• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

YURT Beta Testing

mikeblas

[H]ard|DCer of the Month - May 2006
Joined
Jun 26, 2004
Messages
12,777
If you've been watching the YURT Design Thread (and even if you haven't), the YURT program is about ready for prime time. In the next week, I'd like to collect about 25 beta testers for a wider test release before we call the thing finished.

What's YURT?
YURT is an idea that enhanced08 had to collect statistical information from Folding at Home clients. The program sends information about your machine -- processor info, memory, clock speed, and so on -- to a database. It also monitors your FAH clients to watch them complete work. As work units and frames finish, the program sends completion information to the site so the amount of time spent completing a given protein is tallied for each machine configuration.

The database has a website front end, where you can view a variety of reports to see how fast your machine is, how much work its doing, how often different proteins complete, and so on. How fast are Intel machines compared to AMD machines? How fast is a 3000+ compared to a 3400+? Does L2 cache make a difference? Which proteins are fastest? Which are the most popular these days? What kind of machines are participating in the project? Who has the most memory?

And so on... In fact, there's probably plenty of interesting reports we haven't written yet, but will do so as we start collecting more data and get user requests.

YURT doesn't cost anything to use. After installed, it runs as a service. It requires Windows; specifically, one of these versions or newer:

  • Windows 2000, Service Pack 3 and later (except Datacenter Server)
  • Windows XP SP1 and later
  • Windows Server 2003 SP1 and later

In respose to enhanced08's solicitations, I stepped up to write the code. It's almost done; it needs some more testing, but we're ready to release what we think is version 1. The testing part is where we need your help the most.

What is a Beta Tester?
A good beta tester is someone who understand computers well, and knows something about the application being tested. They take the time to read the manual and provide coherent feedback on it. They observe what the program does and comment on how it might be made better.

When errors or crashes happen, they don't freak-out: they collect all the information they can and report the problem so that it can be fixed. They work with the developer (me!) to get the problem reproduced, understood, and addressed.

They don't give copies of the beta to their friends who are outside of the beta program. A good beta tester realizes that beta testing is a responsibility, not a status symbol. Beta testers don't use illegal copies of Windows.

They try to put the program through its paces as a normal user would, and for this application, that's "set it and forget it". They also play around and explore, trying to break the program. They look at the features and suggest ideas for new ones, but don't throw a tantrum when their idea is punted to the next release.

Beta testers try out the software on a variety of hardware.

When and Where?
If you're interested in being a beta tester, please send an email to yurt at blaszczak.com and explain what you'll do to help. Or, make a post here with your information. Don't hesitate to ask any questions you can think of.

After I select the testers, they'll have a week (or two) to run and use the software. I'll try to fix whatever problems are reported, and we'll go from there. If we don't find anything super serious, we'll probably call YURT done and release 1.0 to everyone, take a break, then get busy on 2.0. Don't bother to beta test if you're going to be travelling, busy, offline, or indisposed over the next three weeks or so.

I'll expect to take a week making a list of testers, then send out the beta on the weekend of April 29. The beta will go out on May 6th (unlikely, since that's a race weekend), shortly thereafter (more likely) or the weekend of May 13th (less likely, though possible if someone finds a juicy bug). I figure the beta will last less than two weeks.
 
This is also reserved for future expansion and not a duplicate of a post that I made in the last five minutes. Seriously.
 
sent

and man you are long winded

edit oh yeah I also run windows 2k3 and win XP and can install it on both
 
Duster said:
and man you are long winded
It's disappointing to hear you feel that way. I doubt you'll like the docs, then—the above post is about half a printed page; the docs are more than ten pages.
 
oh i read it... just saying

i don't mind reading something just gives me something to do at work

 
I am interested in being a part of this beta.

Current machines on the farm:

Opteron 165 @ 2.4GHz
Dothan 1.73 GHz

Next week adding:

AXP M 2600+ @ 2.2GHz
Pentium D 805 @ 3.2+ GHz

I figure it's a pretty wide range, and it should get a lot of data out in the two weeks of testing.
 
