News February 2003

Saturday 01/02/03

Top 500:

Last weeks xylus 500 has been uploaded.

 

Monster WUs:

Overnight Stanford released some monster work units with a range of 47 to 73 points. These huge units have deadlines to match as well, 50 days. Now that's something chewy to get your teeth into.

 
  P626_vfd_nat: 73.0 points
  P627_vfd_ext: 73.0 points
  P628_vsd_nat: 55.5 points
  P629_vsd_ext: 55.5 points
  P630_vmut_nat:47.0 points
  P631_vmut_ext: 47.0 points

Also there were a couple of small one pointer Gromacs released for the beta users.

 

KDFold V6.0 Beta 4 Web Enabled:

WinKDFold (Windows) V6.0 Beta 4 F@H 1,2,3 G@H 1,2(F@H3) DF & F-A-D
Release 6.0 Beta 4 01/29/2003
Monitor Multiple Projects and Clients from one application instance-
KDFold Currently Supports- Folding@home 1,2,3 & G@H 1,2(F@H3) F@H & G@H
Autoswitching in newer clients & Distributed Folding & Find-A-Drug

1.By user request I have modified the way clients are added into a step by step process
with hints at each step. I have also included an autosearch function to quickly locate
the proper directory.
2.By user request I have added the Folding 2&3 stored count to the quickstats in the
progress bar.
3.Added a web server to KDFold for simple viewing and remote control of the desktop
from any browser. KDfold can also function as a regular simple web server.
4.Via web You can view and Click on the desktop or the various stats file.
5.Remote by clicking on still images(Jpeg's) of a remote desktop those clicks are
applied on the remote machine.

HTTP Quick Help-
You will find the web server settings on the Web Services tab.
Try on your local machine by
http://localhost
Remote machine by IP address or url or machine name.
IP/URL/Machine Name example
http://192.168.1.2 http://www.kdfold.org http://kidsiii

Download WinKDFold v6.0(Windows)
http://www.kdfold.com/winkdfoldv60b4.zip

Home Page-
http://www.kdfold.com

Support Forum-
http://www.plasticstudios.org/forums/index.php

Forum thread here

 

Sunday 02/02/03

Top 500:

This weeks xylus 500 has been updated.

 

Thursday 06/02/03

New Gromacs core FAHCore_78 important:

Stanford released a new Gromacs core (FAHCore_78) today and I found out the hard way that you need to have the latest client 3.24 for it to work properly. I was using 3.14 and after downloading the new core it would come up with an error and download it again. Fortunately 3.14 downloads the core 5 times then it goes to sleep.

[Feburary 6 10:27:01] Working on Unit 02
+ Working ...
[10:27:01] Folding@home Gromacs Core
[10:27:01] Version 1.41 (January 28, 2003)
[10:27:01]
[10:27:01] Proj: work/wudata_02
[10:27:02] Looking at optimizations...
[10:27:02] Created dyn
[10:27:02] Files status OK
[10:27:02] Error: Work unit read from disk is invalid
[10:27:02]
[10:27:02] Folding@home Core Shutdown: CORE_OUTDATED
CoreStatus = 6E (110)
+ Core out of date. AutoUpdating...
- Attempting to download new core...
+ Downloading new core: FahCore_78.exe
+ 10240 bytes downloaded

Download the new client from here http://folding.stanford.edu/download.html

 

Friday 07/02/03

DG closing in on 150K:

At the 17:00 WST Stanford update DGROMS.com had 149,500 points and closing in on the 150K target really quick. So the Race To 150K will be over sometime tonight with plext some 23,000 points behind. Well done DG that is an awesome weekly output you have there and you have really put the foot down over the past couple of weeks with the availability of the Gromacs.

 

Saturday 08/02/03

FAH Benchmark Top 100 list:

Since we now have a reliable folding benchmarking tool for our high power super computers I thought it would be a good idea to have a Top 100 list of the fastest computers folding for OCAU.

Post your benchmark scores and machine details like this

Machine: XP1800@1669Mhz (Palamino)
Benchmark: 5488

Machine: Cel900@1063Mhz (Coppermine)
Benchmark: 3296



To do this you need the latest Folding client 3.24 and the lastest version of EMIII (v2.3.1). The benchmark information is stored in the queue.dat info which you can access through the first options page of EMIII.



I will collect the scores and put them on the Statsbar above broken down into the Top 25 and Top 50 etc like the other stats. So post your fastest benchmark or email them to me here and who knows you could add another line to your forum sig like this Fastest Folding@Home Benchmark Score xxxx or even Folding@Home Benchmark Score xxxx.

 

FAH Logstats v0.3 Released:

Thanks to Buzzard our local programming guru he has released a simple monitoring program that was aimed at users have Smoothwall boxes. You can grab a copy locally here as well as from Buzzard.

Both EMIII and KDfold are good programs for monitoring but I wanted something simpler and needed something that would easily monitor my Smoothwall/IpCop box ( to use EMIII you need to set up samba ) so I wrote my own monitoring program that only used FAH log file.

Below is an example of the configuration file:
 

code:

"XP2100+","E:\Folding@Home"
"K6-III 450","\\The void\FOLDING@HOME"
"Smoothwall","E:\Folding@Home Stats\Smoothwall"
"Dual PPro 200 CPU1","\\DualPPro\\F@H\CPU1"
"Dual PPro 200 CPU2","\\DualPPro\\F@H\CPU2"

 

The first part is just the name of the computer. The second is the directory (without a trailing \) of the F@H client. The config file (config.ini) and a temporary log file (templog.txt) are saved in the same directory as the program.


Setting up for linux
To use with Smoothwall/IpCop, you have to set up a FTP server on a computer, and get the linux box to upload log file every 10minutes. To upload the file I used ncftp and a cronjob.

code:

*/10 * * * * /home/ncftp/bin/ncftpput -u smoothwall -p fahstats 
192.168.0.50 / /home/folding/FAHlog.txt > /dev/null &

 

The above is probably a very bad example but it works for me

Download here: http://www.holohost.com/~mdavey/Console Stats 0.3.zip

As always... Comments and suggestions are appreciated ( Along with bug reports )
 


 

 

DG wins the race to 150K:

DG passed the 150k mark at 23:00 WST yesterday and won the race to 150k. He is going to keep the engine running at full RPM and try and get to the 20,000 Folding Club. Well done David.

 

Sunday 09/02/03

Top 500:

This weeks xylus 500 has been updated.

 

Monday 10/02/03

FAH Benchmark from the [H]orde.

Over at the [H] forum the highest benchmark score is 7488 from MrMgrMatt and his machine is an AMD Athlon XP 2700@2366MHz. Which is just a tad faster than Teddy's P4. So the gauntlet is down to find the fastest Folding machine, I think DG has a some quick machines that should nudge the top spot when he gets around to posting his scores.

 

News from TheWeatherMan:

Work Unit news.
New proteins released and problems
More new proteins have been released, but one of them, the p395 Gromacs protein, has been causing serious problems. As far as anyone can currently determine, no one has been able to successfully return one of these work units. They have been pulled by Stanford and they are looking into the issue.

One other issue that I have reported it the possible shut down of the gromacs core during the folding process sometimes causes the work unit to fail a test when it completes. Just be aware that there may be a problem. My suggestion is don't shut down the client if it is doing a gromacs work unit until it finishes and sends it in. I have currently switched almost all of my systems back to just using the Tinker core. That means running the client without any flags.

Link: http://home.attbi.com/~wxdude1/emsite/

Download New Protein file for EMIII here

 

FAH Benchmark Top 100 list:

I have added new pages to the benchmarks list above. It now contains separate lists for Intel and AMD processors as well as the overall list. Post your scores in the forum thread or email the screen shots to me.

 

Wednesday 12/02/03

New Folding Club Badges:

There are some new Folding Club badges. the rest will gradually be changed over to the new design.

 

New Protein design core

It looks like there is some new proteins on the way with the release of a new core (FahCore_ca). I'm not sure if this is just for the Genome users or for general distribution, time will tell.

We're very happy to announce the release of our new protein design core. Designated as FahCore_ca, it's been in beta-testing for several weeks, and it's now stable enough for general release.

This core allows us to greatly augment the scientific work that we began with the G@h project, and Core_c9. The new protein design engine allows us to tackle the design of much larger proteins, and even complexes of several proteins, a very exciting new research direction for our group.

I'd like to thank all the beta-testers for their input and bug reports, all of which were very helpful.

Good luck - Stefan

 

Gromacs on hold for a while:

The Gromac proteins have dried up for a while from server 114.

Server .114 will not be assigning or appearing on serverstat for a short period.

It's back in serverstat again. We'll use that server to test out a new project before bringing it to advanced methods in the next day or two -- in the meantime, it will continue accepting all work units outstanding.

 

The future of Folding@Home:

There is an interesting thread over at the Folding Community forums regarding what the future holds for the FAH project. Here is Vijay's replys

This is a good question. First and foremost, my job is to plan and architect for the future. I'm really looking forward to this speed increase (both from Moore's law as well as from more clients perhaps). We have roadmaps on all levels -- web, server, client, core, db backend -- to handle this.

What really helps too is that as we get more and more prominent, several vendors have been offering us funds, great discounts or donations of hardware. We'll be announcing a major one pretty soon (I think -- we've been ok'd for the donation and are just waiting for h/w delivery).

I'm really excited about the future. There are several "landmark calculations" -- comparable to the successful results published recently in Nature and other journals -- that are on that road map.

 


I think we can safely say that F@H has been scientifically the most productive of all science-based DC projects by far. In a time when the science of the space missions are constantly being examined ("what did we get for all those billions of dollars spent?"), it's interesting to ask various d.c. projects "what results have been learned considering all of the billions of CPU days spent".

We can point to major scientific results, in prominent journals (Science, Nature), and with a lot of press activity, and donor, academic, and industry interest. I'm really proud of my team, the forum mods, and the F@H community as a whole for making this a reality.

 

Friday 14/02/03

Milestone Badges:

You can now collect Milestone Club badges to go along with your Folding Club badges. They are available in a range of colours to mix and match with your current Folding Club badge. They make nifty forum avatars and can be animated to show off your Folding achievements. They can be found here, what are you waiting for start your collection today.

 

 

Statsman:

Statsman has going missing again, but some how I managed to get access to the html stats from 10am WST this morning. I have saved the page here for all the stats junkies out there.

 

 

Electron Microscope III v2.3.2 (Update):

-nostealth flag added
13 client view now shows the box name
Computer status color change
Added CPU Type & Speed under Benchmark score
Made another change to Genome points calculation

So many people accidently put EM into stealth mode (completely invisible) and don't know how to get it to reappear, that I have added a new flag, -nostealth. This is just a method you can use to clear the internal stealth setting when you start EM without clearing all the other information. There is also a new warning message that will pop up when this tray option is selected. I'm hoping this will cut down on the email I get on this issue.

I finally got around to making EM display the box name when using the 13 client view. This makes it a lot easier to identify the box. Should have done it a long time ago, just now got it the way I like it so I'm passing it on to everyone.

I changed the computer status icon colors that show everything is fine. Light Green denotes a good Tinker core running, Light Blue denotes a good Gromacs core running, and Light Purple denotes a good Genome core running. Now, at a glance, you can distinguish what type of core is running without really having to look at the protein to figure it out. The latest snapshot of EM running shows a sample of the status icon showing a Tinker core and a Gromacs core running.

It seems there is a lot of activity centering around the benchmark scores EM displays as a result of the test run just before each new protein is started. Note that as far as I know there is no relationship between the protein you are about to work on and the benchmark score. There does seem to be a huge difference if the advanced assembly loops are not active on your system. If you start the client using the -forceasm flag it will override the fallback to normal loops. The flag of course can't give your CPU things it doesn't have, but the client will try to used the advanced methods if possible. This can improve your score, and the rate at which you finish a frame will be reduced when running the Gromacs core.

I added the new Genome points calculation formula, but it seems the format of the file I use to determine how many Amino Acids (the chain length) keeps changing. This makes it difficult to determine. Once I see that it doesn't change I will fix the points display issue. EM will track the protein, it just won't display points and the number of Amino Acids. Hopefully by the next release this will remain constant and I can fix it.

A big thank you to everyone who has contributed a few dollars to help me keep the EM-DC web site up and running. Every little bit helps and I appreciate the donations.

Get the latest version here:
http://www.em-dc.com

 

 

KDFold V6.0 Beta 5:

WinKDFold (Windows) V6.0 Beta 5 F@H 1,2,3 G@H 1,2(F@H3) DF & F-A-D
Release 6.0 Beta 5 02/13/2003
Monitor Multiple Projects and Clients from one application instance-
KDFold Currently Supports- Folding@home 1,2,3 & G@H 1,2(F@H3) F@H & G@H
Autoswitching in newer clients & Distributed Folding & Find-A-Drug

1.By user request you can now hide the kdfold icon in the systray.
See support forum on how to use this feature.
2.Fixed some missing stat info on Gromac cores.
3.Added support for the new Genome cores.
4.Because of the core change to Genome I have not quite finished some DF
requests, please be patient with me I am trying to get these added.
5.I have delayed converting stats and other files to XML so I can add point tracking,
once again please be patient this is high on my list.

Check the forums for all of the changes and keep current on KDFold News.
 

Download WinKDFold v6.0(Windows)
http://www.kdfold.com/winkdfoldv60b5.zip

Home Page-

http://www.kdfold.com

Support Forum-

http://www.plasticstudios.org/forums/index.php

Have fun,
Robert-

 

 

Saturday 15/02/03

1 million points behind:

[H] have now got a lead of over 1,000,000 points, my how things have changed since the middle of last year when we were neck and neck. Now we have to worry about oc.com overtaking us, at the current rate of production from both teams they should over take us on the 13th of August or 179 days from today. I think its time to do something about it and so does javascripterror, he sent in this picture to try and get a recruitment drive going and bring in some new members.

 

FAH Benchmark:

nikhsub 1 from the oc.com team has posted the highest score so far, his XP2600 is running at 2525Mhz with a 210 fsb.

 

Statsman:

Statsman explains what's been happening with his stats...

My hoster has moved my site to a different server and IP address, if you are getting 404 errors try running 'ipconfig /flushdns' and rebooting your machine. Then try to access my site again.

Also if you are behind a proxy server you might have problems until that server flushes it's dns cache.

 

Sunday 16/02/03

Top 500:

This weeks xylus 500 has been updated.

 

Flogometer:

Since oc.com now have us in their sights and are only 139 days away from catching us. I thought it would be appropriate to have something to keep tabs on the gap and show how many days we have left in 2nd place. Ladies and gentlemen I present to you the Flogometer, the Flogometer will be updated daily and contains the daily deficit or gain by us, the overall gap and how many days before they catch us. Basically it measures how much we are getting flogged by. But hopefully we can increase our output and prevent such a catastrophic event from happening. We can do this by each of us doing our best to attract new members and by using all the available resources we can get our hands on.

 

Monday 17/02/03

The Kaliban Factor:

All today's stats have been skewed by something going wrong with Kaliban's stats. He has dumped over 17,000 points for the day with 8 machines. Here is his explanation in the Forums. Hopefully the stats get sorted out soon so that we can get things back to normal. Besides that he broke the Flogometer, I think I will have to order the deluxe version which has more features and is able to handle big numbers a bit better.

 

Saturday 22/02/03

EM Protein Calculator (v 1.0.0):

Apart from Kaliban's effort this has been a quiet week on the folding front but never fear TheWeatherMan is here and he has a new toy for us to play with. Good onya Larry

This tiny little program calculates the number of points per hour, per day, and per week a computer can complete under ideal circumstances. By "ideal" I mean that it doesn't take into account upload and download time when sending and receiving work units.

If you have EM III installed, simply unzip the calculator into the same folder EM runs in. EM reads the emprotx.dat file and fills the drop down box with all the protein names. When you select Folding Work Unit, the default, then select a protein from the list, the calculator will put the number of frames and the points for that work unit into their proper places in the calculator. Since Stanford has released a few proteins with the same names and different numbers of frames, you might have to adjust the number of frames after selecting the protein. You can also just ignore the protein selection and put your own values into the appropriate text boxes. The item you need to be careful with is the "Time Per Frame" information. The program wants it entered one way only. Hours, minutes, and seconds all need to be entered. I will probably make this part smarter in the future, but for now make sure you enter the information exacly like EM III displays it. If the frame time is 3 minutes and 8 seconds a frame, you enter 00:03:08 If you enter 3:08 it won't work. Once the information is entered, press the Calculate button. Your points per hour, day, and week will be displayed.

The Genome calculation is very similar, and you don't have to select "Genome Work Unit" if you want to enter everything by hand. But, the program will calculate the number of points the Genome work unit is worth if you select Genome Work Unit, then enter the Number of Amino Acids (chain length). You do have to know that information for the calculator to figure out what the work unit is worth. The program is using the latest formula from Stanford so it should be pretty accurate. The default value of 30 frames is assumed, but you can change that once you have selected to do a Genome Work Unit.

That's all there is to this program. It's pretty simple. I wrote it because people are always asking how much a Genome work unit is worth, or how many points per day can a certain computer crunching a specific work unit put out. This little program can pretty much answer your question.

You don't need the emprotx.dat file, and the program will run from any folder on your system. If you click the EM logo a browser will be launched and the EM-DC web site will be requested.

Enjoy,

Larry

Get it here:
http://www.em-dc.com


 

 

Monday 24/02/03

Top 500:

This weeks xylus 500 has been updated.

 

Tuesday 25/02/03

HideIt X v4.0 :

Another of TheWeatherMan's programs has had a makeover. This time HideIt X has gone had the Larry treatment and has come out with a new look and a few more features.

HideItX - Starts, Hides, Shows, and Kills any program (almost) in Windows 95/98/ME/2K/XP!

Due to popular demand I changed the "No Smoking" icon to a "No Eye" icon. That should make a bunch of people happy.

I fixed a small issue that only rarely showed up. Sometimes the program couldn't find a window. That should no longer occur. Since it didn't happen very often, the majority of users never saw the problem.

Two new commands have been added so you can make any window the top most window or remove a window from top most status. When using this command HideItX will not set focus on the window, it just makes it the top most window, or removes the top most status. When you give this command, no matter what condition the window is in, even hidden, the window will be shown, then made top most. Even if you minimize the window, it will remain the top most window when you restore it. A cool feature I think.

Download it here (download area):
http://www.em-dc.com

Instructions:

You can put a shortcut to HideItX in your startup folder to start a program when Windows starts. Put a space between the command and the program info. If you are hiding, showing, or stopping a program, be sure to put the exact title of the process (Upper and lowercase letters don't matter)

Startup Commands :
-start......Starts any listed client. (use full path info)
-starthide..Starts any listed client. (use full path info)
-hide.......Hides all open listed client windows.
-delete.....Deletes any file. (use full path info)
-stop.......Shuts down all listed clients.
-show.......Shows all listed client windows currently running.
-ontop......Shows the client and makes it the topmost window.
-offtop.....Shows the client and removes topmost status.
-yespill....Restores UD's icon to system tray.
-nopill.....Removes UD's icon from system tray.
-help.......Shows the Help message box.
-delay......Puts an X second delay before next command is executed.

Examples:
-start c:\Program Files\F@H2\FAH2Console.exe
-starthide c:\Program Files\F@H3\FAH3Console.exe
-start c:\Program Files\Genome@Home\GHClient.exe -nonet
-hide FAH2Console
-ontop FAH2Console
-delete C:\Program Files\Distributed Folding Project\foldtrajlite.lock
-nopill (nothing else, UD specific)
-stop FAH2Console
-delay 5 (delays for 5 seconds)

Example of a Multi-Command line:
-start c:\Program Files\United Devices\ud.exe;-delay 2;-nopill

Since UD has some unique problems associated with trying to manipulate it, I have had to add 2 more unique commands that are UD specific.
-hide UD
-show UD
-stop UD Not currently implemented

 

 

 

 

Return to Main Page.