NBR Folding@home Team

Your first post is always monitored to make sure you're not a bot. After that, post away

 

Will do when i get home

 

Time for the daily update, kiddies!

Grand Score 3041883 (certificate)
Work Unit Count 3201 (certificate)
Team Ranking (incl. aggregate) 1753 of 208195

Team members
Rank
(within team) Donor Score WU
1 reborn2003 1149340 1100
2 DR650SE 362454 423
3 TurbodTalon 358993 286
4 ganzonomy 326819 383
5 JTK 284114 222
6 Thierry19 203303 169
7 justinkw1 136040 140
8 too456 47761 36
9 TANWare 29328 66
10 Ayle 26243 53
11 privatejarhead 22246 19
12 Chiefule-reborn2003 20947 17
13 Prema 20630 30
14 RodKnock 18435 38
15 reborn2003-M11x 8371 21
16 TONY211 8141 41
17 Courtathor 7238 121
18 Rickards12 5557 9
19 TurbodTalon_XPS 1467 5
20 Anonymous 1325 2
21 TurbodTalon_M1530 867 4
22 H.A.L._9000 778 3
23 too456-3820TG 511 1
24 KoI 511 1
25 anseio 464 4
26 too456-XPS420 0 7

 

I noticed the stats page gives me 3 clients but I only have one system. I think this is the beta as since it downloads a WU while the old has not sent. I noticed the ID of the WU is either 0, 1 or 2 with the qued download becoming the free ID. Anyone else notice this too with the beta?

 

It's so cold in my apartment. My CPU has been running at 100% for 12hrs and it's HOT at 68c. LOL

 

At smp3 my cpu run the three cores 100% and the fourth varies by the system. Highest temp @ 2.66 GHz is 80c on the third core. I have been running it this way for quite a while.

I just got another CPU WU of a base of 1698 credits and at the 1.00% reading said 1.60 days. Estimated credit is 7481 and 23 mins 17 secs TPF with 4627 PPD.

 

Congrats TurbodTalon,

Great stuffs.
From 3rd moving up to 2nd oles!

Cheers.

 

Congrats too, we lowly O'l single lappy users don't stand a chance against you guys..............

We have blown through the 1700's though. Not all too much longer and we should be in the top 1,000. A great feat for such a small team IMHO.......

Edit; it should be mentioned over the last 7 days the top 6 have provided over 85% of the points, DANG............

 

Done, Though I didn't change the screen shot. May do that later if I find time.



Is it me or is Kakao Stats always down? Havn't had access to it since Sat.

EDIT: Now it's back online Added Kakao stats link to OP under the official stats link.

EDIT 2: Nice job TurboTalon!

 

Thanks DR. I have a feeling that you and I will be switching spots a lot though. I can't run all three of my computers full time like I could during the holidays.

 

Yea, I'll be having to give up my Inspiron 1720 in two weeks or so when I give it back to my brother. So You'll probably run off with it. I'll be losing a 2.6GHz Core 2, and the 8600M GT GPU. I reinstalled everything because he had destroyed the OS with viruses and trojans. It was mine orignally and I'm lending it to him. He kills it again and I'm done though. I'll just take it back for good.

 

Oh, so you're the 'family computer fixer guy' for your family too? Funny, when I was home on vacation how everyone seemed to migrate to me and ask why their computer was doing this and that, and if I could 'clean it up' for them.

 

Clean them up... for a fee.

I'm a hobbyist, like most of us, but I've got several women who bring their lappies to me for all kinds of things and tutorials. It's a nice way to make a buck.

 

DR back to 2nd oles!

2 DR650SE 391293 451
3 TurbodTalon 390454 309

Great stuff guys. Yup I trade IT service for food, favours and gifts.

Cheers.

 

Sheesh. People actually pay you to do repairs, I have to repaste computers, update drivers, fix problems etc for free

 

I'm a self-employed massage therapist and times are tough. It's not right to live in an affluent community, struggle to pay the rent, and be expected to give my time to people for free.

The people I work with understand that, in exchange for my various skills, they have to compensate me for it. It's win-win, if you ask me.

Tell the next person who wants you to do something that it's $25/hr and see how they react. If they're up for it, then great. If not, then you don't need to waste your time (which should always be respected).

 

Folding@home V7 client update is now available!

Version: 7.1.43 BETA
Release Date: January 9, 2012

For Windows - https://fah-web.stanford.edu/file-r...32bit/v7.1/fahclient-installer_7.1.43_x86.exe

 

Awesome Justinkw1,

Thanks for the heads up.
I'm going to install my new HS and give it a go.

Cheers.

 

However I'm a student and so are the people I usually help. Them being my classmates usually doesn't help matters. It's kind of bad to ask your classmates for money especially when you've known them for so long

On the other hand, that is a viable technique for people I don't like

 

Gonna DL 7.1.43 when I get home from work. (Hopefully the 5XXm cards are white-listed ).

/start vent

As for doing IT service, I'm pressed into fixing:
NP8130
NP8662
XPS M1710
a Bevy of EEE PCs that aren't being used
an Inspiron N5040
an Inspiron M5030

where all bar the first 3 get their fair share of "females who refuse to google" problems. (I will disclose, in the interest of avoiding undue sexism, that my girlfriend has decided to learn how to be self-sufficient with the computer and has opened up a MBP to clean out the vents and whatnot herself... so I know they can learn if they're willing.) What grinds me is that let's say I'm playing Project CARS (I'm a beta tester), or trying to get a folding problem fixed, or doing HW (I just graduated Queens College with a BA in Political Science), I have to drop EVERYTHING in order to fix what typically is "do you wish to upgrade Adobe Flash?". Granted, how computers work / die / blowup / fillintheblank is a personal interest of mine that dates back to the 286 (I'm 26, so further back predates my use), but to exploit that because you're unwilling to crack open google to type in the crux of the problem, or read a book that gives general teachings on the fundamentals of windows, to me is unexcusable.

Which leads to my mom's dad, for whom while he may have a book explaining how to convert a .doc or .odt into a .pdf, will require a LETTER-PERFECT, written, step-by-step (ie: numbered) direction-sheet for each and every small operation there is. (ie: how to email is ONE sheet, how to attach a document to an email is ONE sheet, etc.,) Failure to do this makes him "confused", even if 10 blank lines are between each instruction set, and renders BOTH (or all sets if more than one on a page) shredded on-sight. Additionally, you cannot say "click NEXT" if it says "Click OK", you have to give each individual button-choice (again, letter-perfect) and the ramifications that transpire by clicking the WRONG button as well.

For him, I have had to do this for the following:
How to make an email
How to attach docs to an email
How to install EVERY program he has
How to UNINSTALL EVERY program he has
How to run CCleaner
How to upgrade his computer from Vista to 7
How to reformat a drive properly
How to use DriverSweeper
How to use Revo Uninstaller
How to use Openoffice / LibreOffice.

I think I've probably written about 250 pgs worth of information for him, and it most likely will grow. With no compensation ($, food, etc.,) for how much time and editing I have to do for him. (Yes, he actually proof-reads the instructions I write for him.)

/end vent


References to how bad it is to run 2 consoles at once:
v7 / i7-2670qm + v6 / GTX 560m = 15,000 ppd
v7 / i7-2670qm ALONE = about 19,000 to 20,000 ppd

Unless your GPU is something so powerful that it can remove the 4,000 to 5,000 ppd difference, until we know for sure that v7 whitelists the 5XXm cards, it may just be better points-wise to run v7 smp only.
Jason

 

Still no gtx580m support...

 

Yea, pretty much. Funny thing is I have another brother who does IT support for Google out in Cali. So he should be the go-to guy, but instead it's me. I also have an Inspiron mini to get fixed as well as the Inspiron 1720.

Thanks, Be interesting to see how it plays out, but I think Turbo will run away with it in the end.

Nice, I may have to give this a try.

@ganzonomy - Thats pretty interesting. You know one thing I noticed, when my heatsink Died (don't know how a or why a heatsink can die , but a heatsink prelacement fixed my issue) and I was only running my GPU's on my M17x R2, my CPU was still at around 50%+ usage. So I wonder if running both CPU and GPU at once actually slows things down. And in that case I wonder if running just GPUs would fetch more points, as I think my GPU's actually fold faster. And the CPU usage remains high to feed the GPUs info. I know that running all 3, my 940XM stays pegged at 100% usage for days on end.

