it is gentoo-amd64.
Printable View
it is gentoo-amd64.
oh yeah you said what CPU you had, didn't you :). That's what I get for only reading threads in reverse order.
linux64 SMP client is def. the way to go for points. I am only using 1/3 of my cluster for folding, all on 64/SMP and it outdoes what my entire cluster did running the old linux client by 50%.
the part of my cluster that is running the client has dual single core opterons in each node. They seem to turn a little less than 1000 ppd each node, something like that.
THere are 4 processes but on a dual core system it assigns 2 per core. It works fine, I'm running the SMP client on a s939 AMD opteron (basically the same as an X2 except 1MB L2 cache standard and typically more OC headroom hehe).
count me in.
I just downloaded the folding@home program. I will start with one machine for now but may add a second machine after I finish doing some work on it.
:welcome: cvoid!
:welcome: to the Team...
welcome, glad to have you cvoid, what brings you in to the folding team? Tell us about your hardware! we're all up for some geek pr0n. :D
:welcome: :wtf:
The reason I picked the folding team was I wanted to do something on my computer when Im not there, such as at work or when Im in school, and this seemed like something worth doing.
Right now I only have one computer running the folding@home and its an older computer:
Mach Speed motherboard (don't remember model number and since Im at work I can't just look at the box :( )
AMD Athlon 64 3200+
2.5GBs of Kingston RAM
Windows XP (may switch it over to run on Suse linux, but when I downloaded the program I had windows up and was also downloading some other things so I didn't want to reboot at the time to set it up on that)
The computer I am working on and will also probable put folding@home on is:
Intel DX38BT motherboard
Intel QX9650
2GBs of Kingston RAM
Windows Vista Ultimate 64bit version
The AMD machine I can't really overclock since it becomes unstable on anything past 2.1GHz but hopefully with the intel extreme I will be able to overclock it to at least 4GHz
hey wow im in the first post, was ranked 25! what happened!?!?!? :D
You fell down... lol I'm dropping my quest for FM's top 20 ( currently just made 28 ) since I moved production here. Will probably put the c2d back for them when I finally get a proper clocking board for this quad :up:
Looks like I can make top 20 here to, never thought XS would allow a single cpu to get on the daily top 20's already was ashame my lone e6600 on occasion led FM's daily's let alone XS :rolleyes:
Fold on people, make me a litle bit more humble :cool: :D
Edit: skype?
Edit2:
I don't want to step on anyone's toes here, but if you guys want to keep a list off active folders and what their running, I'm doing that on FM to and wouldn't mind doing it for XS. Just seems the current roster isn't being maintained atm, it's a diffrent setup also then what I was used to but if you tell me what you want to see I'm sure I could be off help :)
Well, in addition to the obvious stuff, I think it would be cool to have some kind of "horsepower" metric, like a GHz-to-WU ratio number or perhaps average frames/day. I'm really new to all this, but I am curious how to get a folding system that provides the highest return in computational horsepower versus watts/heat/system cost...
I find the "calculation flexibility" aspect intriguing too. I think it would be great if the F@H client could produce some additional stats around types of WUs being performed so we could have teams configure different client types to actually race towards solving particular problems that are high priority....
e.g., if our team could get bonus points for hosting a richer set of hardware/client types, then we could put more of the most valuable calculation horsepower online and thus benefit the project more. Am I making sense? :cool:
I think you are looking for something like this site http://fahinfo.org/
And I believe it's best if Stanford doesn't give us too much freedom. It's not a game they are making and imo it's a lot better if they put time and effort in making faster clients than giving the users a lot of options which not everyone would need.
http://fahinfo.org/index.php?avgscores=true
Points per day per Ghz metrics as per request. Power metrics depend on your particular OC setup.
You change what clients you receive by the flagging of the executable. Computers that can handle the SMP clients are of the greatest interest at the moment I believe and they require a separate client. Most of us who are active in this subforum have tweaked out gear for best results, I think you can see that by sorting through the active folders in the top 20 sorted by daily production.
holy folding stats, batman! That is cool stuff!
I'm feelin like a dunce now
:lastweek:
I'll just start folding now, though I have a crappy Celeron D, kept @ stock. I can only fold w/ 50% CPU usage, though due to temperature issues.
