My Apologies To The Einstein Crunchers
8 Nov 2010 12:48:15 UTC
Topic 195439
(moderation:
Greetings Einstein Crunchers,
Last month, October, I decided to re-attache to Einstein@home since SETI@Home is going through, what I would like to call, a spurt of growth. Many of you know that SETI is in the process of receiving 2 new servers to better serve the project and the thousands of volunteers faithfully crunching SETI WUs.
Not long after I started crunching Einstein on my i7 PC, I noticed some strange anomalous behavior. I would see WUs that were "Waiting for memory". I would also notice that after some time, I would see that text was missing from things like web pages and even things on the i7 such as the Windows start button would no longer say "Start" and tabs in my browser and HTML editor would also be blank. When shutting down a program, it would take a couple minutes for the window to go away. Programs would no longer start. BOINC would no longer make a connection to localhost after restarting it. The i7 re-booted all on its own once. I immediately suspected Einstein.
Now, lets go back in time to May (I believe it was) when I built this i7 PC. I noticed strange things happening with my video display. My mouse cursor would suddenly jump from one point to another across the screen. Programs, like my HTML editor, would auto scroll while I would be doing my coding. When on a website, I would be reading a page and suddenly be transported back a page or 2 as if I hit the back button, which I hadn't. I suspected that there was something flakey about my new PCI-E video card. I lived with it for several months.
When I saw the problems I started having, when starting Einstein again, I figured, even though I still suspected that Einstein was at the root of the current problems, that my video card was getting worse. So, a couple days ago, I bought a new one and installed it. The problems I was seeing, before Einstein, went away. Good call on that one, the "old" card was indeed flakey. But, I was still observing the problems that manifested after starting Einstein again. My suspicion of Einstein was getting greater.
I mentioned these problems in another thread, here, and a few members gave me some suggestions to try to alleviate my problems. A few of them had no affect, but, I did get a definitive explanation for my problem. It was indeed Einstein that was the cause.
But, my i7 was also at fault, sort of. I'm not running a 64 bit OS, so WinXP Pro is not seeing all of my 4GB of RAM. Einstein uses a great deal of RAM, due to the size of the WUs, and perhaps other factors as well. Since my 32 bit OS is only seeing 3GB of RAM and 8 Einstein WUs are taking up 2 of those GBs, the i7 was left with 1GB for everything else including BOINC.
And now for the apology: About 45 minutes ago, give or take, I checked on the i7 and BOINC. You see, after about 8 to 10 hours of continuous running, BOINC would start creating havoc with the i7 and I would have to re-boot the PC. When I checked on the progress of BOINC, I saw that 175 WUs were reporting "Computational error"(s). I had 5 completed WUs waiting to report. Suffice to say, there probably are many of you out there waiting for the results of those 175 WUs, and for that I do apologize.
For the time being, I have resigned myself not to run Einstein. As a matter of fact, BOINC is sitting idle right now with nothing to do. I do plan on upgrading to a 64 bit OS (WinXP Pro, not Win7). I may even decide to double my RAM to 8GB. Maybe then I will start crunching Einstein again. But, right now, I do not want what happened this morning to happen again. It is NOT fair to the rest of you waiting for results so you can get the credit for the work you have already done.
My thanks go out to Mikey, Gundolf and DanNeely for the help they gave me. I really appreciate it! :)
And now, it's time to do my normal routine(s), now that the i7 is free to do my bidding, once again. ;)
Have a great day, everyone! :)
Keep on BOINCing...! :)
CAPT Siran d'Vel'nahr XO
USS Vre'kasht NCC-33187
Siran's website: [ ONLINE! ]
My Apologies To The Einstein Crunchers
)
What about reducing the number of concurrently running BOINC tasks, instead of stopping altogether?
Set "On multiprocessors, use at most __% of the processors" to 50% (25%, 13%) to use only 4 (2, 1) out of your 8 cores. That should reduce the memory burden enough so that you can continue crunching without problems.
And don't forget that running CUDA tasks can also affect the graphic performance up to (seemingly) freezing the screen.
Gruß,
Gundolf
Computer sind nicht alles im Leben. (Kleiner Scherz)
Hi Siran, error message is
)
Hi Siran,
error message is always:
Seems sth. is messing with your downloads, firewall, antiv virus. Try to disable/exclude BOINC and try again!?
Michael
Team Linux Users Everywhere
RE: RE: -[ snip ]- What
)
Greetings Gundolf,
Sorry, but that would defeat the purpose that I built my i7 for to begin with.
My new video card will probably handle CUDA WUs better than the defective one did, but that was not why I got the card. I don't do CUDA, yet.
Another thing that kinda disturbs me is the amount of storage Einstein requires and is currently using. When I had a full compliment of WUs, about 215 or so, Einstein took up almost 2 GB of space. Granted, I have a 500 GB drive, but it's the fact that with all the WUs gone, Einstein is still taking over .5 GB. SETI is taking only 31 MB, Orbit 0(zero) and Virtual Prairie, which I just attached to, 484K. Seems to me, unless there is some underlying logical reason for the over sized WUs, Einstein could reduce the size of them and not put so much stress on users PCs. In my opinion, it's arrogant to assume that EVERYONE has a PC that can handle the work they bulldoze out to us. Or that users will (can) just go out and get what's needed to upgrade their PC to handle the load from Einstein.
When I ran Einstein, in the past, I do not remember having this problem with their WUs. And back then I was running a single core with HT and I doubt even 1 GB of memory.
Einstein will remain on hold, but attached, until I decide on whether I will "upgrade" my i7 to better handle the load or not.
Thanks for the suggestion anyway, Gundolf! :)
Keep on BOINCing...! :)
CAPT Siran d'Vel'nahr XO
USS Vre'kasht NCC-33187
Siran's website: [ ONLINE! ]
RE: Hi Siran, error
)
Greetings Michael,
Well, at this point in time, I believe that error message to be a moot point. That message was on one of the 175 WUs that whacked on me this morning. Up until that point, I had no problem getting, crunching, uploading and reporting WUs as long as I kept a close eye on the i7 and managed to "refresh" it before something really drastic happen like that which happened this morning. I have never had a conflict problem between BOINC and my firewalls or anti-virus software. And, I have both a hardware and software firewall.
Thanks, though, for bringing this to my attention anyway. I truly doubt I need to deal with that error.
Have a great day, Michael. :)
Keep on BOINCing...! :)
CAPT Siran d'Vel'nahr XO
USS Vre'kasht NCC-33187
Siran's website: [ ONLINE! ]
Einstein is taking me 512 MB
)
Einstein is taking me 512 MB of disk space. I have two 160 GB disks inside my workstation and an external disk of 1.4 TB used as a backup disk. Storage is rather cheap today. My RAM is 5 GB out of a possible 8 GB. I am using a 32-bit Linux which is pae and could use up to 64 GB RAM. I am running 3 BOINC projects on my Opteron 1210 with two cores (AQUA, Einstein and QMC). Other 4 projects (SETI, QuantumFIRE, CPDN and LHC are not giving me work at this moment, for various reasons). My last CPDN job took 4600 hours before erroring but, due to the trickle method, I did not lose all my credits.
Tullio
Hi Siran! I can only tell
)
Hi Siran!
I can only tell something from my own experience:
- The 860 is known to be running a bit hot, hope u have a sufficient cooling solution on the chip.
- i looked on the asus webpage about your P7P55D-E PRO, i see only a 4 pin power connector for the cpu. For normal use that is no problem, but for crunchers, its nearly on the edge. i take 4 pin boards only for i5 chips, not for i7. i get the i7 only stable with a 8 pin power supply connector.
With the other problems, dont know, can be all or nothing, for my own, i think, top modern hardware needs an OS wich can handle the new technology. On nearly 10 year old OS like XP, its nearly obvious to see some unwanted effects.
Hope, my brainstorming helps you a bit further.
