Fresh hacks every day
It has only been a month since [BillyTheImpaler] posted his Folding@Home How-to and the Hack-A-Day team has already completed 1,000,000 points. Great job team!
Hey great job guysgals!
I’m getting as many boxes as I can running for Team Hack-A-Day. All willing friends, family and clients that run it all day (and night!).
Sweet. We’re currently averaging 42,000 points per day which puts us in 27th place worldwide for that category. to get into the top 100 we’ll need more than 4,000,000 points which eoc says we’ll have in about 100 days. We’re certainly among 31337 company with numbers like these.
Anyone intrested in creating a Hack a day seti @ home team???????
… if we start runing seti at home that will cut in to cpu time for folding at home.
Join the Folding @ Home team!
We want you!
I think a SETI team would be ok. FAH, in my opinion is just a better porject. If I was devoted to BOINC I’d run some other project like Einstein or LHC and wait for Folding@BOINC which was in closed beta but has since gone dormant (http://fah-boinc.stanford.edu/). Rumor has it that with FAH6 FAH BOINC will debut but Vijay Pande will not comment.
I used to do SETI but came to terms with the needle-in-a-haystack nature of the project. I was tired of a week of downtime every month. I decided that FAH was the project for me becasue I think it’s the poeject most likely to benefit humanity in the here-and-now. teh client is very flexible. You can sign up for a wide variety of work. From timeless WUs that use 5mb ram all the way up to QMDs that use 325mb ram. there’s a FAH setting that’s well suited to every machine i own, from my P2 400mhz all the way up to my dual 3.2 xeon workstation.
Berkley’s servers are down about 15% of the time. Becasue of this, there are all sorts of crazy WU chaching schemes beginning with setiqueue and bringing us to the BOINC platform of today. Stanford has more than 98% uptime due to it’s smaller size and greater scalability. The fact remains that SETI was supposed to be a project that would run and then be over. It has continued to re-scan the same part of the sky over and over again. Mind you, they’re gathering new data all the time, but it’s still the same 30% of the sky that is scannable by the arecibo observatory. (http://en.wikipedia.org/wiki/Arecibo_Observatory)
I think a SETI team would be good but I’d rather see those cpus go toward FAH as we’re already established in this arena and are clawing our way to the top.
*disclaimer* I’m a huge proponent of FAH and probably h-a-d’s biggest fah cheerleader.
Great Work. Guys.
Woot I did my part of 1000… hehe im werking on my fif one WEEEEE GO HACK-A-DAY!!
now I’ve tuned my network over the folding…
so far there is an inspiron600m and a dimension 4600, but a few more p4s are on the way…
The only problem is that I am getting programs that are uber small; using 8mb (worth 41pts) or ram on the laptop and 20mb(worth 149pts) on the desktop… Is it just my processing speed that is to blame? Or am I missing out on some pointage…
I told it to use >5mb proteins…
#10 What’s the processor in your machines? How much Ram do they have? If they too little, then it won’t get certain units. Other than that, it’s probably just having a bad stretch. I went a few weeks on my Smelleron without seeing a Wu worth more than 100 pts and now all I see is 600s. Sometimes it’s the luck of the draw.
Is -advmethods on? Is -forceasm on?
When the process starts does it say:
[12:37:44] Writing local files
**[12:37:44] Extra SSE boost OK.**
[12:37:45] Writing local files
[12:37:45] Completed 0 out of 1000000 steps (0)
If it does not mention that SSE is on, then you’re not using assembly optomisations which will cost you.
I used to be a part of Team Dopefish in SETI@home, but when they upgraded the client and I tried to activate my account, I got errors. Cheap bastards. I’m completely devoted to Team Hack-A-Day in F@H. Great work, guys!
i have access to about 400 pentium 4′s at my college. about 50 of which have over a gig of ram each. most of the rest have 512. Unfortantetly the borging tactics you described in your tutorial dont work on these machines. Do you think it would increase the chances of sucess if i ran my ‘borg.bat’ while in windows safe mode?
this is a novell network.
i have 2 machines folding in my room. but can you imagine the speed and ferocity of 400 or so pentium 4′s that just sit at a login screen 99% of the day.
ill play with it this afternoon, if you have any tips let me know
Jslice, do you have admin rights on those machines? That’s the most important thing to consider when b0rging. If you could, describe the exact way the system there works. I.e. user rights are such, the hard disks are automatically reimaged at the end of the day that sort of jazz.