EDIT: Folding name is my username
That's OK, whatever you have to bring is great! :) The point is that you are folding, that's what is important :up:
Welcome to the team FF>IE :welcome:
Hopefully you more time for posting too. Quite a difference between for example you and Sparky :D
Welcome FF>IE!
Don't worry about what you fold with. Entrance to this club is cheap :). My HS economics teacher used to say "That dime might keep you from becoming a millionaire" - which is to say that no matter how fast you are going through WUs, yours might be the one that unlocks the key to curing cancer.
Thanks for the contribution and welcome to the team!
I'm in. Sorry it took me so long to pull my head outta my arse and put my computers idle time to good use.
real good to have ya - no matter what!
Add me to the roster too... Bacon1187 for the name though.
Welcome aboard mates, take one
http://www.jeremylatham.com/images/c...cookie-jar.jpg
PS3 24/7 and my current rig (see signature) F@H SMP client. Still working on stabile OC, but slowly getting toward 3.6.
spider_d for team 36362
Howdy :D
boingg!!! off the 70 second rev limiter again....
I'm folding project 2665 on the SMP client now and wondering if I can increase the spped from about 13 minutes per percent. Also, I read the thread about 2665 and don't fully understand the issues, or how I am affected by them. Is good science still ahppening regardless of the points?
And SparkyJJO, What is the 70 sec rev limiter that I keep seeing you mention?
Good science is still happening, yes. It is just a larger, longer unit and the points they assigned to it end up being less per day than some others. I just sigh a bit since I do like to see the numbers go up but the science is important and it is still happening :up:
The "rev limiter" I keep referring to is the the minimum time between posts. Many sites have 30 seconds between posts, but XS a couple months ago went to 70 seconds. I tend to read fairly quickly and I can type quite fast so when I have a few different threads open or heck just browsing around quickly I tend to bounce off of that a lot. I just blame overclocking my brain and fingers too much for the site to handle :D
I thought the post limiter had something to do with it. Thanks for the info.
Do you know a good SMP client guide for a novice. I have it folding, but I don't know how to use the cmd line stuff and have had trouble finding a guide. My WU is at 50% now and I am hoping to fold solid for a while. I feel like my stats are terrib;e even with the SMP running and the PS3 24/7 folding for a full day. Am I expecting to much or is 70 points per day going to be it?
With the SMP you'll get much more than 70 PPD, much much more. Just when you first start getting things going the stats take a while to come up to full running points. You'll be well over 1000 PPD, closer to 2000 I'm sure especially as you bring those clocks up!
CHeck in my sig, I have some setup guides linked, I believe there is an SMP setup guide in there :)
Is SMP stable now? It was eating WUs before and I feel nervous leaving my 2x text.
Not having any problems here w/ it. I'm running dual SMP on my dedicated quad cruncher and single SMP on my main rig.
If the team could use somemore CPU's crunching for the team, piont me in the right direction and Ill be happy to help. Have 2 quadcore desktops and a C2D laptop with 6x HD3870's between the 2 desktops (If you can Fold with them also?) that run 24/7. Never done anyfolding so might need a little info.
We always welcome more :welcome:
You can fold with the HD3870's for sure, however i'm not sure on how to set it up or anything since I don't have one of those cards. For just CPU folding though, if you look in my signature I have a link to setup guides :)
ok thanks, Ill see if I can get them set up by tomarow.
Anyone else happen to know if you can fold a ati gpu and cpu at same time? and if so, what about 3 gpu's and a cpu?
Each GPU client requires a CPU core.
ok, last question then (I think). so that means basicly I can eather use my 3 gpu's or just the 9850 phenom. wich would do more?
I think the 3 GPUs would do more. You could actually do 3 GPUs and one standard CPU client to use the last core on your quad.
cool, have 2 systems just like that. looking forward to putting them to use.
Sounds good! I know some here use the new GPU client, so I'm sure they can help if you have any issues :)
they dont seem to have the gpu2 client for vista at the moment. says there working out some bugs so guess ill be using cpu only for awile
Works fine for me in vista64. gpu2 for nvidia is still being worked on.
Download
http://folding.stanford.edu/English/DownloadWinOther
Install
http://www.xtremesystems.org/forums/...3&postcount=17
el cheepo 3850 gets me 1100+ ppd
Yea you can but not with Vista atm
Unless they fixed it it's 1 gpu client per "Vista" machine.
