Adding the -smp flag makes the SMP client unhappy and refuses to start up. Something about not connecting to my machine on a port being that mine actively refused it.

Have to head to work, so I'll have to figure this out later. :thumbsdown:

And, now they gave my laptop one of those big A3 workloads - P6041 this time, not 6040 like before. It's taking this laptop 50 minutes per frame, argh. Luckily the deadline is in 6 days from now.

The 6041 is a little quicker than the 6040, but still a beast on anything but a fairly fast quad. On a stock Q6600 the frame time is 33 seconds lower- from 25:39 to 25:06.

I finally managed a good OC on the new i7-930. I ran with a 3.2 for about a week, and tonight got it to 4.0. The PPD on a 6015 went from 12,400 (3:51) to 17,700 (2:57). Temps are mid 70's. The heatsink, http://www.frostytech.com/articleview.cfm?articleID=2481, is a monster and prevents adding a fan to the side panel of my Antec 1200. I'm running it with 2x Enermax UCMA12 (70CFM) in push/pull at constant full speed. I may need to find higher flow fans, but I like these because they're very quite- <18db.

I still can't get a good OC on the Q9550 (EVGA nForce 780i SLI FTW). It's at 3.0 and likes to lock up the system. I can't seem to get a voltage adjustment to "take". Maybe I should have stayed with my trusty Gigabyte- although I lost my OC (3.4) on the Q6600 in a P35 Gigabyte board, and can't get that back, either.

The heatsink on my i7-860 is a Noctua NH-U12P SE2. It's not quite up to Megahalems level but it's still pretty good and I have 2 low-rpm fans installed on it. Now I have the cpu clocked to 3.57GHz (170MHz base clock) and the temperatures are around 63-70C on the cores. It's still perfectly stable but I think I'll keep it at this setting maximum, because my RAM is not specified to run any higher on this multiplier. I'm quite happy with the performance and stability. The motherboard is a Gigabyte P55 one.

Man, I'm getting sick and tired of the ATI GPU "Missing work files" after I stop it to do other things then restart it. I've lost 80%+ 3 times this week. :facepalm:

could give qfix a try to salvage parts of broken wu?s, admittely only used it occasionally my self on smp-work(it is inluded in one of the client-vm?s).

tried higher oc on a be 940,ended up on 3.5 ghz since it was issues with higher settings,and I?m only use air-cooling (zalman).
well,wish me luck on my intel endevaur mainly to get more gpu?s into the fold,hopefully I get it set up before the weekend,and no,only "cheap" stuff-intel,no i7 or anything

The problem is it deletes the old units then downloads new stuff before I even know what is going on.

My 5870 fan also started to do this very loud annoying rattle yesterday after a reboot. The fan sounds like it's hitting the shroud. Tapping will make it stop for a second or two but it comes back. Then it just stops as suddenly as it started. It did this 2 months ago to, (Again, after a reboot.) and was fine from then til yesterday when it reared it's annoyingly loud head again. It's fine at this point in time but I have a ticket in with XFX anyway. My concern, is this is a known issue with the 5xxx series. I do not want to pay to RMA it, just to get the same issue. (Though I would hope I wouldn't, the possibility is there.)

uhm you mean you get a message about deleted wu?s and it continues after getting new work and all ?
not sure it actually deletes the wu?s despite the message,hopefully it is saved under one of your slots,cd to your folding folder with cmd and it may or may not work.
http://foldingforum.org/viewtopic.php?f=19&t=6042

hopefully someone more experienced could shime in and make sure one way or another,on deep water here,was this nvidia debacle recently where it was several slots in my work-directory since they piled up there then,a -send all worked at that time though when the servers was up again.

I have the same problem as wolfpup, I think. Usually when it happens the exact same work unit appears on two consecutive slots. Like slots 5 and 6. I guess it just restarts working on the supposedly failed unit. Or it just fails to restart it. Except that I have the system running non-stop, and those duplicate work units on different slots still appear.

