• 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.

Newbie needs help with it 6274

T

Thomas R

Weaksauce
Joined
Aug 31, 2012
Messages
110
Hi guy´s

First - sorry for my bad english. I´m from germany and my school-english is´nt the best!
Furthermore i´m a noob in Folding. Only two month´s i´m on board.
Next - i´m a dummy in Linux. I hate it - terminal is not my friend.
And I need your help.

My system:
I have this racket (http://www.supermicro.com/Aplus/system/1U/1042/AS-1042G-TF.cfm) within a H8QGi+-F.
4 x Optereon 6274 Tray.
And 16 x Samsung 4 GB reg. ECC DDR3 DIMM SDRAM M393B5270CH0-CH9 (4096 MB, reg.ECC DDR3 1333, 1 Modul, Dual Rank, x4 )

In this thread: http://hardforum.com/showthread.php?t=1708870 I read, that it is possible to get "turbo boost" with TPC and psmax 1.

Found this How-To an have done it like you can read it there: http://hardforum.com/showthread.php?t=1669698
I was trying Version 0.43 from TPC. With option "tpc -l" i´ve got the answer, that there are no CPU ID´s to read (I don´t remember the real Output - but you understand what I mean)
Removing Version 0.43 and trying V 0.422. Same Problem.
Removing an trying V 0.41 - same Problem.

I don´t know where is my error!
Is it impossible to work with TPC on my Mainboard?
Is it possible at all to get "turbo-boost" with the mix from my mainboard an the installed CPU´s?
Or can´t do one of it the turbo-boost.

If you give me an answer, please don´t use too much specific words. I know this ist hard for you - but it is really harder for me if you use it! ;-)
 
Before running TPC make sure you have package cpuid and msr installed and loaded. If you are unsure how to do this just copy and run these commands (assumes you are running Ubuntu):

sudo apt-get install cpuid msr
sudo modprobe cpuid
sudo modprobe msr

Then try running TPC again (make sure to use 0.43 older versions won't work with your CPU's)
 
Thank´s for your answer.
I´ve tried to do like you wrote.
Here is what I´ve got back:
tpc01.jpg

Only a part of it.
 
You need to run TPC w/ sudo in front of it - that's likely what is causing all the permission denied messages.
 
OK - look´s better!
Can´t give you the hole output.

tpc02.jpg



Should I try it now with "sudo tpc -psmax 1" ?



EDIT:
Now I had a try and the "Maximum pstate" changes from 6 to 1.
Is this OK?
 
Last edited:
Yep sudo tpc -psmax 1 is the right command so you should be all set.
 
Thank you very, very much tkam.

I´ve changed to "-psmax 1" 45 minutes ago and got positiv results.
I don´t know how high the frequency is running now (think 2500), but the TPF at the actual running 8101 is going down from round about 15 minutes to 13 minutes and 6 seconds at the moment.
This is fantastic.
Once more - thank you very much!
 
Glad to have ya folding [H]ard!

If we didn't have new people asking questions about how to get hardware up and running it would be rather boring place. This is one of the few "safe" places on the internet where noob isn't a bad thing. ;)
 
Welcome, glad you got it sorted. If you need to post any more logs try paste bin with a link, or put the output in a quote within your post.
 
Congrats!
However, 13min TPF for 4p 6274@2.5GHz is still a little slow in my opinion. ~12m10s should be possible for it. Have you installed thekraken and is DLB engaged?

Thomas R said:
Thank you very, very much tkam.

I´ve changed to "-psmax 1" 45 minutes ago and got positiv results.
I don´t know how high the frequency is running now (think 2500), but the TPF at the actual running 8101 is going down from round about 15 minutes to 13 minutes and 6 seconds at the moment.
This is fantastic.
Once more - thank you very much!
Click to expand...
 
Thank you for your answers.

thekraken is installed and running. I can read a message in the terminal when I start ./fah6.
So I think it is running.

Because DLB:
Bevor the first frame is in work I can read the message:
"Note: Enable Dynamic Load Balance"
Is it running than or not?
 
That looks like it is working properly.
 
The prompt of "NOTE: Turning on dynamic load balancing" is the proof of DLB has been enabled.
 
OK
In this thread I read that PowerNow should be disabled:
http://hardforum.com/showthread.php?t=1669698
Must I disable PowerNow?
Can anyone tell me how I can write in the terminal this sign: [ and ] ??
What kind of combination I have to use on my keyboard?


And - is it wright that "psmax 0" is not possible?
Or better - is it possible to get out the maximum Frequency of 3100MHz from my 6274 Opteron's?
 
I don't think disabling PowerNow would increase performance on Opteron 6200s (especially when "-psmax 1" is on).
As for "-psmax 0", it would be great if it works as expected. However, I had tested it before but seems that would just make the system run into a ultra-slow state.

EDIT: -psmax 0 is generally not a good idea since its default vcore is too high and would cause damage to the board if it really works.
 
Last edited:
Thank you!

What about this?

[13:07:52] Completed 250000 out of 250000 steps (100%)
[13:08:07] DynamicWrapper: Finished Work Unit: sleep=10000
[13:08:17]
[13:08:17] Finished Work Unit:
[13:08:17] - Reading up to 64340496 from "work/wudata_06.trr": Read 64340496
[13:08:17] trr file hash check passed.
[13:08:17] - Reading up to 31556960 from "work/wudata_06.xtc": Read 31556960
[13:08:18] xtc file hash check passed.
[13:08:18] edr file hash check passed.
[13:08:18] logfile size: 192732
[13:08:18] Leaving Run
[13:08:22] - Writing 96251064 bytes of core data to disk...
[13:08:50] Done: 96250552 -> 91494336 (compressed to 5.8 percent)
[13:08:50] ... Done.
[13:09:51] - Shutting down core
[13:09:51]
[13:09:51] Folding@home Core Shutdown: FINISHED_UNIT
[13:10:00] CoreStatus = 64 (100)
[13:10:00] Unit 6 finished with 76 percent of time to deadline remaining.
[13:10:00] Updated performance fraction: 0.758737
[13:10:00] Sending work to server
[13:10:00] Project: 8101 (Run 23, Clone 1, Gen 42)


[13:10:00] + Attempting to send results [September 6 13:10:00 UTC]
[13:10:00] - Reading file work/wuresults_06.dat from core
[13:10:00] (Read 91494848 bytes from disk)
[13:10:00] Connecting to http://128.143.231.201:8080/
[13:10:00] - Couldn't send HTTP request to server
[13:10:00] + Could not connect to Work Server (results)
[13:10:00] (128.143.231.201:8080)
[13:10:00] + Retrying using alternative port
[13:10:00] Connecting to http://128.143.231.201:80/
[13:10:00] - Couldn't send HTTP request to server
[13:10:00] + Could not connect to Work Server (results)
[13:10:00] (128.143.231.201:80)
[13:10:00] - Error: Could not transmit unit 06 (completed September 6) to work server.
[13:10:00] - 1 failed uploads of this unit.
[13:10:00] Keeping unit 06 in queue.
[13:10:00] Trying to send all finished work units
[13:10:00] Project: 8101 (Run 23, Clone 1, Gen 42)


[13:10:00] + Attempting to send results [September 6 13:10:00 UTC]
[13:10:00] - Reading file work/wuresults_06.dat from core
[13:10:00] (Read 91494848 bytes from disk)
[13:10:00] Connecting to http://128.143.231.201:8080/
[13:10:00] - Couldn't send HTTP request to server
[13:10:00] + Could not connect to Work Server (results)
[13:10:00] (128.143.231.201:8080)
[13:10:00] + Retrying using alternative port
[13:10:00] Connecting to http://128.143.231.201:80/
[13:10:00] - Couldn't send HTTP request to server
[13:10:00] + Could not connect to Work Server (results)
[13:10:00] (128.143.231.201:80)
[13:10:00] - Error: Could not transmit unit 06 (completed September 6) to work server.
[13:10:00] - 2 failed uploads of this unit.
[13:10:00] Keeping unit 06 in queue.
[13:10:00] + Sent 0 of 1 completed units to the server
[13:10:00] - Preparing to get new work unit...
[13:10:00] Cleaning up work directory
[13:10:00] + Attempting to get work packet
[13:10:00] Passkey found
[13:10:00] - Will indicate memory of 64410 MB
[13:10:00] - Connecting to assignment server
[13:10:00] Connecting to http://assign.stanford.edu:8080/
[13:10:02] Posted data.
[13:10:02] Initial: 8F80; - Successful: assigned to (128.143.231.201).
[13:10:02] + News From Folding@Home: Welcome to Folding@Home
[13:10:02] Loaded queue successfully.
[13:10:02] Sent data
[13:10:02] Connecting to http://128.143.231.201:8080/
[13:10:15] Posted data.
[13:10:15] Initial: 0000; - Receiving payload (expected size: 30306571)
[13:12:36] - Downloaded at ~209 kB/s
[13:12:36] - Averaged speed for that direction ~225 kB/s
[13:12:36] + Received work.
[13:12:36] Trying to send all finished work units
[13:12:36] Project: 8101 (Run 23, Clone 1, Gen 42)


[13:12:36] + Attempting to send results [September 6 13:12:36 UTC]
[13:12:36] - Reading file work/wuresults_06.dat from core
[13:12:36] (Read 91494848 bytes from disk)
[13:12:36] Connecting to http://128.143.231.201:8080/
[13:12:36] - Couldn't send HTTP request to server
[13:12:36] + Could not connect to Work Server (results)
[13:12:36] (128.143.231.201:8080)
[13:12:36] + Retrying using alternative port
[13:12:36] Connecting to http://128.143.231.201:80/
[13:12:36] - Couldn't send HTTP request to server
[13:12:36] + Could not connect to Work Server (results)
[13:12:36] (128.143.231.201:80)
[13:12:36] - Error: Could not transmit unit 06 (completed September 6) to work server.
[13:12:36] - 3 failed uploads of this unit.
[13:12:36] Keeping unit 06 in queue.
[13:12:36] + Sent 0 of 1 completed units to the server
[13:12:36] + Closed connections
[13:12:36]
Click to expand...
 
Thomas R said:
Thank you!

What about this?
Click to expand...

That's most likely Langouste in action. There's a post (by musky) that explains
this in more detail (as does the README:) though I can't seem to find it.

If work from the machine in question is getting credited correctly, don't worry about it.

quickz said:
I don't think disabling PowerNow would increase performance on Opteron 6200s (especially when "-psmax 1" is on).
Click to expand...
+1


On retail 6200s we want PowerNow and TurboBoost/CPB on.
On other G34 chips (MC/IL ES, retail MC) we want them off.
 
OK - than I´m fine! ;)

The TPF in the new WU is at 12:40 now.
That´s great and I thank you once more for your help!
 
A correction for "-psmax 0":
Just now thinklet told me that "-psmax 0" did work for his ES 6200s.
It's a good news. However, we still have no way to let it work for retail 6200s.

EDIT: -psmax 0 is generally not a good idea since its default vcore is too high and might cause damage to the board if it really works.
 
Last edited:
3.1 is most likely some-cores turbo. So... AMD marketing information is not false per se,
just of little to no use :rolleyes:

6200 series have two turbo bins: some-cores turbo (pb0 in tpc -l) and all-cores turbo (pb1 in tpc -l).

CPU enters some-cores turbo (pb0) state when certain number of compute units are in C6 (deep
sleep) state (something you won't observe when FAH is running). In other words -- you
can forget about pb0 (3.1 GHz).

Ok, so we settle with all-cores turbo (pb1).
Now, problem is that even all-cores turbo doesn't work reliably with these CPUs [sic!].

Purpose of -psmax 1 is to keep your CPUs in all-cores turbo state at all times (pb1).
Note that this is something that CPU designers most likely didn't intend to permit (CPU
bug), we're just taking advantage of it (thanks to quickz who found it).
 
@tear
This is real a loveley description of the facts! ;)

@402blownstroker
I guess, this is the same what ´tear" want to tell me. :D

Sorry tear.......
 
Thomas, you are new to Folding and are a Linux dummy, yet you jumped right into the AMD 4P Opteron pond. You sir, are brave and ambitious. Salute! :)
 
@Linden

Since TSC ended, I never was crunching anything.
Now I´m back and want to do my best.

And I hope my machine is running the next years without any problems or restart, because when I must start Langouste, TPC and FAH again, I must search and read all the turorials again. :rolleyes:
Linux is not for me and I think Linux is thinking the same about me! ;)
 
My rig is dead !!
Yesterday in the afternoon, my rig was going down.
Today I was trying to start it again, but there is no reaction to the start button.
I can hear and see that the fan´s from the psu are running. Slowly - but they are running.
I disconnected the power cord and tried it again after a few minutes.
Same procedure!
PSU fan´s are running slowly but no reaction to the start button.

Any suggestions?
 
Do any of the heatsink fans spin? Have you tried resetting the BIOS? Any more details on how the machine went down?
 
No spinning heatsinks.
No BIOS-reset, because I think if there is no reaction to the start button, it is no problem with the BIOS. ?!

I can´t tell more about how the machine went down because it ist standing in another building and nobody comes in this room.
 
I would suggest you to check the 24-pin and 8-pin connectors on the motherboard, any damage there?
 
No damage at the 24-pin and 8-pin connectors!

I made a BIOS-reset now - no reaction.

The green light at the PSU is burning. And the control lamp for the network connection is burning too.

:(
 
Last edited:
No - it starts slow and remains slow.

Now I know more!
I´ve disconnected the PSU and have connected a different (only 450W).
The machine starts without a problem.
After the phone call with a technician told me this, that I would have to remove the plug-in module for the PSU also.
Just so he could check if it's the PSU or on the module, without having to send in the whole machine.

Shit!
 
PSU is back now.
And I have a new problem.

Machine is running, but I have noticed that it responds slow.
When I will only start Firefox, it takes a half minute till the window appears.

I´ve started F@H and got a new work unit.
I proved the cpu-usage with htop. It´s between 94% and 99% on all threads.
But F@H is standing now about an half hour.
big.jpg
 
I have big problems!!!

Now I have made new installations from Ubuntu (three times) because I don´t get FAH tu run.
Don´t ask why - I don´t know!

I have done it once more. Got FAH, thekraken, langouste and tpc.
I can start FAH, but I get p6978.
Stopped FAH, deleted the work-file and start FAH again.
Same preciture again, again and again. Every time I get p6978.
Stopped FAH (close Terminal) and made a pause.
Come back, start htop to see what is running and see this:
thekraken.jpg

There are 8 Tasks wiht FahCore_a3.exe an about 20 Tasks with thekraken -FahCo.

What the hell have I done??????
 
Those are the extra processes that the kraken kicks off to engauge DLB. They should go away after a little bit.
 
You must log in or register to reply here.
Back
Top