Also Cat.8.5 DOES NOT work for folding, use 8.3 or 8.4
I've also found that when 2 or 3 cards are folding on 1 machine the total
ppd is a lot less then using 3 cards on 3 machines.Mind you that was
on the MSIK9a2 and 9600be @ 2500mhz. Higher cpu freq. should bring higher ppd
Maybe I should make a gpu2 thread? got lots of picks
so the GPU client is CPU limited? would have thought the gpu client taxed the cpu just a little unless its doing some of the folding with the cpu core also.
Perhaps the GPU takes the data and crunches it, then the CPU processes/compresses it? Dunno, just guessing :shrug:
The old X1900 series client wasn't GPU limited too much, I assumed this to be the same here too but I guess I'm wrong.
I just found my answer in a FAQ, The Gpu is capable of so much work but the data is fed to it by the cpu wich realy cant keep the gpu fed as fast as the gpu can work.
"Dedicating 100% of a CPU core to a GPU client is strongly recommended. You will need one CPU core per GPU core to keep the GPU client adequately fed with data.
Folding performance on high-end GPUs is directly related to the performance of the CPU feeding the new client. Increasing the GPU core speed may not improve folding performance without sufficient CPU speed to support those higher GPU speeds. The reverse scenario may also be true. A very fast CPU helps to get the best GPU performance with this new client."
Ahh OK that makes sense then.
It's true however a lot depends on the given project:
For example some projects will load up the gpu to 100 % with
the cpu clock being low (2.4). Other projects need 3.0ghz in order to
load the gpu 100%
So if your gpu is 80& load then increasing your cpu speed will take the load
higher and the ppd along with it. Recent projects however (4727) don't need
much from the cpu and increasing cpu speed by 400-500 Mhz brings no
ppd increase. Hope this makes some sense;)
Ok, I'm in......running a 512mb 8800GT gpu for F@H. A Q9450 feeding that and WCG....with Dimes running around somewhere! :ROTF:
I folded for a while a few yrs ago, but I cannot recall my username, so I'm just starting from scratch= naja002
I have a few questions, but I'll just start a new thread(s).....:up:
Thanx for the Welcome! :up: I think I've already answered all of my questions, but.......that's not likely......:ROTF:
Everything is running fine atm.....:up:
AHHHHH!!! Watch out!! Snake!!!
:welcome:
Ok just threw a 4870 on F@H. Does this iter /sec number say anything? I think it looks low compared to some numbers I saw around the net. I'm seeing 700 to 1700 iter /sec jumping up and down when I open the viewer. Clocks are at 790 / 1000Mhz. Load is 100% on the GPU along with 100% WCG on the CPU.
The 4870 is one hot card, I'm afraid my poor old PA120.3 cant take it anymore. The air coming out of the rad is very very warm, even without yesterdays 30C ambient here in the crunching room. And that's with 6 Yate Loons at 12v. Temps on the GPU are hitting 45C and the CPU is exceeding 60.
Also, this fahmon application isnt doing anything for me. Keeps saying work is at 0%?I think I might have confused the program when I accidentally installed the CPU F@H client first, and then installed the GPU client in the same folder?
Isn't the gpu client a standalone app?
Try moving it to it's own folder ;)
Hehe, easier said than done no? I cant see what is what, and when i set fahmon to monitor the folder named folding@home-gpu in the F@H directory it doenst recognise anything - so moving that wont work. And I'm afraid I will lose all progress when I just reinstall to a different directory.
In addition to the above little problems, it seems that Vista isnt able to feed the GPU client enough data without one core assigned to it. This pisses me off more than high temps or whatnot. Wasnt this an XP only problem?
Been a long time since I ran it (with a 1900xt) but I just kept the folder on desktop.
Could you not download another copy just to see which files the folder needs, then put them together in new folder?
Sorry, I am so out of touch with folding....I'll stop trying to be helpful now. :)
You must point fahmon to a folder (it's a "hidden" btw):
XP: C:\Documents and Settings\<your_windows_username>\Application Data\Folding@home-gpu\
Vista: C:\Users\<your_windows_username>\AppData\Roaming\F olding@home-gpu\
Have you tried set the gpu-client to "slightly higher" tab on and if it has any difference in performance?
EDIT: You find this via systray and setup (or something like that IIRC, at work right now). Might have to restart the client/windows to take it account.
