need to sign up and make an account if you fancy it :). infact bith of your systems would do even better :D
 
Last edited:
Universe @ home. hmmm. Looks like options for CPU and ATI GPUs, but not Nvidia GPU. This could make things interesting as most of the GPU sprints so far favoured Nvida users. This could bring a different mix of power to the teams.

Oh and it's a project where each WU needs a wingman to correlate results (and thus points). Adds to the lottery factor as to when or whether you get your credits in time.
 
ATi GPU work is an option in the project preferences? They've no doubt given up on that application then.
 
It is an option for testing that started almost 2 years ago but hasn't resulted in a production app yet.
It was the primary reason why I stopped pushing on that project with my CPUs, I put them onto other projects expecting to be able to make up the points when the GPU app appeared :rolleyes:
 
Good luck, gents. A good deal of TAAT power is still building bunkers for the upcoming WCG b-day challenge hosted by SETI.Germany. As the temperatures drop, your rigs are firing up, and for a cool weather CPU race, we really don't stand much chance for 1st or 2nd this time around. Less than an hour until the start, hope you are all getting some sleep, and let's go do some science! (instead of theoretical math!!!!!!)
 
ye I did notice tony that you wasn't bunkering after checking your systems on universe, then seen a post over on another forum. you have a lot of cpu over at TaaT and should do pretty well at WCG :)
 
Anyone would think we want to win this one :D

OK it's early hours with less than 12 hours gone of the 3 days. Plenty of time for others to unleash bunkers albeit that it's challenging building a good bunker as Universe does seem to have limits on the number of WUs you can download per machine.
1 OcUK - Overclockers UK 2,547,100
2 Crunching@EVGA 810,233
3 Overclock.net 673,500
4 TeAm AnandTech 444,700

That's particularly impressive as despite our being in league 2, we'd be in Silver position amongst the big boys of league 1, only Seti.USA beating us. :cool:
 
You must log in or register to reply here.
Back
Top Bottom