TeAm F@H Project Database

[natethegreat]

Since there are so many different kinds of wu's and they are each quite different, I am going to be keeping track of what projects I am running and what the ppd of the projects are netting me and I am inviting all of the TeAm to help me out. Just post the full name of the project and what cpu it's crunching on and I will keep track of it here.

If you aren't using F@HLogstats I highly reccomend it as it makes keeping track of you cli clients a breeze and gives you some great info besides ppd.

I will continue to update this post as new info comes in. I will be making seperate categories for each type of cpu so projects will be listed under what cpu it was crunched on. Dont worry if a certian project/cpu is already listed, ideally I would like to have multiple results from a project/cpu combo and average the ppd out. Here's what I have so far:

A64
Project...........................................................................|.........Core....................|.....Credit.......|................PPD
p1481_tet_2nm...............................................................GROMACS....................364............................407
p1808_Collagen_Brodsky_refolding..........................AMBER...........................153............................160
p1809_Collagen_POG10new_refolding....................AMBER..........................153............................180
p2052_abeta_4mer........................................................GROMACS....................266............................135
p2081_p53_in_AGBNP..................................................GROMACS....................255............................191
p2106_lambda_5way_melt_4_10011.......................GROMACS.....................202............................109
p2107_lambda_5way_melt_4_10011.......................GROMACS.....................404............................110
p2306_BBA5_mutant.....................................................GROMACS....................188............................195
p2306_BBA5_pfold in water.........................................GROMACS.....................188............................184,196
p2364_Protein_A_wt_Sosnick.....................................DGROMACS..................586............................333
p2386_Protein_A............................................................DGROMACS..................742............................336
p2388_CI2.......................................................................DGROMACS..................495............................351
p2390................................................................................DGROMACS..................1112.........................275

-----------------------------------------------------------------------------------------------------------Average ppd=226


AXP
Project...........................................................................|.........Core....................|.....Credit.......|................PPD
p1808_Collagen_Brodsky_refolding..........................AMBER...........................153............................180
p1809_Collagen_POG10new_refolding....................AMBER...........................153............................173
p1850_Myosin6_PT_US................................................AMBER...........................52..............................46
p2106_lambda_5way_melt_4_10011.......................GROMACS.....................202............................116
p2107_lambda_5way_melt_4_10011.......................GROMACS.....................404............................117,181

-----------------------------------------------------------------------------------------------------------Average ppd=136


Thunderbird
Project...........................................................................|.........Core....................|.....Credit.......|................PPD
p2107_lambda_5way_melt_4_10011.......................GROMACS.....................404............................101


P4
Project...........................................................................|.........Core....................|.....Credit.......|................PPD
p2107_lambda_5way_melt_4_10011.......................GROMACS.....................404............................95
p2355................................................................................DGROMACS..................683............................290
p2387................................................................................DGROMACS..................731............................397

-----------------------------------------------------------------------------------------------------------Average ppd=261


P4-HT
Project...........................................................................|.........Core....................|.....Credit.......|................PPD
p1808_Collagen_Brodsky_refolding..........................AMBER...........................153............................72
p2051_abeta_4mer.......................................................GROMACS.....................267............................59
p2087_Protein.................................................................GB GROMACS..............172............................88
p2106_lambda_5way_melt_4_10011.......................GROMACS.....................202............................60
p2107_lambda_5way_melt_4_10011.......................GROMACS.....................404............................65
p2306_BBA5_pfold in water.........................................GROMACS.....................188............................65
p2355................................................................................DGROMACS..................683............................166
p2369_ACBP...................................................................GROMACS.....................153............................51
p2383................................................................................DGROMACS..................468............................180
p2387................................................................................DGROMACS..................731............................229

-----------------------------------------------------------------------------------------------------------Average ppd=104


Xeon
Project...........................................................................|.........Core....................|.....Credit.......|................PPD
p2107_lambda_5way_melt_4_10011.......................GROMACS.....................404............................78
p2361................................................................................DGROMACS..................689............................210
p2389................................................................................DGROMACS..................529............................190

-----------------------------------------------------------------------------------------------------------Average ppd=159


P3
Project...........................................................................|.........Core....................|.....Credit.......|................PPD
p1808_Collagen_Brodsky_refolding..........................AMBER...........................153............................51
p1809_Collagen_POG10new_refolding....................AMBER..........................153............................52
p2391................................................................................DGROMACS..................592............................74

-----------------------------------------------------------------------------------------------------------Average ppd=59


Cores Reported:
TINKER..........0
GROMACS......16
DGROMACS....12
GB GROMACS.1
AMBER...........8


Don't be shy, post what your working on now This could turn out to be a great resource for the TeAm!