You are the man sc00p :up:
Agreed... a bit of mess again, UnG was on the track for a while :shakes: What can you do... we could use dedicated and hired mod :p:
Welcome back from rehab MtM :rofl:
You're free to pm me for example if you need to ventilate sometimes so it wouldn't lead to so drastic consequences. I plan to stick around here again now that I have had my own holiday used this summer.
Yeah I miss him to. I am so freakingly preoccupied with personal issues ( severe one's... ) that plays a role, but also I am not the kind of person suited to do what he did. If I consolidated knowledge like that, even when I gave references I would get burried under accusations of stealing other people's intelectual property :ROFL:
Atleast it happend in the past, and with me being kind of stuck thinking in circels it make me pretty sure it will happen again. I seen all the info concerning the client posted, here on xs and other forums and I know Mike seen it to and commented on the disabillity currently showing in the fah community to consolidate everything. I think two things are happening, to many teams want to keep the juicy stuff for themselfs and I think allot of people think that with the rate Pande Group is updating clients it's almost no use as it will be obsolete the moment it is posted.
That does not mean there are no good public guides out there, you just need to know where to look :shrug:
IIRC, in XP both the nvidia and ATI clients need a whole core. In vista, nvidia just uses a couple % while ATI still uses a whole core. I think that's what I've heard.
What we need is a good setup guide written up in the thread linked in my sig, Then Jimwah to edit the first post with the direct link to it. I think that would be helpful. Then a MOD to sticky that thread as well!
Wow... do I see 3 consequitive posts about the same subject yet no one has done so yet. I'm willing to contribute if something is asked but I'm pretty sure one of the reasons I been a lil out off things last week ( besides my vacation offcourse ) is that my medication has been alterd ( again, and without my consensis or approval ) and I really feel incapable of doing more then formulating a semi english looking post. Which is pretty rough because I just bought my own domain and need to put content on that asap ( personal, related to the stuff in my signature amongst others ).
Story of my life maybe, always have this stuff happen to me at moments I whished I could have been more myself.
Sparky don't you feel up to it? I know you got the enthousiasm, and if you don't know anything sc00p, mike me and everyone else would help where needed. Just start the thread and I'm sure everyone will pinch in and proof read/correct. If it's all done to satisfaction, we get one off our long lost mods to edit it in? I'm not trying to put you in a position where you would feel saying no would reflect badly on yourself, I just thought about it and then realized you yourself must have thought about it and just needed a lil incentive/push :up:
15% usage on ATI in vista? Not bad, apparently it doesn't need a whole core then. Cool beans.
Yeah newer cores on the Red side made cpu dependancy a lil better almost comparable to the Green team.
How much of a hit did you take on wcg? I mean we can't directly compare points/scientific value but I would like to know the drop.
I hear quads do nice in wcg, if give you my quad ppd in wvg ( eg change client on it from fah ) if you give me equivalent ppd back the quad does in wcg in gpu2 clients for fah, deal?? :D
Edit:
mHouston isn't sitting on his @ss all day long ( or... who stands when coding :rofl: ).
Their making big improvements to both clients all the time :up: I'm missing about 2 new cores if I don't log into the forums daily lol.
Tbh I don't exactly know how much ppd a q6600@3g does in wcg either, I just very much like the idea. I think the advancements made through the gpu clients are more important scientificly then the advantage of my quad on the project, so I'll happilly trade my quad to another folding project if it means getting more gpu's onboard this one :up:
Gpu folding is the way to go, I think wcg and any other major dc project will go ggpu eventually ( anyone heard anything about this?? ) but atm it's just this project and it's one I like very much because of the open results it obtains ( which could then also benefit other folding related dc projects ).
If you want to take me up on the offer shoot me a pm we'll talk about it in private, might be a nice incentive to others if we can work something out :wtf:
All in the name of science :fact:
I'm very much liking this idea. It's an ideal solution for both teams.
Lets try it out for a week or so and see if my card can hold the load without dying. :D:eek:
People folded with x19xx cards untill Pande Group discoverd DirectX wasn't producing scientific results, think Ati has more off a proven track record here then Nvidia :up:
Wait wait wait, i need something cleared up now.....i was just going to switch to Vista because i heard it doesn't need 1 CPU core, how much does the Nvidia client use in XP?
