F@H: Big WU + advmethods = 150% points

[tyan]

Got my first big WU´s if p130_1RYP_AAAA_UM counts as one. At least it´s taking 2 days on my main rig... Hope the 182 points is worth it 😉

 

[GLeeM]

Originally posted by: tyan
Got my first big WU´s if p130_1RYP_AAAA_UM counts as one. At least it´s taking 2 days on my main rig... Hope the 182 points is worth it 😉

If your main rig is the one listed in your sig, it should not take two days to finish, maybe one.

Could you look at how long each step (percent) is taking?

A non-OCed A64 should finish in a day or less. My P4 3.2 takes 25 hours with one running.

Maybe something is not set right.

 

[tyan]

One step takes about 15 to 30 minutes so I guess the whole time it takes is quite close to 24h. It´s just that I can´t keep it running during the night... 🙁

 

[lobadobadingdong]

my 2500+ barton @ 12.5x168 is completing the big ones (p130_1RYP_AAAA_UM) in about 13-14hrs.

 

[GLeeM]

Now that I've got everybody doing these, there are more Double Gromacs going out. 😕
If you have the SSE that these use, the "bonus" for doing them is even better than the BigWUs. The only trouble is getting them.

@lobadobadingdong
🙂 It's not nice to fool with Mother Intel! 😉

That's an awesome time.

 

[lobadobadingdong]

😀 that's the time I pulled down from the log file on that one.

I went back over that one, 😱 I jumped the gun, that one only finshed 42 of 100 steps. So that would really translate to ~26-28hrs.

[09:09:30] Protein: p130_1RYP_AAAA_UM
[09:10:02] Extra SSE boost OK.
[09:10:03] Writing local files
[09:10:03] Completed 0 out of 100000 steps (0)
...
...
...
[21:25:18] Completed 42000 out of 100000 steps (42)
[21:41:56] Quit 101 - Fatal error:
[21:41:56] Step 42992, time 42.992 (ps) LINCS WARNING
[21:41:56] relative constraint deviation after LINCS:
[21:41:56] max 0.089698 (between atoms 45947 and 45948) rms 0.000554
[21:41:56]
[21:41:56] Simulation instability has been encountered. The run has entered a
[21:41:56] state from which no further progress can be made.
[21:41:56] If you often see other project units terminating early like this
[21:41:56] too, you may wish to check the stability of your computer (issues
[21:41:56] such as high temperature, overclocking, etc.).
[21:41:56] Going to send back what have done.
[21:41:56] logfile size: 38575
[21:41:56] - Writing 39259 bytes of core data to disk...
[21:41:56] ... Done.
[21:41:56]
[21:41:56] Folding@home Core Shutdown: EARLY_UNIT_END
[21:41:59] CoreStatus = 72 (114)
[21:41:59] Sending work to server

 

[Pokey]

I have a question.

Now that I am set up for big WU's what am I looking for to identify one? I use F@H LogStats to survey my Lan. Am I looking for credit, something in the name, core, other?? 😕

Thanks 😛

 

[lobadobadingdong]

look at the protein, it tells you what you'll be working on. see the bold below for the part to look up on the F@H home site.

Protein: p130_1RYP_AAAA_UM

 

[Insidious]

anyone know what this one is worth? (output from F@HLogStats)

Protein p130_1RYP_AAAA_UM
Protein Core Gromacs
Credit
Deadline
Current Frame 73 of 100 (27 left )
Time Per Frame 17 mins, 5 sec
Time Left 7 hours, 41 mins
Client Version
Core Version Version 1.68 (August 18, 2004) SSE Enabled
User Name Insidious
Team Number Team 198
Est. PPD 0.00 (0.84 WUs)
Est. PPW 0.00 (5.90 WUs)
Uploaded Projects 591

-Sid

1700 minutes..... whew I'm gonna have to pour more :beer: on this thing to cool it off!

 

[ProviaFan]

Originally posted by: Insidious
anyone know what this one is worth? (output from F@HLogStats)

Protein p130_1RYP_AAAA_UM

-Sid

1700 minutes..... whew I'm gonna have to pour more :beer: on this thing to cool it off!