This would be interesting to see, because then we can maximize points. I may have to do some testing. Now as for having to type all those steps that is way to crazy. At that point I'd be going crazy, which I have no doubt you are. You should condense all of that into a book, and print it for him for his birthday. Maybe He'd realize how much work your actually putting in.

 

Wow 1 million points reborn!!!

I got my 6990M today and it works but still have problems so I think the CPU is not functioning properly....

Break the 1500 mark soon folks!!!!

 

Yea, reborn has been killing it! I think he has a ton of work systems folding.

Thanks to justinkw1 for fixing the OP after I messed up all the code using "switchmode" +rep Not sure if anyone saw this yesterday, but he did a lot of work to get the look and feel of the OP back. He also added the following stats links to the original post to make it easier to find stats. I know it makes it useful for me on my phone when I'm on the go.

TEAM STATS: Stanford (official) | EXTREME Overclocking | Kakao Stats[/span]

 

Bummer. Are there any improvements though? At least the ticket has our GPUs noted.

He's the sort that does things to the letter. As far as I'm concerned, if in 100 pgs of typing i type "next" as "nwxt", you should be able to get the gist of the point (i didn't write "back", after all). Nope... in his eyes, everything is strict compliance (you'd think he's a lawyer, he's actually a graphics designer) or the service wasn't rendered correctly. (He's also the sort that calcuates tip % at restaurants to the penny.)

As for CPU + GPU use, from what I've noticed, I think it weighs more heavily not so much on if CPU and GPU run together, but if they are on different consoles. This is because each console needs AT LEAST 2 cores, so you're affording 2 to v6, and 6 to v7 (with an i7-2670qm). By affording all 8 to v7 and turning off the v6 GPU, yes you may do fewer WU / day on the machine, but the SMP that does get done gets a much bigger QRB (Quick Return Bonus) that can EASILY make up for points lost by a 560m, and perhaps even a 580m. The reason on my NP8662 I keep v7 for GPU and SMP is because it's a 24/7 Folding Rig, and because the 260m was whitelisted for v7 (which apparently the 5XXm's still haven't received). It's also only a dual-core, so I don't find that it would take much of a hit running SMP and GPU on the single console.

For reference (ppd on smp)
P9700: 4600 to 4800
i7-2670qm: 18,000 to 20,000

Test this out by finishing the WUs for your GPUs, and then pause them. Let the CPU run in SMP ALONE for a few days, and see what the ppd for the GPU is. If you have GPU in v6, after you turn it off you should notice a pretty healthy SMP spike in v7 SMP. If you have GPU also in v7, it may not be as noticeable a spike because v7 also has the GPU console... but try it, it may spike anyway.

As far as killing your heatsink, I wonder if you did the "clip removal mod"... perhaps the brace is prone to torque-induced distortion that can be just enough to form an airbubble between the CPU, the thermal paste, and the copper heatsink takeup. It could also just be a poorly made heatsink, who knows.

Jason

King,

Have you tried turning OFF the OC on your CPU when you fold? F@H says that the way the CPU is stressed during folding can cause even mild OCs to make failed WUs. They claim the only thing more intense than F@H is intelburntest (in fact, they recommend IBT because if the OC can survive IBT for 24 hrs, odds are it will be fine for F@H).

Jason

 

Hmm, I may have to do some testing then. Although all my systems are running SMP and GPU on V7, so V7 is all I'm running. But it would still be good to test anyhow.

As for the heatsink, nope, never removed the clips or did the retention mod. But I'm definately glad to have my 940XM back in. I had to run my 540M dual core for a bit, just so my GPUs could keep folding, and even that a 35w chip was breaking 80°C. But at least I was able to keep the system up for a few days as I waited for the heatsink. Now I have an extra heatsink that I don't want to sell because it doesn't seem to work, but can't bring myself to throw away

Thats a good point. I was testing my new heatsink and was running wPrime 1024 at 90w TDP, and even F@H was more stressfull on the CPU than wPrime.

King, what are your CPU settings and what issues are you having with the chip?

 

It's the PERFECT technique for people you don't like. "Sure, I'd love to help you, but I'm very expensive!" HAHAHA

Finding that bigadv's don't get me more PPD. They take so long. I turned it off last night, but one more slipped through it's a 24hr one that's still got 15hrs to go. Will finish it. Update to v7 and get back to my standard WU's.

 

Thank you.

PUBLIC SERVICE ANNOUNCEMENT

Pursuant to this Post, in order to SUCCESSFULLY complete bigadv's on time, you will need MINIMUM 8 core / 16 thread of SMP power. Although it says "no earlier" within the text", assume that it will be changing Jan 16. They want the bigadv's done faster.

Jason

 

So does this mean that we need to change our slot options if we have added "client-type bigadv" for our SMP? I'll also need to make a change on the original post.

I only wish I had an 8core/16thread system.

I'd rep you but I need to spread the love apparently.

 

Heh, the love;s there. I wish i had an 8C / 16T system too (SAGER, are you LISTENING?! ). It does mean that you have to change the slot option by removing it, since if you get the bigadv unit, the timeframe will be too small for successful completion. As it is now, it takes an OC'd i5-2500k to squeak it under the buzzer according to F@H, and that's with a fairly healthy OC. Unfortunately for us notebook folders, that removes us from being able to employ bigadv... but at the same time, look at what some folders use. Awachs has something like 6,600 CPUs folding when he does an SMP (it's gotta be like SMP: or something like that), and many SMP folders who have highly competitive machines will get something like a quad-socket Magny-Cours rig, or a dual-socket intel xeon rig. The latest of the dual-socket Intels can be 10C / 20T, so doubling them up makes effectively 20C / 40T. F@H wants to take advantage of the highly evolving nature of CPUs as even the mainstream now would be considered supercomputer-level say 5 or 6 years ago. Supercomputer now, is like a Cray 10 years ago.

But alas, I digress. It also seems to be in the rumor mill that the "reference computer" is going to change at some point as well. What we as folders are benchmarked against is this (scroll down to "How do you decide how much credit a work unit is worth? How do you determine how many points a work unit is worth? " for veracity of statement). This is why those of us with i7 laptops are experiencing CPU ppd's that could close in on 20,000ppd with the right tweaking. The way baseline points are established is that the WU is 110 times the number of days it took THAT particular P4'd machine to complete one WU. However, with most folders going SMP due to being C2D or newer, that's going to change. Can you blame PandeLabs though for wanting info that could solve pandemics more quickly? Of course not, and that's why they're encouraging the bigadv's to more "robust" computers than what is merely "mainstream" or "performance".

 

Ahh, yea it makes complete sense and it's crazy to think how far things have come. Which brings me to another question.

Can a quad core like the i7 940XM be run as four single uniprocessors? From my understanding, since 8 threads is hyper threading, they share resources as a quad can run 4 floating point calculations at once due to only 4 floating point units. I doubt I have my understanding right, but something I read put the idea in my mind that maybe it would be better to run four uniprocessors rather than one SMP using all CPU resources. not sure if I'm even making sense.

What I read was the first line in this thread that got me a bit confused.

 

Heat pipes can loose effectiveness. The problem is the liquid and preasure to calibrate them. Like a rubber tire they can loose preasure either positive or negative and this effects their performance. Also even the liquid can can be lost. Either way the temperature can start going haywire.

Thanks, I pulled bigadv out, hadn't started one but did leave big for packet size....................

 

Ahh gotcha. Maybe thats what happened. I wonder if F@h had anything to do with that I guess we'll see if it happens again My CPU hovers around 75°C 24/7 due to F@h.

 

Same here, but b/c of Sager's design, when I finish folding (and quit)... about 15min later it's back down to about 40 deg C.

 

Yea, mine would but I set it up to download a new WU when the current one is at 90% so my CPU is pretty much always folding. If there is no folding going on (rare the last few weeks) my CPU idles around 45°C.

 

Time for the daily update

Grand Score 3443115 (certificate)
Work Unit Count 3579 (certificate)
Team Ranking (incl. aggregate) 1612 of 208305

Team members
Rank
(within team) Donor Score WU
1 reborn2003 1266068 1199
2 DR650SE 405938 465
3 TurbodTalon 403780 317
4 ganzonomy 368107 420
5 JTK 291454 226
6 Thierry19 248361 205
7 justinkw1 147888 175
8 Ayle 57371 85
9 TANWare 54242 94
10 too456 49059 37
11 Prema 33055 41
12 RodKnock 30920 49
13 privatejarhead 22246 19
14 Chiefule-reborn2003 20947 17
15 Courtathor 8935 122
16 reborn2003-M11x 8371 21
17 TONY211 8141 41
18 anseio 7216 14
19 Rickards12 5557 9
20 TurbodTalon_XPS 1467 5
21 Anonymous 1325 2
22 TurbodTalon_M1530 867 4
23 H.A.L._9000 778 3
24 too456-3820TG 511 1
25 KoI 511 1
26 too456-XPS420 0 7

Almost at 3.5m!

 

Nice! Looks like Turbo and I have gone back and forth a few times today

 

Hmmmm... maybe what I'm doing now isn't a bigadv then. It's a 13m:51s simulation, but the timeout is 14 days from now. Not big enough, eh?

It's currently 57f in my apartment (I like it cold, and am doing cleaning anyway), and my CPU is at 100% and 66c. It never really goes about 70c anyway.

 

Nice. I like it cold in my computer room too.

I just read that I can run my T9500 on both cores at 2.8GHz, as opposed to 2.6GHz by using throttle stop and disabling EIST and disabling EIST in the bios as well. And I can undervolt. Looks like my T9500 may be a little better for folding now

 

BigAdv's are BIG to the point that on even a 2670qm, it should take something on the order of an hour and a half to two hours to do 1%. The minimum is an i5-2500k that's been OC'd to 4.7GHz, with an 8-core, it's something like an i7-2600 @ 3.6. However, because F@H isn't keen on OC'ing because of stability issues, it would probably be best just to use a 3930k or 3960x as the fastest single-cpu folder (they avg OVER 40k ppd in SMP), before going to an 8 or 10-core xeon solution.

 

See here http://forum.notebookreview.com/har...23-baffling-hardware-fault-my-cpu-faulty.html

Basically I can't run in any other mode other than safe mode. Already tested different ram, gpu, HDD, wifi and BT, different windows and no windows so I think the cpu just isn't working properly anymore. Safe mode works no problem and can route around in bios etc lol

CPU failures I know are uncommon but I think when windows loads normally it uses turboboost or hyperthreading which safe mode probably doesn't and that causes immediate shutdown at exactly the same point during boot up.

 

a notice for DR650SE

as of 7.1.43, the NP8130 is fully whitelisted (the 560m works fine!)

That is all

Jason

 

Interesting that you mention the Uniprocessors DR.

I actually added them to a couple of my machines a couple of days ago. They seem to register as another Client. They seem to be working fine as well. Pulling in small cpu WU's. I will see if after 10 WU's have been completed whether or not the bonus points are worth it.
Pity about the Big Adv WU's, I was really hoping for massive 70k WU's lol

I'm down to folding on only 1 notebook at work now. It's my new work lappy. It seems to probably finish may 3 WU's in 24hours or maybe just 2. Quite slow but every little bit counts ey. Not too sure how long I can keep it running though as I checked the temps and they are running constantly at 98 degrees for both CPU and GPU. I better not break my new work lappy, I only just got it lol!

HAHA also its great to see DR and Turbod going toe to toe lol. I have a feeling that ganzonomy will be battling for that top 2 spot really soon as well! WATCH OUT

Cheers.

WOW nice man. So no more running 2 client consoles then? One for SMP and one for the GPU?
That would be nice if it recognizes my NVS 4200M GPU at work as well.
I'll give it a console upgrade and see if it picks it up later on.

Cheers.

 

Hey Chief, and all who foldeth.

I have come to the conclusion that running GPU and CPU on one machine is detrimental to the ppd output. . Instead of 18k to 20k ppd on the i7, i'm getting 5k on the i7 now. As such, I'm going to stick SOLELY to cpu folding on the NP8130. It is good to know though that both CPU and GPU can use the same client.

Jason

 

Interesting Jason,

I'll be sure to try that out.
So the higher the PPD score/number the better ey?

ATM my SMP one is at PPD = 89111.69 however the credits seem quite low for it. I'm not really too sure what to make out of it.

Edit: It's dropped down to PPD 24189.08 (it seems that it is still calculating)

Cheers.

 

the way scoring works is like this: rather than thinking of a "y = 0 when x = number of days point, and y = points when x = 0 number of days" and then making a normal slope, imagine half of a parabola that works this way, but "y = infinity when x = 0 days" and "y = 0 pts when x = deadline" and you'll have an idea. The initial drop is HUGE, and slowly tapers off as you approach the deadline. Because of this, that's why multi-socket machines do well both for Pande (they get info faster) and guys with multi-socket rigs (they get more points because the clock is stopped more quickly).

You get tons of PPD by either doing LOTS of small WUs really quickly (At one point I got 25k from my NP8130, because it got GOBS of tiny WUs and dispatched them in about 80 min each), or you get a ridiculously large WU and finish it "awachs style" (ie: sheer brute force of 6,600+ i7's). PPD is also based on the urgency of the WU. If you finish a low-priority WU quickly, it may garner fewer points than a high-priority one as well (provided that the high-priority one has a correspondingly proportionate due date). Ergo, PPD is a hybrid of speed you fold it, the amount you fold per day, and how long it took a baseline machine to fold the WU you're currently folding. (Faster, QUicker, More Plentiful = more PPD).

 

Look at EXTREME Overclocking stats. A huge noticable dip in PPD from when Chief pulled back on his folding Now is when I make my move!

 

Time for the daily news.

Grand Score 3627963 (certificate)
Work Unit Count 3749 (certificate)
Team Ranking (incl. aggregate) 1563 of 208354

Team members
Rank
(within team) Donor Score WU
1 reborn2003 1332823 1258
2 DR650SE 426424 482
3 TurbodTalon 417722 325
4 ganzonomy 383074 435
5 JTK 305802 232
6 Thierry19 277673 229
7 justinkw1 159865 194
8 TANWare 61029 107
9 Ayle 59415 89
10 too456 50407 38
11 Prema 33055 41
12 RodKnock 33036 52
13 privatejarhead 22246 19
14 Chiefule-reborn2003 20947 17
15 Courtathor 8935 122
16 reborn2003-M11x 8371 21
17 TONY211 8141 41
18 anseio 7982 15
19 Rickards12 5557 9
20 TurbodTalon_XPS 1467 5
21 Anonymous 1325 2
22 TurbodTalon_M1530 867 4
23 H.A.L._9000 778 3
24 too456-3820TG 511 1
25 KoI 511 1
26 too456-XPS420 0 7

 

Woot Woot! Top 1500 here we come!

 

A small observation having a dualcore / 260m and a quadcore/octathread / 560m:

It is better to have too much gpu and a puny cpu than vice versa. Running the 2 clients together, my ppd for the cpu dropped by over 12k (hence my puny score on kakao for the past 24 hrs). The gpu can only make up about 7k of this when it is a 560m. Because of this, we can say the 560m is NOT enough gpu to compensate for loss of cpu performance due to the extra console running. I suspect the minimum at the i7 SB-class is something around a 580m (a la prema), where the cores, VRAM, and the frequency are so plentiful they can successfully cancel out the 9k to 11k ppd deficit experienced. (See my previous post on the parabolic shape of the QRB point curve for why.). At the same time, for the relatively puny P9700 (a Core 2 duo), the 260m that is dragging along has MORE than enough power to compensate for the ppd loss on that computer.

Obviously for those with dual 6990s / 580m's, there is simply so much gpu power that this is irrelevant, but those of us with robust cpus and relatively puny gpus (anything from a gtx 560m / radeon 6870m and below) may want to stick to smp to maximize points. No change in improvement really even from 7.1.38 to 7.1.43.

I'm only SMP'ing the NP8130... the gpu is now whitelisted, but the ppd loss is too great (as is the fact that my electric bill is gonna spike). The NP8662, SMP and GPU.

Prema, if you can, please run your laptop with the client for the GPU off... let me know if the total ppd you get goes up or down between your 2 clients (I ask this since you have a 580m, and I wonder if the 580m is where the cpu loss is cancelled out by a robust gpu). ie: what does 7.1.43 get for the cpu with gpu on ppd (combined), and then what does the cpu get alone (gpu client completely off). Is the 580m able to cancel out the loss of cpu ppd?

Sorry for the bad news

Jason