This is logged just after downloading new work:
[10:24:43] Initial: 0000; - Receiving payload (expected size: 97207)
...
[10:24:54] fcCheckPointResume: file hashes different -- aborting.
[10:24:54] mdrun_gpu returned
[10:24:54] Checkpoint failure
[10:24:54] Folding@home Core Shutdown: UNSTABLE_MACHINE
[10:24:57] CoreStatus = 7A (122)
[10:24:57] Sending work to server
[10:24:57] Project: 5739 (Run 0, Clone 23, Gen 769)
[10:24:57] - Read packet limit of 540015616... Set to 524286976.
[10:24:57] - Error: Could not get length of results file work/wuresults_07.dat
[10:24:57] - Error: Could not read unit 07 file. Removing from queue.
[10:24:57] Trying to send all finished work units

... But on the next slot it prints this as if it was starting the same thing again with no problems:
[10:24:59] Initial: 0000; - Receiving payload (expected size: 97207)
...
[10:25:05] Project: 5739 (Run 0, Clone 23, Gen 769)
[10:25:05] Assembly optimizations on if available.
[10:25:05] Entering M.D.
[10:25:11] Tpr hash work/wudata_08.tpr: 1810823123 2680969346 2926436853 3399626860 1268365315
[10:25:12] Working on Protein
[10:25:12] Client config found, loading data.
[10:25:12] Starting GUI Server
[10:26:55] Completed 1%

Or is this just typical behaviour? It can't realize in any other way that maybe there is still old data in the directory. It just checks hash, finds that it's different, but then print oops, "unstable machine" which isn't true. It's not even overclocked and the client does the same when the GPU is underclocked. It's even downloading the same work unit twice, it seems, doesn't it.

Receiving the same WU after a failure is pretty standard. It can send you the same one 5-6 times, giving you additional chances to run it, before it either shuts you down for 24 hours, or gives you a new WU. It's that first failure that seems odd to me.

Last night I had the exact same WU on all 3 GPUs in my Vista box. It took 30 minutes of deleting before I had different WUs on each card, because it kept downloading the same one. :ahhh:

Alright, scrapped the Deino client and grabbed the Mpich one. Both "advmethods" and the "big" settings are on, and just like the Deino client, -verbosity 9 seems not to change anything and -smp simply gives me
http://img685.imageshack.us/img685/8745/smpfail.jpg As it stands, I'm still regularly getting SMP WU's under 100 points each.

C'mon Fuzzy, I know you'll save the day. :thumbsup:

Edit: Oh, and despite deleting both the work folder and the queue.dat file from my SMP folder in hopes of getting a proper WU I'll get the same WU as I just tried to get rid of. Those are the only two things to delete to get a new WU, right?

OK, lets step back a second. Along with the proper client, running SMP requires:
1- being installed in a password protected Admin account
2- being run in that same account, while logged in

Which OS are you using? XP isn't picky, but for Vista, and I think W7, DO NOT install the SMP client in the Program Files. The easiest way is to create a folder for it directly on the C:\ drive, or D:\, or in the case of the machine I'm on, E:\smp\Folding@home-Win32-x86.exe,

The verbosity flag will just put extra lines in the log, once we get you running. It helps with trouble shooting, later on.

How and where did you install this? Did you run the install.bat, and what was the result? I'm lazy and don't like running anything from a command line. I download the client to the desktop, run the extraction process and send the files to the folder I set up. Open the folder and double-click the install.bat file. A command window will open.

-------------------------------
MPICH2 Process Manager, Argonne National Lab installed.
account (domain\user) [****-XXXXXXXXXX\****]: don't type, just hit Enter
password: type password for the Admin account and hit Enter
confirm password: do it again
Password encrypted into the Registry.
If you see this twice, MPI is working
If you see this twice, MPI is working These 2 lines are what you need to see
Press any key to continue . . .
--------------------------------

From the bit of log you posted, it looks like you didn't get install.bat run properly. If you don't get this when you start the client, it will not be able to start and run SMP WUs.

---------------------------------------------------------------------------------------------
Note: Please read the license agreement (Folding@home-Win32-x86.exe -license). F
urther
use of this software requires that you have read and accepted this agreement.

Using local directory for work files
4 cores detected
If you see this twice, MPI is working
If you see this twice, MPI is working

--- Opening Log file [April 14 23:05:48 UTC]


# Windows SMP Console Edition #################################################
###############################################################################

Folding@Home Client Version 6.29

http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: E:\smp
Executable: E:\smp\Folding@home-Win32-x86.exe
Arguments: -smp -verbosity 9
---------------------------------------------------------------------------------------------------------

Without the -smp flag, this client defaults to being a fancy CPU client.

Sometimes you have to delete those several times before it will give you a different WU. It like to make you try to run a WU 4-5 times before it lets you give up. If it's running a cpu WU right now, stop the client and add the -oneunit flag. When it finishes, make sure the account is set up right and run install.bat. Remove -oneunit and add -smp. Fire it up and see what happens.

If you don't want to log in every time you start the machine, like me, try Start>Run>control userpasswords2
This will let you set it to log you into the Admin account automatically when Windows starts, without ever seeing the proccess.

:facepalm:

I might have overlooked the .bat file. You know the one that's clearly named at "install.bat"? But yeah, after putting the PW in, both lines appeared and my SMP folded a 400-something point WU yesterday and is in the process of folding a 484 point one now.

The one thing I am curious about, however, is that when you set the SMP client to get "big" files, the "-advmethods" setting is on, and I have the -smp flag put in, doesn't that mean I'm supposed to get heftier sized WU's? I mean, most of my GPU WU's are easily over 400, normally 500+ points each. The SMP ones should be quite a bit larger, right?

:brokencomputer: :whistling: :lmao: :goodjob: :cookie:

This is where the A3 WUs get a little strange. They are low point value, but carry a nice bonus, based on how fast you can send them back. The newest ones, 6040 and 6041, are 3 times the size and worth 1395. They take up to 3 days, depending on your cpu, and the upload is around 68 MB.

To qualify for the bonus, you need to use a passkey and complete 10 A2 or A3 WUs within the deadline. I had been running the A2 WUs in a VM for quite a while, so my first A3 got the bonus.

Fahmon will not show the bonus, so most of us are running http://foldingforum.org/viewtopic.php?f=14&t=9903. Download and install the latest version. It is similar to fahmon to set up, and it will calculate the expected bonus for the WU you're running.

Ahhh, that's a crafty little scheme they threw together. HFM is up and if what I'm seeing is right, I have project 6014, worth 1336 points, and due tomorrow at 7pm. A couple hours shy of two days. It's up and running at least, that's what counts.

I'm not getting any new work for my main system today. This started this week. At first it only took a few tries, but now the cpu has been running idle since the morning because it resorted to trying only once an hour after a few tries. I have restarted the client but it's doing the same thing again. Worked fine for a few months. I only changed the settings slightly now because someone on the official forums suggested that, but that had nothing to do with - I think the servers are just overloaded or something.

on one client I had/have trouble getting smp work lately,last days or so, and got work in form of a1 just now,not the best but better than it sitting idle.
seems they had a shortage of smp work from the look at it,also seen it mentioned that new work is just around the corner,hopefully it is so.

while the smp client was unemployed I ran a few ordinary cpu units,wasn′t soo bad,a few 100′s ppd on those,did them with oneunit flag though,some overlap when the a1 got in but only for a few hours until it was finished.

I had the -forceasm parameter set for the GPU client. After I removed that, I've had no failed GPU work units at all. So, maybe that was it, but after 6 units it's a little early to tell.

We made a brief visit to #136, but are back at #137 for another few days. It's still looking good to break 130 within a month. I really want us to break into the top 100, so I need several of you to buy dual quad servers and run -smp 16. Please start now. :stare:

Adding to a couple other posts, there is a shortage of both SMP and CPU WUs. Besides waiting for them to set up more projects, here are a few tips on maximizing you chance of getting work. I've been pretty lucky and haven't had too much idle time.
http://foldingforum.org/viewtopic.php?f=24&t=14255

Milestones
(TES)_wolfpup 1,250,000 04-18-10
Floating_Jet 40,000 04-18-10
Patton1243 20,000 04-17-10
Chaul 1,250,000 04-15-10
WTFimLEET 70,000 04-13-10
jaxawier 2,000,000 04-12-10

No new members, so I'm keeping the cookies this week. :frog:

Fold On! :goodjob:

sry, a tad over my budget, got a package of demo stuff including an asus P5N-D mb with an E5300,also a demo used gt 240 which was a fail,so it is only that E5300 and a 9800gt for now,guess it is better than nothing, btw increased available memory on the vm and suddenly it got work,had enough of protomol-work for now :hehe:
kind of waiting for thuban′s to appear in the local shops,then it depends how expensive they get :shrug:

Sadly my roommate came an repossessed the PS3, so it's no longer folding for me. (Til I can snag it back. ) He got a new game, that's why. It has been folding for some time for me. (100+ WUs complete.)

At least I have a new game I can play, Medal Gear Solid 4. I love that I do not pay for PS3/PS2 games, I just play the ones he buys. All my money is dedicated to hardware/PC games ATM. (Getting a 40" Samsung HDTV tomorrow. : :celebration: ) Too bad it can't fold.

After 19 completed and 0 failed GPU work units I can confirm that my problem really was caused by the -forceasm parameter I had used with the client. This caused HFM.NET to count the units as failed and I could see something about unstable machine in the logs at the start of the work unit followed by downloads of the same unit again. I don't have any problems even after overclocking the GPU from 850 to 920MHz now.

I want to say congrats, but that means you're going to pull away from me in points now. (I was only 20,000 behind last week, now I am 50,000 behind. THIS MEANS WAR!!!! )

That aside, I truly am glad you figured it out. :celebration: Go Team Elder scrolls!

Have to congratulate both of you, big hd tv is always nice
and very nice to get the folding sorted out there :goodjob:

guess I′m quite in the same situation about money spend on hardware and games,thought I finally get to play mirrors edge,but when gaming smp -3 works on most games like 7 kingdoms frythan wars or 7k aa,doesn′t work on oblivion and newer titles.mw is on the border,in all cases besides 7k aa two gpu clients get shut down,both tops around 8k a day.

still haven′t got any good single slot gpu′s ideas for sure,to bad buying cards and it doesn′t fit,a 4850 would fit but not much bigger than that,tips if any would be welcome :brokencomputer:
nvidia that is though,gave away my 4850.

We are at #135. We're due to move up 2 more spots next week, and hit 130 the week after. Yay, team :hot cheerleader smilie:

If any of you are stat junkies like me, and wondering what the hell I'm doing, I just started running -bigadv on the i7. Instead of getting 15-18K per day every day out of it, I'm getting about 26-28K per day in the form of one lump of ~55K every 2+ days. My chart is going to look like a saw blade. I'm 8 hours into my second one, and they take about 53 hours on my rig- i7-930 @ 4.0, 12GB triple channel @ 1600, W7 Ultimate 64 running http://www.evga.com/forums/tm.aspx?m=4464 in a VM. So far, so good, for a 10K PPD increase.

Milestones
Patton1243 50,000 04-24-10
Shanater 175,000 04-22-10
MistaMo 200,000 04-22-10
Milanius 35,000 04-21-10
WTFimLEET 80,000 04-19-10
gigabyte 1,000 04-19-10
(TES)_wolfpup 1,250,000 04-18-10
Floating_Jet 40,000 04-18-10
:cookie: :cookie: :cookie:

I'll pat myself on the back for a stat that doesn't show on any lists- I just passed 50K WUs for about 3 years of folding. :flamethrower:

No new members this week.

Fold on! :goodjob: