New release of diskless folding suite

[bollix47]

@ notfred .... thanks very much ... appreciate all the work that you do.

@Flying Fox .... my thoughts re creating a text file with the current client names would not involve having to know future names. The idea was that if a new client came out the user could change the name in the text file and notfred's kernel would read that file and re-create the links accordingly before downloading the core. It was just a thought ......

[Flying Fox]

@Flying Fox .... my thoughts re creating a text file with the current client names would not involve having to know future names. The idea was that if a new client came out the user could change the name in the text file and notfred's kernel would read that file and re-create the links accordingly before downloading the core. It was just a thought ......

You will still have to put that text file onto the disc image, meaning as a user you will have to regrab the image anyways.

Read it from TFTP? That's too much dependency.

Prompt for the file name? Then it's not as "lights out" as the original idea.

So what are you left with? A "tracker" site, with all its issues and maintenance headaches (someone has to go update to the new file names).

In the end I agree with notfred, we may have to just bear with it until the software comes out of beta.

[notfred]

New release is out with updated URL for the client download.

If you really want, you can always edit the source yourself and rebuild. Alternatively the download location is in the init file in the initrd, you could always extract it out, edit it and repackage it. Putting it as a parameter on the startup line isn't worth it at this point, it doesn't change often enough and I can normally get a release out pretty quickly.

[bollix47]

Thanks a lot for your quick attention to this.

Back to folding.

[Flying Fox]

Is there any plans to incorporate SMP WUs in the benchmark? Those multicore CPUs seem to be screaming for that.

[notfred]

That's on the really long term list, I think we would need Stanford to get the Linux SMP client out of Beta first otherwise the results could change every 3 months when they change the client.

[acruxksa]

Greeting NotFred,

I'm a bit hesitant to have my first post on the forum be a trouble shooting question, especially since you have already done so much to help folding@home, but here it is anyway. I was successfully using your usb key folding image for about a week when the latest update from stanford changed the url. No worries, I just checked back here regularly and you wasted no time getting the changes out. This is where the problem is now though, I'm receiving "Hard drive image is corrupted, not backing up" error on all four of my folding machines. It doesn't seem to affect folding since a all machines are humming along nicely and a couple have successfully completed and returned work, but I'm wondering if you could shed some insight into the situation.

Thanks for your hard work!!!

[notfred]

Hmm, you should only see that error message if you are running under VMWare. It's because I was in the middle of getting the VMWare folding appliance out when Stanford changed the URL. I wouldn't worry about it for now.

[acruxksa]

OK,
No worries then. Thanks for your quick reply.

[bollix47]

FYI

Just a confirmation that I too am getting the same message with the latest version and if I power off and on the work starts at step 0 again. Of course I won't be powering off again until the WU is finished and uploaded.

[Karamiekos]

TTT
First time post here, so first off a big thank you to Notfred and anyone else who has contributed. This is a great program!

Because of this new update I too get the Warning about not saving. Is there a fix to this? I like to start up my pc sometimes for some quick work in windows, maybe a game or two, and then go back to folding, but with the backups not working I am afraid it has crippled my folding, because I lose all that precious work unit progress.

I have breifly tried to edit the old image I have, the initrd in particular, to just change the web address and see if that works, but I am not savy in that area. Please help.

[jeffry55]

I have a curiouity question on folding in Linux. Can you run the GPU client in Linux using Wine?

[Optimum1a]

I'm also seeing the "not backing up message", but I'm able to do a backup tar from the web interface. What's the procedure for restoring from that tar? Can we just untar it to the USB stick and start again?

Thanks

[Flying Fox]

I'm also seeing the "not backing up message", but I'm able to do a backup tar from the web interface. What's the procedure for restoring from that tar? Can we just untar it to the USB stick and start again?

I would think so, but that was one reason why I set up TFTP in the first place instead of untaring it to a USB disk. I just have to manage the files on the TFTP server and it will deal with all that stuff.

[theMASS]

I'm also seeing the "not backing up message", but I'm able to do a backup tar from the web interface. What's the procedure for restoring from that tar? Can we just untar it to the USB stick and start again?

I would think so, but that was one reason why I set up TFTP in the first place instead of untaring it to a USB disk. I just have to manage the files on the TFTP server and it will deal with all that stuff.

I get the "Hard drive image is corrupted, not backing up" message and backups to USB sticks still work fine. As notfred mentioned, the message is due to unfinished code for the VMWare folding appliance, and it's really just a cosmetic issue.

You can untar from a network backup to a USB stick ../folding/1/ or/and /2/ if 2xSMP.

