Posts by Mr. Kevvy


log in
21) Message boards : Science (non-SETI) : The computer that crunches cloud data to heat your home (Message 1640696)
Posted 12 Feb 2015 by Profile Mr. KevvyProject donor
I've been doing this for years... turned off all the radiators in our apartment and it's been heated by computers ever since. Not pleasant in the summer though. :^)
22) Message boards : Number crunching : >1 Task per CPU? (Ans: Probably Not) (Message 1639610)
Posted 10 Feb 2015 by Profile Mr. KevvyProject donor
How did you measure % utilization on the GPU?


EVGA Precision. Essential in the cruncher toolkit. :^)
23) Message boards : Number crunching : >1 Task per CPU? (Ans: Probably Not) (Message 1639606)
Posted 10 Feb 2015 by Profile Mr. KevvyProject donor
However doing so will likely have little to no gains. More than likely it would just result in less work done in a given time if gone to far. Much like I suspect is likely happening with the iGPU running 5 tasks at once.


This is pretty much it. CPUs run BOINC code natively so are always very close to 100% utilization; there's no need to run >1 instance per core.

Conversely, GPUs have to be loaded with work by the CPU; you can see these helper processes feeding work to the GPU in Task Manager->Processes when you have BOINC GPU work running. If they aren't allocated enough due to the CPU core running the process not getting enough timeslices, being too slow, or due to bus overloading, the GPU will run at less than 100% utilization. In this case, running >1 instance per GPU does have significant gains. When I got my first 980, I tested it. With 0.4 cores reserved per GPU thread, I found with 1 instance on the GPU there was 45-55% utilization, with 2 was 70-75% and 3 is 95-100%. So I stuck with 3. Anything more than this is just going to slow it down and waste memory, as the GPU is always maxed out and then other factors such as bus oversaturation can occur. If someone has a slow CPU and a fast GPU, 5 instances could actually be worthwhile, if 4 are tested to be consistently and significantly less than 100% utilization, but this would usually be unlikely.
24) Message boards : SETI@home Science : ET, is that you? Scientists launch plan to beam welcome messages to far-off star systems (Message 1639573)
Posted 10 Feb 2015 by Profile Mr. KevvyProject donor
A statement on this just went on the main page with an associated thread. I recognize three of the signatories are from this project (Andrew P. V. Siemion, Eric J. Korpela, Dan Werthimer).
25) Message boards : SETI@home Science : ET, is that you? Scientists launch plan to beam welcome messages to far-off star systems (Message 1639388)
Posted 9 Feb 2015 by Profile Mr. KevvyProject donor
Agreed, but a century is such a short time, astronomically speaking.


It is, but it's nigh-impossible for us to ever make those signals unbroadcast, so probably too late to worry. That ship has sailed, barn door horse etc.
26) Message boards : SETI@home Science : ET, is that you? Scientists launch plan to beam welcome messages to far-off star systems (Message 1639382)
Posted 9 Feb 2015 by Profile Mr. KevvyProject donor
If there is an alien species out there who runs contrary to the "default position of the egghead set"* and is both technologically sophisticated enough that we have no chance of defeating them, and are still merciless usurpers of other species' planets, then it's too late already to cower in the dark, as they would also be sophisticated enough that our leakage of everything from morse code crackles to thousands of simultaneous digital video broadcasts has been whetting their appetites for a century.

*Contact, 1997
27) Message boards : Number crunching : Will switching from 32-bit OS to 64-bit OS help? (Message 1638046)
Posted 6 Feb 2015 by Profile Mr. KevvyProject donor
Thanks for the explanation. :^)

32 bit ? well 4Gig Video RAM equipped cards need more than that, simple as that really.


They are all 2048MB or less, being Beaters and all.
28) Message boards : Number crunching : Will switching from 32-bit OS to 64-bit OS help? (Message 1637978)
Posted 6 Feb 2015 by Profile Mr. KevvyProject donor
Now that I think of it, there are two problems that I constantly and regularly used to have with the "Beater Crunchers" with XP that I have yet to experience once since reimaging them with 7 (32-bit for both) and that did actually negatively affect crunching performance. Your mileage may vary of course... personal experience only:

First, every few weeks, they would exhaust some sort of system resource (I don't think it was memory... handles perhaps?) and every work unit would crash, menus wouldn't display properly, and I'd even get "Access Denied" trying to reboot. This was a definite detriment to crunching as it would lose all the work units in progress (very sad with the AP ones with >10 hours on them) plus being unable to do anything more and sitting idle if this wasn't caught. It happened on all of the XP boxes and pretty regularly. Of course as luck would have it the cycle time was just shy of the monthly "Patch Tuesday" which is the only time they were ever rebooted, and mostly then just to power down to blow the dust out of them. So my long uptimes with heavy processing was exposing some bug that most others would never see.

Secondly, every few days, the wireless LAN/WiFi drivers would fail and the machine would no longer be able to connect to upload/download work until the connection was disabled and re-enabled. This also happened on all of the WiFi XP boxes, though all of their WiFi cards and thus drivers were different. Again, long uptimes were probably exposing a bug that shorter sessions would obscure.

Of course, in a perfect world, these bugfixes would have just been rolled into XP and it would have continued on its merry way. But Windows isn't Linux, and Microsoft eventually decides that the bugfixes are going to go into the newer OS to encourage paid upgrades and because fixing an old OS just isn't as profitable.
29) Message boards : SETI@home Staff Blog : What does loss of net neutrality mean for volunteer computing? (Message 1637414)
Posted 4 Feb 2015 by Profile Mr. KevvyProject donor
FCC Chairman Tom Wheeler: This Is How We Will Ensure Net Neutrality

The important bit:

That is why I am proposing that the FCC use its Title II authority to implement and enforce open internet protections.

Using this authority, I am submitting to my colleagues the strongest open internet protections ever proposed by the FCC. These enforceable, bright-line rules will ban paid prioritization, and the blocking and throttling of lawful content and services. I propose to fully apply—for the first time ever—those bright-line rules to mobile broadband. My proposal assures the rights of internet users to go where they want, when they want, and the rights of innovators to introduce new products without asking anyone’s permission.


Yay! :^)
30) Message boards : Politics : The Venus Project-The Choice Is Ours (Message 1636993)
Posted 3 Feb 2015 by Profile Mr. KevvyProject donor
@Sarge (et al): Just to explain my now oddly out of place post, the person's name that prompted it has changed, so it may be disregarded. :^)
31) Message boards : Number crunching : Will switching from 32-bit OS to 64-bit OS help? (Message 1636752)
Posted 2 Feb 2015 by Profile Mr. KevvyProject donor
XP-32 drivers certainly exist for the first round of mid-range Maxwells - I'm running three GTX 750 Ti cards in old hosts upgraded from 9800 GT doorstops (four times the productivity at half the power).


Heh... didn't even know they were Maxwells (thinking of the 9x0 series.) And yeah the old 9800s are doorstops now. Have two of them in the closet as they aren't even worth the power and the PCIe slot.
32) Message boards : Politics : The Venus Project-The Choice Is Ours (Message 1636668)
Posted 2 Feb 2015 by Profile Mr. KevvyProject donor
33) Message boards : Number crunching : Will switching from 32-bit OS to 64-bit OS help? (Message 1636349)
Posted 2 Feb 2015 by Profile Mr. KevvyProject donor
I do hope to catch up with the rest of you one fine day.


<<Looks at Mark's RAC>>

Um, I think it's the other way around. ;^)
34) Message boards : Number crunching : Will switching from 32-bit OS to 64-bit OS help? (Message 1636347)
Posted 2 Feb 2015 by Profile Mr. KevvyProject donor
For crunch only rigs, there is only XP as the best option...


There are no XP NVidia drivers (32 or 64-bit) that will support any of the Maxwell cards. (Pretty sure of this... please let me know if there's any out there.)

I know that this planned obsolescence is done on purpose with hardware and OS companies (OK, who am I fooling... there is only one "OS company": Microsoft... the other OSes are free with the hardware or to download) symbiotically dropping support for older products to force upgrades but, well, there it is.
35) Message boards : Number crunching : Lots of invalid results? (Message 1636095)
Posted 1 Feb 2015 by Profile Mr. KevvyProject donor
I'm happy to report, for this problem at least, that the problem is definitely solved and it was an insufficient power supply/PSU.

This machine which generated well over a hundred Invalid results a day on its GTX980 (until throttled) as of this writing hasn't generated a single one in two days on a new PSU, moving from a no-name 680W PSU to one of these.

Note the NVidia GTX 980 specs.:

Thermal and Power Specs:
Graphics Card Power (W) 165 W
Minimum System Power Requirement (W) 500 W


For most if not all 500W PSUs for anything but casual gaming, this is underrated, as the quality of power they are going to deliver is going to be poor close to their limits (read this article on PSU Ripple for an example... as Mark so appropriately puts it they "cough up a furball" on your work units.) So if you are having this issue I would suggest getting a better PSU. Do some research on the make/model and of course: you get what you pay for. There are numerous quality brands and models (for example I've had very good performance for years and recently with 2x980s on this PSU) but two brands that seem to be mentioned the most are Corsair and Seasonic, for consistency at least (some makers have good high-end PSUs but their lower-end are substandard.)

(If you can't afford a new one, you can get a good used one (scored two excellent ones this way including the RM850) by being a "smart shopper" and following a few common-sense guidelines: buy from a seller in their residence not from "meetups", make sure it comes with a sales receipt to allow RMAing if it fails and show when it was bought, which also shows that it isn't stolen property... hot PSU in any sense is bad! (original box and all packaging/cabling is good too), bring a PSU tester (I paid $12 for mine years ago and it's been a very useful investment) or demo it running a system under load, should be clean/scratchless and of course not make any odours!)

And as noted earlier, for some reason, MB (multibeam=standard SETI) CUDA work is the only thing that has pushed the GPU/PSU over the threshold and caused this effect, on both stock and anonymous/Lunatics platforms. Astropulse CUDA doesn't seem to and neither did any Einstein@Home CUDA application. MB CUDA is the vast majority (95% or more) of what our GPUs will be doing here of course.

Hope this helps. :^)
36) Message boards : Number crunching : Main cruncher dead (Message 1634565)
Posted 29 Jan 2015 by Profile Mr. KevvyProject donor
Motherboard: MSI Z87 MPOWER MAX AC


If it does turn out to be the motherboard, please give this a read. I will personally never buy an MSI product again.
37) Message boards : Politics : Your patience levels? (Message 1634162)
Posted 28 Jan 2015 by Profile Mr. KevvyProject donor
Discovery of Extraterrestrial Intelligence - 15 years (and counting) :^)
(But I don't count not being a Brit. Commonwealth with the Queen as head of state close enough?)
38) Message boards : Number crunching : Lots of invalid results? (Message 1628225)
Posted 16 Jan 2015 by Profile Mr. KevvyProject donor
I think I have part of the answer here. It's just a matter of the server not taking no for an answer. I am the 8th one to get this work unit and in about 2 days it will be crunched and I will get no credit for it.


I think this is a different issue... that one is a bunch from one day. This issue is Maxwell 970/980 cards generating 80% invalids from any time.

I guess the MB CUDA just stresses the GPU more than AP or the Einstein code.
39) Message boards : Number crunching : Lots of invalid results? (Message 1628221)
Posted 16 Jan 2015 by Profile Mr. KevvyProject donor
OK, I moved both 980s to my main machine. In doing so I noticed the elephant in the room that seems to have been overlooked:

Power.

I had one on a 525W named PSU and another on a 680W no-name, but I bet neither one could deliver the sustained ripple-free 14A the cards need. (Also NVidia has really pushed the reduced power requirements of the Maxwells... maybe to the point of "exaggeration.") Now that they are on the 1000W Azza Titan with the 86A 12V rail, all the work units are completing.

What make/model of PSU do you have Ryan?
40) Message boards : Number crunching : Lots of invalid results? (Message 1627514)
Posted 14 Jan 2015 by Profile Mr. KevvyProject donor
how much CPU are you giving each MB?


0.4 on both. GPU util. usually about 78-80%.
(Edit: running with count=0.5 as they aren't very powerful.)


Previous 20 · Next 20

Copyright © 2015 University of California