If the drives are remiaged daily or some such thiung, it’s possible, as I recall, to set up FAH to save all the work and that sort of jazz onto a networked machine and have all the rest copy the work data over to ramdisk. I’m not sure how well that would scale to a b0rg of that size though. You could snag an older machine from central stores (i doint know about you, but they hand out p3 systems like they’re water around here) and set up a network drive, then create a script to run on each borg that automagically maps the network drive. You can hide the drive from users in Window$.
This all relies on you having admin rights.
Give us the specs here in the comments and we’ll try and work out a solution for you.
What exactly happens when you try to run the .bat file? If you can’t read fast enough to know what it’s saying, take a screenshot while the code flies by.
BTW, a b0rg of that size would probably put you in the top 10 worldwide individual producers!
a pause at the end of the batch file may also help you to debug. it’s a bit easier than taking a screenshot, eh?
wake me up when you cure cancer…
I’m running some computers on the team as of this post…
But, at work I have to do ‘cpu burn-in’ tests to test computers with heat problems….
We usually use a DVD and let it run overnight…
I was thinking that it would be cool to use something like folding@home so it would actually be doing something productive. I tried to make a boot floppy that boots the ME boot disk, and runs the win98/ME console version, but it won’t run in “dos mode”.
I can get it to run on knoppix, but to save partially complete work units, I have to have a writeable media like a usb drive.
Maybe someone can make a small linux boot floppy that will work? Or a distro like overclockix that doesnt suck ass, and happens to boot right into the folding, with a config file that can be set to user name and hackaday group, etc….
Let me know if anyone has ideas. I know network access can be a problem with the wide array of cards out there.
I’ll send in a ‘tip’ if I get anything working.
I just joined the HAD f@h team. My CPU temp is around 51C. What are you guy’s CPU temp at? I am wondering if it is bad for it to be that hot for a long period of time and if I should set up a better cooling situation.
play alot of soccer or something? (lol)
pat, please go read about the post before you post dumb questions we all have to read. “Points” are awarded to you (and your team) when you complete work units. Without adding too many details, you pretty much get more points the more complex the work unit is. This is how you can keep track of how your team is doing compared to other teams… so you can see how your help elevates the team’s standings, etc.
it would be very simple to make a linux bootdisk that would boot straight into the client but it would still need a place to save partially completed work units.
#21, there already is such a linux distro. it’s called Overfoldix and can be found here http://sourceforge.net/projects/ofdx
It runs like knoppix; without a hard disk. It has a bunch fo utilities to help set up a folding farm too. Try it out if you like, it’s quite cool for a dedicated folding machine.
continuing from post 11 by billy, they both have .5GB of ram, the laptop is a PM 1.66GHZ, the desktop is a P4 2.66GHZ
I get ‘extra sse boost ok’ message, and the desktop has both -forceasm and -avmethods on, and the laptop only has -forceasm since you said for PMs you had better luck w/out it…
The first set of proteins finished, and now they both got measly 50 pointers…
Thanks billy, that overfoldix sounds about like what I want… only thing is, how does it save partially completed packages? Is that possible? It would be cool, if no other solutions were avalible to use gmailfs on linux and automount a gmail account for saving partially completed units. then again, with a time constraint on so me of the units, you may not get to finish them anyway if it is a week or two until you resume the process.
smartwhiteboy243, if these are among your first units, don’t sweat it. You won’t always get the best work that’s worth the most points; it’s the luck of the draw. No matter what settings you enter sometimes you won’t get the primo units that you could be getting. If you’re not seeing any good work in after a few weeks, then I’d be concerned. If you run FAHmon, leave it alone and check your benchmarks to see what kind of ppd you’re seeing. If it’s anything over 100 on average, you’re going to be ok.
Tired2 and drpepper, I
Since this is a tiffy post I figure that this is fairly relevant.
If you want to post a tiffy and want to spice it up a bit, try adding &bg=x to the end of the URL. x can be a number from 0 through 7 to give you different pictures in the background.
so yeah now everyone should join my team laff team:46894
now to set up my borgs instead of using them as a clusternode lol
Fill in your details below or click an icon to log in:
You are commenting using your WordPress.com account. ( Log Out / Change )
You are commenting using your Twitter account. ( Log Out / Change )
You are commenting using your Facebook account. ( Log Out / Change )
You are commenting using your Google+ account. ( Log Out / Change )
Connecting to %s
Notify me of follow-up comments via email.
Notify me of new posts via email.
More Posts from this Category
Return to top of page
Copyright © 2014 · Genesis Framework · WordPress · Log in