• 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.

115k Short - damn

T

TheMTZ

Limp Gawd
Joined
Jan 17, 2011
Messages
189
Missed my goal to score 1 million in a day by 115.539 points. If the current 6901 would be ~3 hrs faster i would have made it. It's kind of annoying, but on the other side i am also happy with 884.461 points in one day - it's my best per-day score so far and i think i mowed a couple of people without them noticing me :cool:
 
overclock MOAR!!!!!!! 100mhz more would of gotten you that much closer to getting a 1 million point update!
 
Wow, the old timers are looking slow compared to you new whippersnappers.
 
Zink said:
Wow, the old timers are looking slow compared to you new whippersnappers.
Click to expand...


nein, we just know our place.. when our rigs continue to just hum along without any issues 24/7 we just sit back and watch as all these high PPD people lose half their points because 1 rig went down.. :p
 
sirmonkey1985 said:
nein, we just know our place.. when our rigs continue to just hum along without any issues 24/7 we just sit back and watch as all these high PPD people lose half their points because 1 rig went down.. :p
Click to expand...

On a side note: All my clients are running base clock. It's one dual Xeon system on BigAdv and the rest are just hundreds of Core2Dou Boxen.
 
Wow...how many folks do we have now with a legit shot a 1 million points in a day? I am not really one of them...the stars would have to line up perfectly for me to do it right now. very impressive MTZ!
 
Did yoy ever calculate your points per watt?
Nice bit of mowage this weekend.
 
TheMTZ said:
On a side note: All my clients are running base clock. It's one dual Xeon system on BigAdv and the rest are just hundreds of Core2Dou Boxen.
Click to expand...


aww i don't get to watch an overclock system crash and burn.. no fair damnit!!!
 
Isn't that like $30k a year greater than idle for the juice for those machines? It's more than $50 a day anyways. I guess it's still winter so spread over enough buildings it might not cost much.
 
Zink said:
Isn't that like $30k a year greater than idle for the juice for those machines? It's more than $50 a day anyways. I guess it's still winter so spread over enough buildings it might not cost much.
Click to expand...

where are they located...

when you use lots of juice... you typically don't pay consumer rates...
 
Awesome output! It explains why we achieved such a high production yesterday. Two people close or above the 1M mark made a huge difference. :cool:
 
That is awesome! How did spike your output up for just one day? I see you averaged maybe 5-10 wu's an update, but shot out 206 on one, and 37 on the next. Did you borg a bunch of computers for one day?
 
Patriot said:
when you use lots of juice... you typically don't pay consumer rates...
Click to expand...

Correct, we got a special deal for power, water, etc. Last time i checked we consume about 70.000 to 90.000 kilowatt hours a month. So a few hundred more or less fah clients won't make a real impact on the cost's.
 
Wyxe said:
That is awesome! How did spike your output up for just one day? I see you averaged maybe 5-10 wu's an update, but shot out 206 on one, and 37 on the next. Did you borg a bunch of computers for one day?
Click to expand...

I borg'ed 287 C2D 2.3 GHz for one weekend. For me it's really easy to get clients assembled into the folding horde because everything is set up to boot via PXE from a NFS share straight into a custom build Linux environment with the fah client. So what i did was i checked out what PC's are not used over the weekend and booted them up over wake-on-lan.

If people are interested, i can do a little write-up about this madness which took me 4 week's to get it working and is still far away from finished. There are still some problems and limitations ... that's why i only usually run 200~300 clients on the weekend (if i get a chance at all).
 
TheMTZ said:
Yes, however the amount of usable clients will vary. Also i have to remember that i need to boot everything up on Friday's :)
Click to expand...
That's great news if everything can be managed that easily. We will get a nice boost every weekend to counter big E's beginning of the month mega boost. :cool:
 
interested in the pxe image you used...
I might have trouble with nic drivers... but I guess I could always break open the image and throw in drivers...
 
Patriot said:
interested in the pxe image you used...
I might have trouble with nic drivers... but I guess I could always break open the image and throw in drivers...
Click to expand...

There is no PXE Image for the way i set it up, it's all heavy custom builds. It works like this - in short terms :)

-> Client starts and does the DHCP broadcast from the PXE rom.
-> DHCP server answers with the client IP and additionally with a pointer to a TFTP server where my linux kernel is located. (i used the standard pxelinux.0)
-> Linux kernel gets loaded (4 mb file) which takes about <10 seconds.
-> My Kernel has build in NFS support so it mounts its root directory from a NFS server to / and continues to boot from there. The NFS server only offers a RO export of the root file system, so /etc, /tmp and so on gets moved into a ram disk directly after the kernel was done loading (via linuxrc) and then the system continues to run the init scripts.
 
that would be how a basic pxe works...

we use dhcp server that also does the tftp... then have a seperate nfs server...

so do you pack everything in the initrd?
 
Nope, not even a initrd. I am using a kernel with everything it needs compiled in - no modules. This is my pxelinux.cfg/default:

Code: 
DEFAULT diskless64
TIMEOUT 3
PROMPT 1

LABEL diskless64
    KERNEL diskless64-3
    APPEND ip=dhcp root=/dev/nfs nfsroot=100.0.1.194:/mnt/diskless/diskless64 init=/linuxrc

/diskless64 contains a Gentoo installation (which is used in read only mode) as base system.

Edit: The part in my DHCP config

Code: 
... range specs ...
        next-server 100.0.1.194;
        filename "/pxelinux.0";
 
Last edited:
And on the DHCP server you need to configure option 066 (bootfile server) & 067 (bootfile name). Then you should be good to go.
 
You must log in or register to reply here.
Back
Top