Page 3 of 3
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 5:16 am
by Amaruk
ElectricVehicle wrote:I also suspected that the bigadv to SMP rollover when bigadv aren't available might be controlled by the assignment servers and you just confirmed that. So it looks like there's a little problem in the new assignment servers to handle that rollover.
I'm not sure the problem is with the AS. Right now it appears the status on all the bigadv servers is 'accept' which means the are not assigning work. But the AS thinks
some (but not all) of them are 'green' and is assigning them anyway.
Code: Select all
--- Opening Log file [June 1 01:27:00 UTC]
# Windows SMP Console Edition #################################################
###############################################################################
Folding@Home Client Version 6.34
http://folding.stanford.edu
###############################################################################
###############################################################################
Launch directory: C:\Users\Blitz\SMP634
Executable: C:\Users\Blitz\SMP634\FAH6.34-win32-SMP.exe
Arguments: -smp -verbosity 9 -bigadv
[01:27:00] - Ask before connecting: No
[01:27:00] - User name: Amaruk (Team 50625)
[01:27:00] - User ID: 7DE849C1xxxxxxxx
[01:27:00] - Machine ID: 1
[01:27:00]
[01:27:00] Loaded queue successfully.
[01:27:00] - Preparing to get new work unit...
[01:27:00] - Autosending finished units... [June 1 01:27:00 UTC]
[01:27:00] Cleaning up work directory
[01:27:00] Trying to send all finished work units
[01:27:00] + No unsent completed units remaining.
[01:27:00] - Autosend completed
[01:27:07] + Attempting to get work packet
[01:27:07] Passkey found
[01:27:07] - Will indicate memory of 12279 MB
[01:27:07] - Detect CPU. Vendor: GenuineIntel, Family: 6, Model: 12, Stepping: 2
[01:27:07] - Connecting to assignment server
[01:27:07] Connecting to http://assign.stanford.edu:8080/
[01:27:07] Posted data.
[01:27:07] Initial: ED82; - Successful: assigned to (130.237.232.237).
[01:27:07] + News From Folding@Home: Welcome to Folding@Home
[01:27:08] Loaded queue successfully.
[01:27:08] Sent data
[01:27:08] Connecting to http://130.237.232.237:8080/
[01:27:08] Posted data.
[01:27:08] Initial: 0000; - Error: Bad packet type from server, expected work assignment
[01:27:08] - Attempt #1 to get work failed, and no other work to do.
Waiting before retry.
[01:27:19] + Attempting to get work packet
[01:27:19] Passkey found
[01:27:19] - Will indicate memory of 12279 MB
[01:27:19] - Connecting to assignment server
[01:27:19] Connecting to http://assign.stanford.edu:8080/
[01:27:20] Posted data.
[01:27:20] Initial: ED82; - Successful: assigned to (130.237.232.237).
[01:27:20] + News From Folding@Home: Welcome to Folding@Home
[01:27:20] Loaded queue successfully.
[01:27:20] Sent data
[01:27:20] Connecting to http://130.237.232.237:8080/
[01:27:20] Posted data.
[01:27:20] Initial: 0000; - Error: Bad packet type from server, expected work assignment
[01:27:20] - Attempt #2 to get work failed, and no other work to do.
Waiting before retry.
[01:27:34] + Attempting to get work packet
[01:27:34] Passkey found
[01:27:34] - Will indicate memory of 12279 MB
[01:27:34] - Connecting to assignment server
[01:27:34] Connecting to http://assign.stanford.edu:8080/
[01:27:35] Posted data.
[01:27:35] Initial: ED82; - Successful: assigned to (130.237.232.237).
[01:27:35] + News From Folding@Home: Welcome to Folding@Home
[01:27:35] Loaded queue successfully.
[01:27:35] Sent data
[01:27:35] Connecting to http://130.237.232.237:8080/
[01:27:35] Posted data.
[01:27:35] Initial: 0000; - Error: Bad packet type from server, expected work assignment
[01:27:35] - Attempt #3 to get work failed, and no other work to do.
Waiting before retry.
[01:27:56] + Attempting to get work packet
[01:27:56] Passkey found
[01:27:56] - Will indicate memory of 12279 MB
[01:27:56] - Connecting to assignment server
[01:27:56] Connecting to http://assign.stanford.edu:8080/
[01:27:56] Posted data.
[01:27:56] Initial: ED82; - Successful: assigned to (130.237.232.237).
[01:27:56] + News From Folding@Home: Welcome to Folding@Home
[01:27:56] Loaded queue successfully.
[01:27:56] Sent data
[01:27:56] Connecting to http://130.237.232.237:8080/
[01:27:57] Posted data.
[01:27:57] Initial: 0000; - Error: Bad packet type from server, expected work assignment
[01:27:57] - Attempt #4 to get work failed, and no other work to do.
Waiting before retry.
130.237.232.237 is in accept status, but marked green. At this time so are:
128.143.48.226
128.143.199.96
128.143.199.97
129.64.95.82
129.74.85.15
171.64.65.84
171.64.65.79
171.64.65.92
171.67.108.22
There are also some servers with 'full' status marked blue, but most of them seem to be marked properly by the AS
whether the AS is ignoring the WS status, or the WS is not reporting it's status to the AS correctly, is anyone's guess.
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 5:19 am
by bruce
slugbug wrote:One of my machines has been waiting for the past 8 hrs or so for work.
If you read all 3 pages of this topic, you'll see that there's a shortage of WUs that may continue for a while. Switch to -smp.
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 5:37 am
by Leonardo
Regarding the assignment servers, they should automatically roll over from bigadv to normal SMP, but that functionality doesn't appear to be working properly.
Yes, I've had two machines' CPUs go idle. All's good though. I am fortunate to be able to monitor my farm even when at work, so I manually changed flags (removed -bigadv) and the clients instantly picked up SMP2 units.
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 5:41 am
by n31l
I'm hoping the -bigadv's come back soon, my 128 thread beastie is currently bored silly.
I bought an i7 2600k a few weeks ago so I could use it to download -bigadv's which I then manually fed to the beastie (it doesn't have internet access currently). It
was running 2x 64 thread Windows 6.34 client's and it spat them out in around 13 hours for around 120k points each
.
I gave normal WU's a try yesterday but even when I run 4 clients with 32 threads it only takes 2 hours and the points are tiny in comparison.
I'm taking it offline for a while today as this 'situation' is the perfect opportunity for me to try and get Ubuntu onto it - I really want to see what 128 threads will do to a single -bigadv WU
** thanks go out to HardOCP forum members for helping me get things going initially, a really helpful community but I'm obliged to join the 'works' team now.
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 9:25 am
by k1wi
With a machine like that I'd be taking a couple of days down time as a small price to pay for some much larger work units that will start out following this (which will come in all sorts of handy in terms of sneakernetting)...
Although having said that, is a machine that powerful being used in an production setting or not (I can't recall from the other thread you had)? Running bigadv might be risky if it is because if FAH brought it down for any reason there might be trouble...
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 10:20 am
by kromberg
So anyone know when WU will start to be assigned again?
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 10:33 am
by ChasR
They haven't stopped being assigned, there is merely a shortage of bigadv WUs. The real problem is that lacking bigadv Wus, the AS isn't switching to regular smp work, so bigadv machines sit idle until they get lucky and are assigned a WU. I got one a few hours ago:
[07:31:40] Connecting to
http://130.237.232.237:8080/
[07:31:46] Posted data.
[07:31:46] Initial: 0000; - Receiving payload (expected size: 24878043)
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 3:33 pm
by n31l
k1wi wrote:With a machine like that I'd be taking a couple of days down time as a small price to pay for some much larger work units that will start out following this (which will come in all sorts of handy in terms of sneakernetting)...
Although having said that, is a machine that powerful being used in an production setting or not (I can't recall from the other thread you had)? Running bigadv might be risky if it is because if FAH brought it down for any reason there might be trouble...
Nope just using it for folding at the moment..
Will I be able to sneakernet the newer bigger -bigadv with an i7 2600k, I've seen 12+ cores mentioned? It's not a major issue as I have a dual Xeon box I can setup to feed the beast with but it's not quite as convenient as using my home PC.
If I could get hold of an Itanium client, I think this x86 box would quake in it's boots
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 6:46 pm
by Grandpa_01
n31l wrote:k1wi wrote:With a machine like that I'd be taking a couple of days down time as a small price to pay for some much larger work units that will start out following this (which will come in all sorts of handy in terms of sneakernetting)...
Although having said that, is a machine that powerful being used in an production setting or not (I can't recall from the other thread you had)? Running bigadv might be risky if it is because if FAH brought it down for any reason there might be trouble...
Nope just using it for folding at the moment..
Will I be able to sneakernet the newer bigger -bigadv with an i7 2600k, I've seen 12+ cores mentioned? It's not a major issue as I have a dual Xeon box I can setup to feed the beast with but it's not quite as convenient as using my home PC.
If I could get hold of an Itanium client, I think this x86 box would quake in it's boots
I very seriously doubt that a 2600K will run them they have said 12 core minimum.
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 11:07 pm
by huyetma
I very seriously doubt that a 2600K will run them they have said 12 core minimum.
isnt that the same with current bigadv, required 8 threads cpu and successfully folded on x6 amd ?
super oced 2600K can match 12 threads cpu mid oc performance.
4.8ghz 2600K =~ 3.6ghz 980x ?
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 11:32 pm
by Grandpa_01
huyetma wrote:I very seriously doubt that a 2600K will run them they have said 12 core minimum.
isnt that the same with current bigadv, required 8 threads cpu and successfully folded on x6 amd ?
super oced 2600K can match 12 threads cpu mid oc performance.
4.8ghz 2600K =~ 3.6ghz 980x ?
I did not say anything about a 2600K being able to run them as far as performance goes. As far as I know there is no VM that you can assign 12 cores to an 8 threaded processor 8 is the max on the VM's I have seen. Thus 12 core requirements cannot be met by a 2600K.
Re: What 's up with the bigadv server(s) ?
Posted: Thu Jun 02, 2011 11:43 pm
by huyetma
why vm and not native linux ?
Re: What 's up with the bigadv server(s) ?
Posted: Fri Jun 03, 2011 12:18 am
by Grandpa_01
Native Linux reports what you have you can not change it as far as I know a 2600K would show as 8 core.
Re: What 's up with the bigadv server(s) ?
Posted: Fri Jun 03, 2011 1:17 am
by Amaruk
huyetma wrote:isnt that the same with current bigadv, required 8 threads cpu and successfully folded on x6 amd ?
Folding bigadv requires 8 cores. Running them on X6 requires lying to Stanford in order to meet that requirement, using VM.
Re: What 's up with the bigadv server(s) ?
Posted: Fri Jun 03, 2011 4:53 am
by codysluder
I'm sure that people will figure out how to lie and tell Stanford they have 12 cores when the really don't. In any case, Stanford just published a Best Practices guide.
http://folding.typepad.com/news/2011/05 ... s-faq.html