Hey Mike,

I am interested in helping you out in this endavor. I have been around and have worked with computers for some time and am willing to take the time to try and break the program.
I can install it on about 9 machines at my home(amd and intel) and maybe up to 30 over time.

e-mail = william@ahpathfinders.com
 
I have a question that's not directly related to the beta testing but somewhat overall. Is this supposed to run on a machine that does nothing but fold? I have a few windows machines but one of them gets used frequently for gaming and another sees light use (work system). A couple more do nothing but fold but one of them I only see every copule of weeks or so.

What I guess I'm trying to find out, is a system that's used a lot or not always on (no, I don't have any of these abominations) going to mess with the results? It's an issue that we have discussed before when contemplating other projects like this.

 
When sending mail for the beta, please remember to include your HardForum screen name so I know who you are. Thanks.
 
SmokeRngs said:
Is this supposed to run on a machine that does nothing but fold?
YURT can run on any machine that's folding.

SmokeRngs said:
What I guess I'm trying to find out, is a system that's used a lot or not always on (no, I don't have any of these abominations) going to mess with the results?
You can find all the details in the design thread. At the end of the day, the idea is that the results will "average out" and show a reasonable prediction for performance of the system. If you're worried that you're poisoning the results, you can either not run YURT on that machine or pause FAH while the machine is "busy" and resume it when you're done.

Unfortunately, there's no way to watch how much CPU time another process actually gets that's both reliable and inexpensive.
 
Sorry if this is a dupe; I know many of you might not be reading the fully detailed design thread.

What's hopefully the last small beta went out just now. I'd like to give the core beta testers about a week with it, and then release it to the wider audience (that's ya'll). Then, after a couple weeks with it, we can smack it in the ass and send it out to the world.
 
I can beta test on a variety of machines... not all of my machines obviously since it would take weeks to install and some are production machines, but I could easily install it on ~20~50 machines depending what you want

lmk if you are interested
 
I've sent mail to everyone who asked. (If you asked and you didn't receive the mail, please send another note; it's just an oversight, don't panic.)
 
The first race of the season is this weekend (If you're in Portland this weekend, come by and watch!), so I'll be online only sporadically and not likely to be working on the software from tonight to Monday. There's only two issues that I'm tracking, meanwhile.

First, unhappy_mage noticed a problem with the *.DAT file when restarting the monitor application. I think this is fixed by a seperate change that I made.

Second, unhappy also noticed that YURTInstaller uses a different Internet connection library than YURTMonitor does. I'm going to fix YURTInstaller to use the same library that YURTMonitor does, so that it's easy to know if the proxy settings on the machine are correct or not.

With preparing for the race, I doubt I'll get this all done this tonight. So I should be able to drop our first release candidate at the beginning of next week. (Or, tonight, on the outside chance I have time to work on it.)

I'll reset the database fully, again, at that time. Then we'll let it run. If there's no complaints after a week, say, we'll call it version 1.0 and move everything to yurt.blaszczak.com, and release to web.
 
I've posted Version 0.94 and reset the database. You should have mail about it. (If you don't and are supposed to be a beta tester, send me a line; it's just an oversight.)

This is our first release candidate. If testing this goes well for a week or so, we'll call it good and RTW.
 
The ShowListOfMachines page now has sortable headers.
 
Work Rate by Frequency Range doesn't make divide by zero errors any more.

Rate Per Protein by Vendor doesn't make divide by zero errors anymore, and is now just narrow enough to fit on a 1280-wide display without horizontal scrollin'.
 
Well, I haven't heard back from anyone. Either the beta is going great, or nobody is actually using it.

Hoping for the best, I'll try to release YURT 1.0 this week; tonight, in fact. It'll be a bit of a challenge because I have a very busy week, but you should soon see an announcement here when I post the bits, flip everything over to yurt.blaszczak.com, and so on.
 
Everything is working fine for me Mike, and I love the sortable headers!
 
Thanks for the feedback, Sobe. I think there's a couple more reports that need sortable headers, and I'll try to add them in the coming weeks. Also, there's a bit of a bug in that we default to the opposite order of sorting when changing columns. (That is, if you sort by A ascending, then click on B to sort by that column, it is descending. That's a bit jarring.)

If anyone has ideas for new reports, or notices problems with the existing ones, please don't hesitate to let me know.
 
Added a box will add more later. It looks nice and I look forward to seeing it filled with statistics.
 
Back
Top