For some unknown reason, I can't reach anything at Stanford right now, so I can't help you out, but try going in to F@H LogStats, Settings, General Settings, and press the Update... button. That will pull down the latest list of WU scores, which should then display in the stats reports.

 

[Insidious]

Thanks pro..er..... J :beer:

tried it... still the same... also, when I go to the projects page, this one isn't listed.

Well........... gotta love a mystery.... right?

-Sid

 

[ProviaFan]

Originally posted by: Insidious
Thanks pro..er..... J :beer:

tried it... still the same... also, when I go to the projects page, this one isn't listed.

Well........... gotta love a mystery.... right?

-Sid

If you're using -advmethods, then it could be a beta WU, in which case it won't be listed in that page. There's another page that lists the beta projects, but I have no idea where it's at, and wouldn't have much luck trying to locate it right now. 🙁

 

[GLeeM]

@Insidious
All of the p130* WUs are 182 points

I got the same WU, I had to open extra-nfo.csv file in the FAH LogStats folder with Notepad. I copied the line with p1301 in it and pasted it below and edited it to look like what it is identified as by LogStats.

@lobadobadingdong and everyone
There will be quite a few LINCS WARNING EARLY_UNIT_ENDs with these and they expect them to happen. It is not something wrong with your computer.
They are trying something with these that will result in 2x to 4x more science being done, and they still get results from the early ends.

 

[GLeeM]

@ProviaFan
LogStats update gets all WUs even the beta ones.
The -advmethods flag only gets late beta WUs, no beta WUs

They fixed the naming glitch after a few got out. I suppose as they are returned and prepared to be sent as the next Gen, they MIGHT be renamed, but I doubt it.

The beta projects are listed here: http://vspx27.stanford.edu/psummaryB.html

Edit: some of them are also listed with the regular projects, those listed in both are advmethods

 

[Khyron320]

I think i have one of these WUs

is it worth 182 points after the 150% bonus?? or before

Box is a 2.6c@3.25ghz 250fsb:250mem 512ram

both Folding windows got the same work unit its taking 22 minutes per step...

DC1:[19:54:07] Project: 1301 (Run 1055, Clone 1, Gen 1)
DC2:[19:52:55] Project: 1301 (Run 1023, Clone 1, Gen 1)


[19:54:32] Protein: p130_1RYP_AAAA_UM

[01:55:06] Completed 21000 out of 100000 steps (21)
[02:10:08] Timered checkpoint triggered.
[02:17:12] Writing local files
[02:17:13] Completed 22000 out of 100000 steps (22)
[02:32:14] Timered checkpoint triggered.
[02:39:19] Writing local files
[02:39:19] Completed 23000 out of 100000 steps (23)
[02:54:20] Timered checkpoint triggered.
[03:01:19] Writing local files
[03:01:20] Completed 24000 out of 100000 steps (24)

 

[GLeeM]

@Khyron320
Looks perfect, just a tiny bit faster than my 3.0c@3.2. They will vary a little though. Also, Gen 1 WUs will be less likely to early end than Gen 0. Supposedly, the higher the Gen # the more stable.

Are you getting higher points per day now than you would with other WUs?

Edit: I have 1 GB ram. Do you notice the BigWUs affecting how the computer responds when you use it? I don't.

The WU was originally 118 points.

 

[tyan]

Originally posted by: GLeeM

Are you getting higher points per day now than you would with other WUs?

Edit: I have 1 GB ram. Do you notice the BigWUs affecting how the computer responds when you use it? I don't.

The WU was originally 118 points.

date score WU
08.25.04 206 2
08.24.04 166 7
08.23.04 214 6
08.22.04 228 4

So far its producing pretty much the same score. My main rig (see sig) has been doing these for a few days now, and the other (Pentium M 1,6 running 24/7) started yesterday. My only worry is that when running these big WU´s there may come day when neither of them complete a WU 🙁 (I´m running F@H only on these two).

But now that both of these are crunching p130_1RYP_AAAA_UM´s I guess time will tell if its worth the effort... 😎

Oh, and my home cp has only 512mb memory, but I haven´t noticed any difference. Even when playing Neverwinter nights at the same time.

 

[GLeeM]

@tyan
If the timing works out that there is a day that a WU doesn't complete, the next day will be all the better!

Thanks for the input about the computer with 512mb memory and not noticing any difference 🙂

 

[Insidious]

Thanks for the help GleeM :beer:

I see that each of my computers finished one of these 'p130' WUs during the night and I received the 182 points for each of them.

-Sid

 

[Pokey]

Thanks Loba............... 🙂

I seem to be doing the p724_Abeta21-43amberGS Protein with 63 points credit.................and the same protein is on three machines...............

Also, I looked it up on the two summary lists and the p724 appears on both lists, but the points don't seem particularly high.

Pokey (still trying to get an understanding of this)😕


Edit: I may have answered my own question. On the descriptions page for this protein :

These are among the most demanding WU's so far because of the large system size.

So I am assuming it is one of the big'uns even though the credit appears rather smallish.....................

If anyone can confirm or deny, I would appreciate the input.

 

[Khyron320]

Gleem i only have 512mb of ram in the system and i dont notice any slow down in responce time. 2 clients are eating up almost all the ram i have though. I do not play games on this box though just surf and burn cds.

I have no stats program installed right now due to a format so i dont know my PPD

edit: Just so you know these WUs do slow down games quite a bit because they are hogging ram. I play warcraft3 all the time with friends over on a 3.0c@3.0 1gb ram and i ALWAYS load a good 5 seconds faster then them but with these new WUs going they all beat me. 🙁

 

[GLeeM]

@Pokey & all
I don't think p724 is a BigWU. I think just: p1113@40 points, p1122@111.6, p1301@182, p1302@182, & p1303 a DGro worth 517 which I want and have not had!
I just looked and there are over 2,000 BigWUs availiable on the .154 server.
It was my understanding that if you have settings right and there are BigWUs to be had, you will get them.
So Pokey, unless they prepared and added a new batch recently, you don't have settings right. Have you gotten any that you know for sure are BigWUs yet?

@Khyron320
Thanks for the valuable info 🙂

@all
Could someone with an AMD rig confirm that these BigWUs give more points per day than all other WUs (except DGro). I've heard of awesome times on some Tinkers with some AMD rigs. I just want to make sure the BigWU is what AMD users want to be doing.

 

[csyberblue]

Originally posted by: GLeeM
@Pokey & all
I don't think p724 is a BigWU. I think just: p1113@40 points, p1122@111.6, p1301@182, p1302@182, & p1303 a DGro worth 517 which I want and have not had!
I just looked and there are over 2,000 BigWUs availiable on the .154 server.
It was my understanding that if you have settings right and there are BigWUs to be had, you will get them.
So Pokey, unless they prepared and added a new batch recently, you don't have settings right. Have you gotten any that you know for sure are BigWUs yet?

@Khyron320
Thanks for the valuable info 🙂

@all
Could someone with an AMD rig confirm that these BigWUs give more points per day than all other WUs (except DGro). I've heard of awesome times on some Tinkers with some AMD rigs. I just want to make sure the BigWU is what AMD users want to be doing.

Mine took me over 24 hours to complete 1 full WU on an AMD XP 1800

 

[Pokey]

Thanks GLeeM. 😀

At the office now and I have one machine here with the "p130_1RYP_AAAA_umbrella" protein.

It is showing 1hr, 13 min. per frame with over 34% done and 81 hrs left to go. Seems pretty big but no credit listed........................

I looked in the folding folder and there is a "wudata_03.xyz" file that's over 21MB in size if that means anything.

I'm going to check all my settings and restart each "folder", and do a little watchful waiting............



(having fun with the spell checker) 😛


Edit: What setting should be in the Configure Advanced Tab/Deadlines check box? The machine with the p130 is unchecked and the preferrences diaologue box is blank.

Also picked up another p130 with no credit (points) listed, and about 38 hrs to complete.

I'll be quiet now.......................