 

[GLeeM]

P3 laptop 1 Ghz
p2391 ________ 592 points _________ 74 ppd

P4C 3.0@3.3
p2387 ________ 731 points ________ 397 ppd

p2387 and p2383 at same time on HT
p2387 ________ 731 points ________ 229 ppd
p2383 ________ 468 points ________ 180 ppd

Dual Xeon 1.6@2.5
p2361 ________ 689 points ________ 210 ppd
p2389 ________ 529 points ________ 190 ppd

P4E(?) 2.4
p2355 ________ 683 points ________ 290 ppd

 

[CupCak3]

hmm i'm going ot have to look into installing it this weekend

also, too bad it doesn't keep track of the time per frame

 

[natethegreat]

Originally posted by: CupCak3
hmm i'm going ot have to look into installing it this weekend

also, too bad it doesn't keep track of the time per frame

Actually F@HLogstats DOES keep track of time per frame

@thanks for the info GLeeM I will edit it into the top post after work today :thumbsup:

 

[Insidious]

This thread is a great idea Nate! :thumbsup:

:beer:

 

[RaySun2Be]

I'm using F@HLogstats, but I don't seem to have all the information. And my PPD shows things like 0.57, or 3.57

And it doesn't give me the total credit for the project I'm crunching. This is on a P4 2.4Ghz and an Tbird 1Ghz.

I'll try and do a screen print when I'm at home and post.

Here are the 2 screen captures:

P4 Stats

Tbird

 

[GLeeM]

Originally posted by: RaySun2Be
I'm using F@HLogstats, but I don't seem to have all the information. And my PPD shows things like 0.57, or 3.57

And it doesn't give me the total credit for the project I'm crunching. This is on a P4 2.4Ghz and an Tbird 1Ghz.

I'll try and do a screen print when I'm at home and post.

Here are the 2 screen captures:

P4 Stats

Tbird

Two things:

1. The P4 is giving 95.376 ppd

The Tbird is giving 101.472 ppd

This looks wrong! I will guess: either you are running two instances or another project on this CPU too or the Tbirds' 2107 is a first (zero) generation WU. By accident, the first gen were only half the length, but the same points.

2. FAH LogStats is not updating for some reason.

It is possible to edit the "extra-nfo.csv" manually. Open it in Notepad, copy the protein name from the LogStats display, look for that protein on the Currently Running Projects page, and add the necessary info, using what is already in there for an example.

Well three things: 3. There is a Protein Calculator here that I use. If the protein is not listed just type in the Work Unit Point Value and # of frames and the frame time.

 

[natethegreat]

RaySun2Be,

You will need to edit this into you extra-nfo CSV file in FAHLogstats:

"p2107_lambda_5way_melt_4_10011","114.00","404.00","100","GROMACS"

then check your summary tab again

edit: Looks like GLeeM beat me to the editing thing

 

[natethegreat]

Thanks for the help RaySun2Be and GLeeM, my top post has been updated with the new info.

Looks like the DGROMACS core is getting the best ppd right now at an average of 259 ppd over 10 wu's, GROMACS comes in with a 104 ppd average and I don't have enough info to avergae out the ppd on the AMBER or TINKER cores

Interesting that my AXP is not getting much better ppd on the p2107 than RaySun2Be's Thunderbird.

Quick Question, Since I am basically clueless about Intel cpu's could GLeeM's P4E be lumped into the P4 category or are the two cpu's quite different?

 

[Spacehead]

I'm wondering... do the PPD calculations change during the processing? I guess since it's based on the time per frame, does the time per frame change. Or is it pretty consistant thru out the WU?


Nate,
MY comps are both 3.0Ghz P4 with HT enabled running F@H & SOB.

Protein: p1808_Collagen_Brodsky_refolding
153 points
PPD 72

Protein: p2369_ACBP
153 points
PPD 51

 

[natethegreat]

Yes your ppd can vary drastically if you are folding on a rig that is used for other things besides crunching. When I used to game or even browse on my xp3200 I could go from 23 min/frame to 29 min/frame (while surfing) or over 1hour/frame (while gaming).

Would you be able to tell me if either of the projects you posted are running on the physical cpu or the HT(logical?)CPU? Looks to me like both of your F@H wu's are running on the HT.

Thanks for the info!

 

[Spacehead]

Originally posted by: natethegreat
Yes your ppd can vary drastically if you are folding on a rig that is used for other things besides crunching. When I used to game or even browse on my xp3200 I could go from 23 min/frame to 29 min/frame (while surfing) or over 1hour/frame (while gaming).

Would you be able to tell me if either of the projects you posted are running on the physical cpu or the HT(logical?)CPU? Looks to me like both of your F@H wu's are running on the HT.

Thanks for the info!

How do you tell???

About the other question, i meant is the time per frame about the same 2% into the WU as it is like 85% into it?

 

[natethegreat]

As long as you have completed a couple os frames, the time/frame should stay the same as long as the core is supplied 100% cpu power. I kind of remember something about some wu's last frame can take a lot of time but I don't know where I read this... Maybe GLeeM would know.

As far as how you tell what process is running on what (physical/logical)cpu, I don't know I haven't had the pleasure of owning a modern Intel cpu yet and I'm not sure if there is a way to tell the difference anyways

 

[RaySun2Be]

My P4 2.4Ghz, 512M Ram isn't doing anything else except sitting there on the workbench, crunching F@H. I'll go check to see if there are any unneeded services running.

Thanks for all the info!

 

[GLeeM]

Originally posted by: RaySun2Be
My P4 2.4Ghz, 512M Ram isn't doing anything else except sitting there on the workbench, crunching F@H. I'll go check to see if there are any unneeded services running.

Thanks for all the info!

I'm still betting that it is the first (zero) generation thing for the reason the Tbird is doing so much better.
If you look in fahlog.txt, near the beginning of the WU you will see something like:

[03:19:23] Project: 2389 (Run 3, Clone 97, Gen 2)

Your Tbird is probably running a Gen 0 or Clone 0


Frame times usually are close throughtout the WU. The first few frames of QMDs can vary from the rest.

FAH LogStats uses the most recent frame to calculate PPD.
You will notice more difference running two programs on HT.

On Intel HT processors, two processes use 50% of the CPU. If the two processes use different parts of the CPU, you can gain ppd. Also the CPU stays busy with the other process if there is a miss in the guess of the next needed instruction. It takes a long time for the correct instruction to get through that looooooong pipeline (which is what allows the higher clock). I don't really understand it all

Originally posted by: natethegreat
Interesting that my AXP is not getting much better ppd on the p2107 than RaySun2Be's Thunderbird.

Quick Question, Since I am basically clueless about Intel cpu's could GLeeM's P4E be lumped into the P4 category or are the two cpu's quite different?

RaySun2Be's WU is probably the first (zero) Gen WU, your's is probably different.

The P4E is Prescott with bigger L2 cache and other differences from P4C. I don't know if there is a big enough difference. Maybe we can find out?

 

[RaySun2Be]

RaySun2Be's WU is probably the first (zero) Gen WU

Yep.

The Tbird: Project: 2107 (Run 240, Clone 62, Gen 0)

The P4: Project: 2107 (Run 35, Clone 10, Gen 1)


Oh, and where do you go to get the project information like
"p2107_lambda_5way_melt_4_10011","114.00","404.00","100","GROMACS" ?

I added an XP2100 this morning, and it's working on a protein:
p1850_Myosin6_PT_US that I need to add to that F@HLogstats csv file.

Thanks for all the help!

 

[natethegreat]

Originally posted by: RaySun2Be
Oh, and where do you go to get the project information like
"p2107_lambda_5way_melt_4_10011","114.00","404.00","100","GROMACS" ?

I added an XP2100 this morning, and it's working on a protein:
p1850_Myosin6_PT_US that I need to add to that F@HLogstats csv file.

The extra-nfo CSV file can be found in you FAHLogstats folder, it should be one of the document type files in the programs main folder. I just double-click it and it opens automatically in notepad, then I just copy/paste the "p2107_lambda_5way_melt_4_10011","114.00","404.00","100","GROMACS" under the last entry. Hope that helps

 

[GLeeM]

Originally posted by: RaySun2Be
Oh, and where do you go to get the project information like
"p2107_lambda_5way_melt_4_10011","114.00","404.00","100","GROMACS" ?

Nate got the information to put that string together from Currently Running Projects page and by looking at the entries that are already in "extra-nfo.csv" file to make the string match.

Originally posted by: natethegreat
I just double-click it and it opens automatically in notepad

If a spreadsheet program is installed on computer, it will probably open in that if you double-click it. I have not tried editing in Excel, but it would probably work.

Anybody else have any WU timings?

 

[RaySun2Be]

Thanks GLeeM, that's what I was looking for. I had already updated the CSV file with the other project info, and wanted to add the project the XP2100 was working on. Got it!

As far as timings You can add another one, an AMD XP2100+, 1G RAM:

Protein............................Credit......PPD
p1850_Myosin6_PT_US......52.00......45.52

Most of the day it will be sitting here crunching, but I use it a lot at night.

 

[natethegreat]

I thought you wanted to know where to put the info not where I got it :laugh: Guess I need to pay better attention

Thanks for the info, I am editing it in now :thumbsup:

 

[RaySun2Be]

 

[Spacehead]

Computer info is same as my above post.

Protein: p2107_lambda_5way_melt_4_10011
404 points
65 ppd


Protein: p2355_ADAh2
683 points
166 ppd


I had 2 WUs in a row abort early. Is that common?

Project: 2374 (Run 0, Clone 29, Gen 7)

[20:11:44] Completed 190000 out of 500000 steps (38)
[20:44:27] Writing local files
[20:44:27] Completed 195000 out of 500000 steps (39)
[21:14:35] Gromacs cannot continue further.
[21:14:35] Going to send back what have done.
[21:14:35] logfile size: 32966
[21:14:35] - Writing 33502 bytes of core data to disk...
[21:14:35] ... Done.
[21:14:36]
[21:14:36] Folding@home Core Shutdown: EARLY_UNIT_END
[21:14:40] CoreStatus = 72 (114)
[21:14:40] Sending work to server

Project: 2369 (Run 1, Clone 53, Gen 4)

[02:43:31] Completed 20000 out of 500000 steps (4)
[03:25:48] Writing local files
[03:25:48] Completed 25000 out of 500000 steps (5)
[03:55:44] Gromacs cannot continue further.
[03:55:44] Going to send back what have done.
[03:55:44] logfile size: 9604
[03:55:44] - Writing 10140 bytes of core data to disk...
[03:55:44] ... Done.
[03:55:44]
[03:55:44] Folding@home Core Shutdown: EARLY_UNIT_END
[03:55:48] CoreStatus = 72 (114)
[03:55:48] Sending work to server

 

[natethegreat]

I have wu's abort themselves from time to time, I guess I would start to worry if it happened back-to-back...

Thanks for the info, it's going into the top post here shortly :thumbsup:

 

[GLeeM]

Originally posted by: Spacehead
I had 2 WUs in a row abort early. Is that common?

If that computer has been folding for a while, as I suspect, without problems, I would not worry about a few Early Ends.

Especially as there have been several projects that are having many of these. I guess they usually mean the end of the trajectory for that run or clone or generation or whatever. They still give valuable science for the researchers.

You get points for those that say "Going to send back what have done." Whatever percent finished they are is what percent of total points you get.

 

[natethegreat]

**Update**

added:
A64
Project...........................................................................|.........Core....................|.....Credit.......|................PPD
p2390................................................................................DGROMACS..................1112.........................275 Thanks Scott66
p2306_BBA5_mutant.....................................................GROMACS....................188............................195
p2306_BBA5_pfold in water.........................................GROMACS.....................188............................184


I noticed that a couple projects have been reported on P4's and HT P4's, the p2355 and 2387. On the HT cpu's both are giving 57% of the ppd of a single threaded cpu, and if you could run 2 of these at the same time on a HT cpu you would get a 14% ppd increase, but we already knew that didn't we
P4
p2355................................................................................DGROMACS..................683............................290
p2387................................................................................DGROMACS..................731............................397

P4-HT
p2355................................................................................DGROMACS..................683............................166 (166x2=332 ppd)
p2387................................................................................DGROMACS..................731............................229 (229x2=458 ppd)


Average ppd for project cores:

TINKER..........n/a
GROMACS......126
DGROMACS....237
AMBER...........97