[theMASS]

A member on the F@H forum posted the following issue:

I had this running on two machines with the .iso and fold64.vbx in the VMPlayer folder. On one machine all was well and on the second all was running fine except that for some reason it seemed to be reading the system clock about 3.5x faster than reality. For instance, it entered a 15-minute checkpoint about every 4 minutes. At first I just thought it was FahMon orienting itself until some time went by and it was still happening, yet the other machine was reporting correctly.

At just under a day of this and on Frame 71 of a 2653 it reported itself as being past the deadline (due to the steroids clock) and deleted the WU and tried to get another.

I'll bet you have an AMD cpu. There a CPU driver that fixes the crazy clock.

...you're absolutely right. Surprisingly, there's nothing on the AMD site other than the standard clock utilities, nothing in the forums that is showing up in a search. Is there anyone that's familiar with this that can point me? I would have thought a driver would probably be on the AMD site. Funny it's affecting the X2 4200+ and not the X2 4000+...

Anyone have any advice?

He has tried running with VMWare Server, updated BIOS, and disabled Powernow and CoolnQuiet...

[Karamiekos]

Cool beans. Over the night one computer finished a work unit. It was at 6 per cent so I figured I would try a reboot. I confirm the backup system still works, damn what the screen says. I was just a little leary last night of gambling with 80% of an SMP.

[Flying Fox]

A member on the F@H forum posted the following issue:

I had this running on two machines with the .iso and fold64.vbx in the VMPlayer folder. On one machine all was well and on the second all was running fine except that for some reason it seemed to be reading the system clock about 3.5x faster than reality. For instance, it entered a 15-minute checkpoint about every 4 minutes. At first I just thought it was FahMon orienting itself until some time went by and it was still happening, yet the other machine was reporting correctly.

At just under a day of this and on Frame 71 of a 2653 it reported itself as being past the deadline (due to the steroids clock) and deleted the WU and tried to get another.

I'll bet you have an AMD cpu. There a CPU driver that fixes the crazy clock.

...you're absolutely right. Surprisingly, there's nothing on the AMD site other than the standard clock utilities, nothing in the forums that is showing up in a search. Is there anyone that's familiar with this that can point me? I would have thought a driver would probably be on the AMD site. Funny it's affecting the X2 4200+ and not the X2 4000+...

Anyone have any advice?

He has tried running with VMWare Server, updated BIOS, and disabled Powernow and CoolnQuiet...

The 2nd quote answered it, by "driver" the poster meant the Dual Core Optimizer. They obviously haven't been on TR! I have been talking about this problem and the X2's for almost a year (hint: search for my posts and ("Audacity" or "ping"))!

[theMASS]

A member on the F@H forum posted the following issue:

I had this running on two machines with the .iso and fold64.vbx in the VMPlayer folder. On one machine all was well and on the second all was running fine except that for some reason it seemed to be reading the system clock about 3.5x faster than reality. For instance, it entered a 15-minute checkpoint about every 4 minutes. At first I just thought it was FahMon orienting itself until some time went by and it was still happening, yet the other machine was reporting correctly.

At just under a day of this and on Frame 71 of a 2653 it reported itself as being past the deadline (due to the steroids clock) and deleted the WU and tried to get another.

I'll bet you have an AMD cpu. There a CPU driver that fixes the crazy clock.

...you're absolutely right. Surprisingly, there's nothing on the AMD site other than the standard clock utilities, nothing in the forums that is showing up in a search. Is there anyone that's familiar with this that can point me? I would have thought a driver would probably be on the AMD site. Funny it's affecting the X2 4200+ and not the X2 4000+...

Anyone have any advice?

He has tried running with VMWare Server, updated BIOS, and disabled Powernow and CoolnQuiet...

The 2nd quote answered it, by "driver" the poster meant the Dual Core Optimizer. They obviously haven't been on TR! I have been talking about this problem and the X2's for almost a year (hint: search for my posts and ("Audacity" or "ping"))!

Thanks! This sounds like what he needs. Guess he didn't look too hard on the AMD site

[Flying Fox]

...
He has tried running with VMWare Server, updated BIOS, and disabled Powernow and CoolnQuiet...

The 2nd quote answered it, by "driver" the poster meant the Dual Core Optimizer. They obviously haven't been on TR! I have been talking about this problem and the X2's for almost a year (hint: search for my posts and ("Audacity" or "ping"))!

Thanks! This sounds like what he needs. Guess he didn't look too hard on the AMD site

I'll take payment in the form of that VM running the SMP client for UnitedGerbilNation and team 2630 for at least a month. If he joins permanently then it is even better.