Greetings
Chris
RE: Hi Siran! I can only
)
Greetings Chris,
Here are my i7 SPECIFICATIONS:
Intel i7-860 2.80 GHz CPU
Corsair H50 CPU cooling system
Asus P7P55D-E Pro MoBo
Galaxy nVidia GeForce GT 430 PCI-E 1MB GDDR3 RAM (CUDA Enabled)
Kingston 4GB PC3 10600 DDR3 RAM (Dual Channel)
Western Digital 500GB 3G SATA HDD x 2 = 1TB Storage
Micro$oft WinXP Pro 32 bit
BOINC v6.10.58
Yes, I know about the 860 running hot. :( That is why I installed the Corsair cooler for it.
According to my motherboard manual, and what I saw physically on the board, if the 8 pin connector you speak of is for ATX power, not to be confused with the 24 pin ATX power connector, then yes it is 8 pin. But, I only have a 4 lead cable/4 pin connector running to it. 4 of the connector's pins, on the board, have a block-off plate covering them. The manual does not specify removing it, just that if the 8 pin is not connected, then the PC will not boot. Well, it boots, no problem. :)
I have a modular PSU. I'm not sure if it has another port for a 4 pin ATX connection or not. If it does, I may have to use the Google machine and see what I can find on the Internet about my motherboard. Maybe someone else asked about the 8 pin connector on the board, and the cover over the 4 pins.
Another thought, I have a newer modular PSU and looked at my unused cables. I do have an 8 pin to 6 pin cable and an 8 pin to 6/8 pin cable. Perhaps the 8 pin connection came out after the PSU was built that is currently powering my i7. Perhaps it is time to get a new PSU for the i7 as well. Scratch those cables! Upon further investigation (looking at the specs on the PSU box), those cables are for PCI Express. It does show that an 8 pin to 8 pin cable is part of the inventory, but it is not in my spares. Perhaps it is being used in my Linux PC that the PSU went into.
So, what you're saying, about my OS, is that it's time I upgrade to Win7 Pro 64 bit? ;) Well, since there are things I do that I cannot do with the Linux PC, I guess the upgrade is inevitable, huh? :)
Thanks for your "brainstorming" Chris! Gave me more to think about. :)
Keep on BOINCing...! :)
[edit]
edited for clarity...
[/edit]
CAPT Siran d'Vel'nahr XO
USS Vre'kasht NCC-33187
Siran's website: [ ONLINE! ]
RE: then yes it is 8 pin.
)
That 4+4 pin power connector has parallel connections in the two halves. So populating both halves reduces resistance in the power input to the board, which lowers heat (power dissipated) at the connection and improves onboard supply voltage and reduces variation. As with the "optional" portion of the wide ATX power connector, which also has extra power connections in parallel with the main part, it is not surprising to have the system run with it not populated, but you are better off hooking it up if you can.
My i7 motherboard had stickers over both of these optional sections--I just took them off as I inserted the proper inputs from my power supply.
RE: Greetings Chris, Here
)
Yes, sry, was my fault, was time to clean up my TFT, but for being sure, i DLed also the manual :-)
like archae86 said before, try to populate them, its better.
I dont see your PSU on your Component-list; so following only guessing:
If you take your equipment and let the power consumption calculate on http://support.asus.com/PowerSupplyCalculator/PSCalculator.aspx, a minimum of 550 W is required. Imho some safety adding, you need to have something over 600 W.
For the OS: I can only say, W7 64 bit is really nice to have, but main point are the 64 bit. Even with XP. The memory border with 4 GB is not really up to date with todays performance monsters.
Hope my english bubbling make some sense :-)
Chris
You're definitely seeing
)
You're definitely seeing paging / swapping to disk. It's normal that this brings any HDD-equipped system to a crawl.
There's a far easier solution than restricting the amount of concurrent threads: restrict the amount of main memory used by BOINC to ~33% (=1GB) or whatever number you like (is probably 50% now) and attach to at least one other project which uses less memory. Personally I like QMC: computationally intensive, stable work supply, good credits, interesting science and German ;)
You may need to adjust the ressource share, so that BOINC doesn't insist on trying to run 8 Einsteins and keeping 4 of them "waiting for memory" forever. Not sure, though.
MrS
Scanning for our furry friends since Jan 2002