Posts by mmonnin
log in
1) Message boards : Number crunching : Can't get work for NVIDIA GPU (Message 67276)
Posted 25 Mar 2018 by mmonnin
I will mention that the developers are reducing work units to GPUs without very high double precision capability. Pretty much all of the NVIDIA GPUs mentioned in this thread have *very low* FP64 (double precision) capability, which would explain why you haven't been getting many work units. You should allocate your GPUs elsewhere.


Only GPUs without ANY FP64 capability and thus produce 100% errors. If it can complete work there's no reason to ban hosts.

Those that said they could not get work before have received work since.
2) Message boards : Number crunching : AMD GPU: Computation Error (Message 67265)
Posted 18 Mar 2018 by mmonnin
My GPU WUs always end like this: https://milkyway.cs.rpi.edu/milkyway/result.php?resultid=2287474945
I'm using Fedora 27 without proprietary fglrx driver.


That does not support double precision on the APU:
http://www.gpuzoo.com/GPU-AMD/Radeon_HD_7540D_IGP.html
3) Message boards : News : Testing Some New Plan Classes (Message 67260)
Posted 18 Mar 2018 by mmonnin
Hey BeemerBiker,

So the runs I put up a couple of days ago have a few extra calculations in them, so they should take a little bit longer. Your credits should have been adjusted accordingly.

Please let me know if you aren't getting an increase in credits for the increased work.

Jake


yea, probably will get more credit, thanks, but why are 4 of my Tahiti class GPUs running at 300mhz instead of 850 or 900. They are not being fed properly I suspect. Could be a problem at my end as I don't see anyone else reporting stuff like I am getting.


I have a 280x in Win10 running at 1070 mhz.
4) Message boards : Number crunching : DB crash and 400 "In progress" limit (Message 67259)
Posted 18 Mar 2018 by mmonnin
Even if it had double, 14*80 is more than what I saw. 1382 in progress now and very few completed.
5) Message boards : Number crunching : DB crash and 400 "In progress" limit (Message 67254)
Posted 17 Mar 2018 by mmonnin
Hi,

There seams to be a 400 "In progress" limit at this project right ?

1) I have changed my configuration and could never get more than 400 "In progress" WUs.
2) Milkyways database is often down some hours which make new WUs unavailable
3) My computer is quite fast ( 5 GPUs, 25 WUs at time )

When database is down my computer will finish all "In progress" WUs in one hour and I think it's the same for all computer having a "Recent average credit" around 2M.

I know the are some changes comming for database server ( 2nd CPU etc ...) but would it be possible to increase "In progress" limit for computer having a high "Recent average credit" ?
Decreasing "In progress" under 400 for low runing computer, using unadapted cards, may also help increasing validation time, and redistribute WUs, wouldn't it ?

Regards,

Raphael


How did you get 1400 tasks on a single 7 GPU system?
6) Message boards : News : Testing Some New Plan Classes (Message 67253)
Posted 17 Mar 2018 by mmonnin
Sounds more like a driver crash than a project problem. Even if load did go back up with collatz.
7) Message boards : News : Reverting Change to Remove Unreliable Hosts (Message 67238)
Posted 10 Mar 2018 by mmonnin
All or nearly all errors on these hosts.

https://milkyway.cs.rpi.edu/milkyway//results.php?hostid=512556
https://milkyway.cs.rpi.edu/milkyway//results.php?hostid=606991
https://milkyway.cs.rpi.edu/milkyway//results.php?hostid=191911
https://milkyway.cs.rpi.edu/milkyway//results.php?hostid=737902
https://milkyway.cs.rpi.edu/milkyway//results.php?hostid=718669
8) Message boards : Number crunching : there is no my name in the statistic (Message 67237)
Posted 10 Mar 2018 by mmonnin
I see it, 19th:
19 Grzegorz Roman Granowski 1,090,474 2,005,595,867 Poland 20 Jul 2009, 20:24:21 UTC

And by total credit:
6 Grzegorz Roman Granowski 1,090,474 2,005,595,867 Poland 20 Jul 2009, 20:24:21 UTC
9) Message boards : Number crunching : How to throttle AMD GPU load? (Message 67206)
Posted 5 Mar 2018 by mmonnin
Not that I know of. GPUs are designed to display output in real time so there's no type of way to run at 50% all the time. You can only run at 100% then 0, then 100%, etc. Some people use TThrottle from the same maker as BOINCTasks to keep temps in check by controlling usage.

https://efmer.com/tthrottle/
10) Message boards : Number crunching : CPU configuration for GPU apps (Message 67188)
Posted 2 Mar 2018 by mmonnin
Thanks for replying.
Also I think there are OS requirement to enable/use AVX. Not sure what the other users computer has for an OS.


My system is using Win7Pro and my friend's system is using Win10Pro. Both are consumer-grade OS.
I also believe the general consensus for the best performance for AMD Bulldozer and its derivatives is using SSE4.1 instruction set, not AVX. That's why I'm a bit loss here.


What you are asking for could be beyond MW's capability to do, they may not be able to only send AVX work to certain cpu's and only SSE4.1 stuff to non AVX capable cpu's.


It seems to be related to PCs with NV and/or new hosts. Minus periods where the database was down I've been able to get tasks on my 280x in Win10.
11) Message boards : Number crunching : CPU configuration for GPU apps (Message 67179)
Posted 1 Mar 2018 by mmonnin
Some projects send both AVX and SSE apps to the same PC. I have had both kinds of apps from several other projects.

Also I think there are OS requirement to enable/use AVX. Not sure what the other users computer has for an OS.
12) Message boards : News : Scheduled Server Maintenance 2/21 (Message 67178)
Posted 1 Mar 2018 by mmonnin
That's really what I meant, to move the conversation to another thread so it can receive the proper attention.
13) Message boards : Number crunching : XFX AMD HD6950 2Gb Install (Message 67172)
Posted 1 Mar 2018 by mmonnin
Mine always updates to WCG for some reason. Its a PIA even w/o using web preferences. Some projects through a fit since some don't have all of the locations.
14) Message boards : Number crunching : Not able to download work for AMD GPU (Message 67171)
Posted 1 Mar 2018 by mmonnin
I wonder if it's limiting connections to the db.
15) Message boards : News : Scheduled Server Maintenance 2/21 (Message 67170)
Posted 1 Mar 2018 by mmonnin
Immer noch keine Tasks für den Win 10 PC

Anscheinend legt hier keiner mehr Wert auf unsere Mitarbeit !


They said(according to Google translate):

"Still no tasks for the Win 10 PC

Apparently no one here attaches more importance to our cooperation!"

Computers are hidden so no one could be of help.
16) Message boards : Number crunching : XFX AMD HD6950 2Gb Install (Message 67153)
Posted 27 Feb 2018 by mmonnin
There have been some database issues. What does your event log say? There were also some aborted tasks and the admin has tightened down on errors.

https://milkyway.cs.rpi.edu/milkyway/results.php?hostid=765772
17) Message boards : News : Scheduled Server Maintenance 2/21 (Message 67151)
Posted 26 Feb 2018 by mmonnin
Last update, and in the hope of narrowing the problem:
-I again deleted milkyway, re-attached the project. The milkyway folder still being empty after that I pasted the contents of the milkyway folder from my working old host.

Same result. Still reporting 0 tasks.


Please split your issue into a new thread as it is not related to the Server maintenance.
18) Message boards : News : Scheduled Server Maintenance 2/21 (Message 67146)
Posted 26 Feb 2018 by mmonnin
Yes the database was unavailable for a while. And indeed my old host can now again get new tasks.
However this is entirely a different problem from the fact that NEW hosts seem not to be able to get either masterfile nor any task at all at the moment.
I really hope the admins can look into it...


Please post the 1st 100 lines of the log after starting the client in another thread.
19) Message boards : Number crunching : New Benchmark Thread - times wanted for any hardware, CPU or GPU, old or new! (Message 67138)
Posted 25 Feb 2018 by mmonnin
Hmm. If milkyway desires double precision and my nvidia 750ti is pretty limited on that compared to amd cards maybe my card is better served serving other projects. Ehh, the quick gratification getting 200 points in rapid succession is still nice.

Modestly my evga 750ti superclocked at stock gets these done in 665 seconds with an i5-4460 running at stock 3.2 GHz (one core free)

Yeah somewhere I just learned that amd cards have far more double precision compute power and that milkyway uses it. And so I came over here if there's anything about it. I now have more respect for amd hardware.

I have wanted to upgrade to an 8 core Ryzen. Didn't think I'd have reason to consider amd in the gpu slot (other than the driver interface looks pretty nice). Too bad modern amd graphics cards are pretty much all gone. Doesn't seem like that's gonna get much better soon.


Older AMD cards have better double precision, FP64. The 7970/280x generation or any derivatives. All newer consumer cards from NV and AMD have poor FP64 performance including that maxwell 750Ti. That's left to prosumer or above cards like the Nvidia Titan Volta for $3k.
20) Message boards : News : Scheduled Server Maintenance 2/21 (Message 67136)
Posted 24 Feb 2018 by mmonnin
I think this says BOINC client sees CPU and AMD cards:

24/02/2018 15:57:41 | Milkyway@Home | Requesting new tasks for CPU and AMD/ATI GPU

But preferences say you are only set to receive NV tasks:
24/02/2018 15:57:42 | Milkyway@Home | Tasks for CPU are available, but your preferences are set to not accept them.
..so nothing is downloaded.

766764 has tasks but 766741 has nothing. Is Win10/766741 the one with the above line about NV tasks?


Next 20

Main page · Your account · Message boards


Copyright © 2018 AstroInformatics Group