[EllisD]

Hi Everyone,

I registered because i stumbled upon this excellent tool and wished to thank notfred and everyone involved in making this project better.

I have tried the VMware+Linux route, but had nothing but troubles and lost my cool, so went looking for alternatives. Needless to say, this filled the void. I am only really on my PC for about 6 hours a day, and its folding the rest of the time, so this is the perfect solution that offers the best results i have seen. On my main rig, it does about 18mins per % which is leaps and bounds ahead of the winSMP client as well as the VMware route. It almost matches my server that puts out 16.5mins per WU within 64bit ubuntu.

Again, i wish to say Thanks and keep up the stellar work!

[Flying Fox]

I have tried the VMware+Linux route, but had nothing but troubles and lost my cool, so went looking for alternatives. Needless to say, this filled the void. I am only really on my PC for about 6 hours a day, and its folding the rest of the time, so this is the perfect solution that offers the best results i have seen. On my main rig, it does about 18mins per % which is leaps and bounds ahead of the winSMP client as well as the VMware route. It almost matches my server that puts out 16.5mins per WU within 64bit ubuntu.

I run the diskless ISO inside a VM.

[RAH]

I have been getting the "Hard drive image is corrupted, not backing up" message also.
Doesn't seem to be much of a problem, except when the machines get restarted. They
go back to zero. Since I don't shut them down, no big deal, usually. Last night had a
power outage. All went back to zero.

Since they are inexpensive, would putting usb sticks in them stop this?
If so, would I need to set the sticks up with the folding on them before hand? Or just blank formatted ones.

[Flying Fox]

Since they are inexpensive, would putting usb sticks in them stop this?

That or you can set up TFTP backup.

[RAH]

I thought it had that. I'm running off my Windows server, using the tftpd32.
There are tons of backup folders, but if they get restarted, they start from zero.

I just ran the USB sticks before trying this, and they always started from the last check point.
I think I'll just get some more sticks, and go back to them.

[Flying Fox]

I thought it had that. I'm running off my Windows server, using the tftpd32.
There are tons of backup folders, but if they get restarted, they start from zero.

Is your tftpd32 set up properly? With it taking over DHCP and not your router, etc.?

I didn't have tftpd32, but Win2K's built-in one. It works.

[theMASS]

I have been getting the "Hard drive image is corrupted, not backing up" message also.
Doesn't seem to be much of a problem, except when the machines get restarted. They
go back to zero. Since I don't shut them down, no big deal, usually. Last night had a
power outage. All went back to zero.

Since they are inexpensive, would putting usb sticks in them stop this?
If so, would I need to set the sticks up with the folding on them before hand? Or just blank formatted ones.

Make sure they are formatted with a FAT partition. The HP formatting tool is the best tool to use:
http://files.extremeoverclocking.com/file.php?f=197

Then if you have the Install to USB drives on boot: option enabled (default setting) you should be able to boot from them as well as have backups written to them.

You don't have to have anything installed on them to use for backups. Just stick them in and they will start having backups written to them.

They don't have to be blank. You can have other stuff on there... I've even used my BlackBerry attached via USB and it worked fine.

[RAH]

I thought it had that. I'm running off my Windows server, using the tftpd32.
There are tons of backup folders, but if they get restarted, they start from zero.

Is your tftpd32 set up properly? With it taking over DHCP and not your router, etc.?

I didn't have tftpd32, but Win2K's built-in one. It works.

Beats me! I set it up with the how to.
Worked fine at first. But since some of the updates, it doesn't do so well, if the client gets restarted.
Having a few family problems right now. I'll go back through it when I have time. The router is set correctly,
unless it has changed.
I could put together a W2K server, and try it too. I have a copy somewhere.

[bj0ng0]

Hi, first i would like to say thanks to notfred for an easy way to pxe boot tons of clients
but after i rebooted the clients, i can see on the main page on the only dualcore in my farm, "it says "Found 2 processors"" but it only has one instance, with prior version i had two, weird...

[theMASS]

Hi, first i would like to say thanks to notfred for an easy way to pxe boot tons of clients
but after i rebooted the clients, i can see on the main page on the only dualcore in my farm, "it says "Found 2 processors"" but it only has one instance, with prior version i had two, weird...

Sounds like you were running the standard client (x 2), and now notfred's software is running the SMP client as it is detecting the dual core processor. If you were running a REALLY old version this makes sense. If your "prior" version was from the last year... then yes weird =:?