I'm getting an Nvidia card real soon and i was getting ready to sell my X-Fi because of the nightmares i had in the past with Vista.
Plus i heard the Realtek on my board is pretty good:shrug: Also i couldn't flash the firmware on my Creative MP3 player, which i need to do in order for Vista to "see" it and charge it :down:
I'm ready to sell that too.....Down with Creative!!:shakes:
Kingcarcas... dont' know if you can still find them but do a search for daniel_k's drivers for vista. They work great. Unfortunately I only have the Audigy 2 ZS drivers backed up. Maybe someone here has the ones for the X-Fi stashed away. I've not been too impressed with realtek audio, it was relatively crappy on my previous board compared to my Audigy 2 ZS - background noise and just not as clear.
In XP I believe that either GPU client requires a whole CPU core. Just the way XP handles CPU processing.
Thanks i'll try to look for them.
Afaik that's only semi correct, both will report the same load but I can recall one side even while reporting the same usage actually doing very little ( less power draw then a fully pushed core ). Had to do with the polling, so I'm guessing Nvidia is the one leading with this aspect again.
If my memory is correct, that also ment under xp that folding smp and gpu is possible but I haven't treid that since... ever? What I am trying to say is, you should be capable of putting that core which shows 100% on nvidia polling to do other stuff without chocking the client.
Someone on xp proof me wrong :D
Mikeb12, just noticed you started folding for Xtreme about a week ago. Welcome and way to make an entrance!
I think the welcome is late, but you're surely right about the entrance ;) howly cows!
thanks for the welcome.... I've been around posting for a while, and been a folder for about a year.... but just started adding units to xs's roster...
I haven't been on the forums much yesterday and today... had a fever yesterday and stayed in bed alot, feel a little better today and no fever, just still not much appetite or energy... must have been a bug or something... I'm sure it''ll pass and i'll get back to reading everything by the weekend... there was a lot of activity in this forum yesterday, I still haven't caught up with all the posts, yall were busy...
Somehow I was thinking Mike was in the team for longer... seems that way in my brain! :doh: Sorry if I didn't give ya a proper welcome mike!
Same here Sparky he was on the forum helping people out for a lot longer :up:
Think he still deserves a proper welcoming thread?
road-runner
http://www.scootplace.net/smiles/original/129.gif
just put the hammer down.... http://folding.extremeoverclocking.c...hp?s=&u=318494
34K ppd yesterday....
:eek2:
Roadrunner music
:wtf:
Hello! I just joined in folding for our team. It's time to use this computer for a good cause!
I'm using smp, and using the rig in my sig below.
It's only using my Q6600 at 25% is that normal?
and another thing, how do I get that sig that shows my status of my F@H accomplishments?
Thanks for your input!
XSbb (F@H n00b)
Hey there :welcome: to the team xsbb :wave:
No, that's not normal. Wel it is, but only if you didn't follow sc00ps guide I think ;)
I bet you forgot to add -smp to Advanced options, additional client paramaters ( though in the guide he only covers adding it to the shortcut, I'll post this as addition as it's easier ) :up:
If you close the client now, create a new shortcut to the executable with -configonly and run it ( it's handy for the future to keep that link somewhere ) and change the option, restart the client and let it finish it's current wu, the next wu will be smp if everything goes right. If it doesn't do a fresh install with the right switches from the start and it should be ok. It's still prety bad they expect you to not delete wu's when they make it harder for the high performance client to get the most performance by default from their installations lol :shrug:
Gah, I tried reinstalling and trying it again. I think im messing up something from the guide.
It's a little frustrating because I want to help out, but if you could make the changes to the guide so that I can fully utilize my Quad to fold that would be great!
I'm a noob at this, but an easier guide could help! :)
Thanks for helping out!
I can't change other people's posts, I posted the additional information though :up:
But I can help you if you tell me where you go wrong. If you installed again and added the -smp option when configuring the advanced options, additional client parameters it should have worked. But there are some setting, like selecting a to small memory size ( which is also a bug, it can on occasion pick up less ram then you actually have ) or not answering Big when asked about the wu size. So, if you could post your fahlog.txt in [ code ] tags, and if that's not enough also a screenshot off you running through the configuration options ( as sc00p did the guide ). Those could help allot trying to figure out where it goes wrong.
:welcome:Full speed ahead! :CTF::cheer: