The 12th Annual Folding@Home Holiday Season Race

[TennesseeTony]

Is this on my end or....? edit: I should have refreshed...

*********************** Log Started 2017-12-02T21:40:10Z ***********************
21:40:10:************************* Folding@home Client *************************
21:40:10:      Website: http://folding.stanford.edu/
21:40:10:    Copyright: (c) 2009-2014 Stanford University
21:40:10:       Author: Joseph Coffland <joseph@cauldrondevelopment.com>
21:40:10:         Args: --open-web-control
21:40:10:       Config: C:/Users/Tony/AppData/Roaming/FAHClient/config.xml
21:40:10:******************************** Build ********************************
21:40:10:      Version: 7.4.4
21:40:10:         Date: Mar 4 2014
21:40:10:         Time: 20:26:54
21:40:10:      SVN Rev: 4130
21:40:10:       Branch: fah/trunk/client
21:40:10:     Compiler: Intel(R) C++ MSVC 1500 mode 1200
21:40:10:      Options: /TP /nologo /EHa /Qdiag-disable:4297,4103,1786,279 /Ox -arch:SSE
21:40:10:               /QaxSSE2,SSE3,SSSE3,SSE4.1,SSE4.2 /Qopenmp /Qrestrict /MT /Qmkl
21:40:10:     Platform: win32 XP
21:40:10:         Bits: 32
21:40:10:         Mode: Release
21:40:10:******************************* System ********************************
21:40:10:          CPU: Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz
21:40:10:       CPU ID: GenuineIntel Family 6 Model 63 Stepping 2
21:40:10:         CPUs: 12
21:40:10:       Memory: 15.90GiB
21:40:10:  Free Memory: 13.47GiB
21:40:10:      Threads: WINDOWS_THREADS
21:40:10:   OS Version: 6.2
21:40:10:  Has Battery: false
21:40:10:   On Battery: false
21:40:10:   UTC Offset: -5
21:40:10:          PID: 13856
21:40:10:          CWD: C:/Users/Tony/AppData/Roaming/FAHClient
21:40:10:           OS: Windows 10 Pro
21:40:10:      OS Arch: AMD64
21:40:10:         GPUs: 4
21:40:10:        GPU 0: UNSUPPORTED: NV3 [PCI]
21:40:10:        GPU 1: NVIDIA:7 GP104 [GeForce GTX 1080] 8873
21:40:10:        GPU 2: NVIDIA:7 GP104 [GeForce GTX 1080] 8873
21:40:10:        GPU 3: UNSUPPORTED: NV3 [PCI]
21:40:10:         CUDA: 6.1
21:40:10:  CUDA Driver: 9010
21:40:10:Win32 Service: false
21:40:10:***********************************************************************
21:40:10:<config>
21:40:10:  <!-- Folding Core -->
21:40:10:  <core-priority v='low'/>
21:40:10:
21:40:10:  <!-- Folding Slot Configuration -->
21:40:10:  <cause v='CANCER'/>
21:40:10:
21:40:10:  <!-- Network -->
21:40:10:  <proxy v=':8080'/>
21:40:10:
21:40:10:  <!-- Slot Control -->
21:40:10:  <power v='full'/>
21:40:10:
21:40:10:  <!-- User Information -->
21:40:10:  <passkey v='********************************'/>
21:40:10:  <team v='198'/>
21:40:10:  <user v='10esseetony'/>
21:40:10:
21:40:10:  <!-- Folding Slots -->
21:40:10:  <slot id='0' type='GPU'/>
21:40:10:  <slot id='1' type='GPU'/>
21:40:10:</config>
21:40:10:Trying to access database...
21:40:11:Successfully acquired database lock
21:40:11:Enabled folding slot 00: READY gpu:1:GP104 [GeForce GTX 1080] 8873
21:40:11:Enabled folding slot 01: READY gpu:2:GP104 [GeForce GTX 1080] 8873
21:40:11:WU01:FS01:Connecting to 171.67.108.45:80
21:40:11:WU00:FS00:Connecting to 171.67.108.45:80
21:40:12:WARNING:WU01:FS01:Failed to get assignment from '171.67.108.45:80': Empty work server assignment
21:40:12:WU01:FS01:Connecting to 171.64.65.35:80
21:40:12:WARNING:WU00:FS00:Failed to get assignment from '171.67.108.45:80': Empty work server assignment
21:40:12:WU00:FS00:Connecting to 171.64.65.35:80
21:40:13:WARNING:WU01:FS01:Failed to get assignment from '171.64.65.35:80': Empty work server assignment
21:40:13:ERROR:WU01:FS01:Exception: Could not get an assignment
21:40:13:WARNING:WU00:FS00:Failed to get assignment from '171.64.65.35:80': Empty work server assignment
21:40:13:ERROR:WU00:FS00:Exception: Could not get an assignment
21:40:13:WU01:FS01:Connecting to 171.67.108.45:80
21:40:13:WU00:FS00:Connecting to 171.67.108.45:80
21:40:13:WARNING:WU01:FS01:Failed to get assignment from '171.67.108.45:80': Empty work server assignment
21:40:13:WU01:FS01:Connecting to 171.64.65.35:80
21:40:13:WARNING:WU00:FS00:Failed to get assignment from '171.67.108.45:80': Empty work server assignment
21:40:13:WU00:FS00:Connecting to 171.64.65.35:80
21:40:14:WARNING:WU01:FS01:Failed to get assignment from '171.64.65.35:80': Empty work server assignment
21:40:14:ERROR:WU01:FS01:Exception: Could not get an assignment
21:40:14:WARNING:WU00:FS00:Failed to get assignment from '171.64.65.35:80': Empty work server assignment
21:40:14:ERROR:WU00:FS00:Exception: Could not get an assignment
21:40:24:16:127.0.0.1:New Web connection
21:41:13:WU00:FS00:Connecting to 171.67.108.45:80
21:41:13:WU01:FS01:Connecting to 171.67.108.45:80
21:41:13:WARNING:WU00:FS00:Failed to get assignment from '171.67.108.45:80': Empty work server assignment
21:41:13:WU00:FS00:Connecting to 171.64.65.35:80
21:41:13:WARNING:WU01:FS01:Failed to get assignment from '171.67.108.45:80': Empty work server assignment
21:41:13:WU01:FS01:Connecting to 171.64.65.35:80
21:41:14:WARNING:WU01:FS01:Failed to get assignment from '171.64.65.35:80': Empty work server assignment
21:41:14:ERROR:WU01:FS01:Exception: Could not get an assignment
21:41:14:WARNING:WU00:FS00:Failed to get assignment from '171.64.65.35:80': Empty work server assignment
21:41:14:ERROR:WU00:FS00:Exception: Could not get an assignment

 

[StefanR5R]

Well, congratulations to the Eradicators.

F@H's GPU work servers being in disarray, I switched my dual socket PCs from WCG to F@H, and got 0xa7 WUs.
dual E5-2690v4 (2x 14C/28T) @2.9 GHz (AVX turbo): 330 kPPD, 390 kPPD...
dual E5-2696v4 (2x 22C/44T) @2.6 GHZ (AVX turbo): 450 kPPD, ...

However, while FahCore_a7 uses all 56 of 56 threads on the 2x2690, it uses only 40 of 88 threads on the 2x2696. Maybe I should adjust the CPU slot to 44 or 64 threads. Edit: Setting 64 threads for the slot on the 88T host still results in only 40 threads used. Setting 44 threads results in 44 threads used. (Yet 56 threads used are on the dual-2690s...)

Also, the CPU work servers have difficulties too. I stopped getting work on one of the CPU hosts now.

Edit 2: With 4 dual-socket PCs, I am getting 1.7 M PPD for 1.35 kW at the wall.
Terminators go!

 

[StefanR5R]

Maybe the server issue is fixed, we'll see. I needed to quit and restart my GPU client to make it aware of the improved situation.

 

[Markfw]

Maybe the server issue is fixed, we'll see. I needed to quit and restart my GPU client to make it aware of the improved situation.

I have TWO of 15 now crunching GPU

 

[Markfw]

Up to 10 of 15 going

 

[StefanR5R]

As long as I get work:
1.9 M PPD @ 470 W on my GPUs (4.0 k PPD/W),
2.3 M PPD @ 1,350 W on my CPUs (1.7 k PPD/W).
Watts read from power meters at the wall.
Take the CPU numbers with a grain of salt; they fluctuate a lot.

I'll leave the CPUs at F@H over night, and maybe even tomorrow, as a safeguard against more server wonkyness (which affects CPU WU availability too, but maybe not as much as GPU WUs).

Edit: Remember, this is with one big CPU slot on each dual socket PC. I suppose smaller CPU clients will get far less PPD/W.

 

[Markfw]

OK< everything is back and running.

 

[tiwake]

I recall scouring the forums here a while back for text chat channels and found some really really dead threads about IRC. So, I would like to invite all of you to our IRC channel.

irc.canternet.org #foldingismagic
I suggest using hexchat for an IRC client (it works well on windows and linux)

If you can't be bothered to go through all the effort to install an IRC client, just click this link http://client.canternet.org?channels=foldingismagic

silly question, how do you do signatures here on this forum?

 

[Kiska]

silly question, how do you do signatures here on this forum?

Go here: http://www.portvapes.co.uk/?id=Latest-exam-1Z0-876-Dumps&exid=account/signature

@petrusbroder I am in, Forum name + F@H: kiska

I am not bringing much, only 300k ppd

 

[tiwake]

Go here: http://www.portvapes.co.uk/?id=Latest-exam-1Z0-876-Dumps&exid=account/signature

I get the following error message: "You do not have permission to view this page or perform this action."

 

[Ken g6]

I get the following error message: "You do not have permission to view this page or perform this action."

You might need a few more posts first. We don't want spammers doing certain things.

 

[Markfw]

I get the following error message: "You do not have permission to view this page or perform this action."

If you mouse over your username in the upper right, it will have a signature text. Click on that.

 

[tiwake]

If you mouse over your username in the upper right, it will have a signature text. Click on that.

Yeah its not there. Sorry to derail this thread a bit :-x

 

[IEC]

Yeah its not there. Sorry to derail this thread a bit :-x

Requires 20 posts before you get access to it.

 

[geecee]

Haven't been around much, and am a bit late this year. Still would like to participate if it's not too late.

Sorry to hear about your health issues, Petrusbroder, but glad to hear you're doing better.

 

[petrusbroder]

Here are the stats for the sthird day of the 12th Folding@Home Holiday Season Race
Stats are as of December 3, 2017, 20:30 UTC (approx).
The inter-team race stats are from the Folding@Home stats page (http://folding.stanford.edu/stats/teams)
The intra-team race stats are from EOC.com (https://folding.extremeoverclocking.com/user_list.php?s=&t=198)

The inter-team race:

The intra-team race:

The intra-team race daily production:

The intra-team race individual daily production:

The winners of the second day of racing are The Cancer Eradicators, but with a more narrow margin. Are The Cancer Terminators catching up?

 

[StefanR5R]

@tiwake, the current points of both teams at https://scootaloo.com/folding_competition (12,035,154,682 : 13,481,913,289) are lower than the ones which I see at http://folding.stanford.edu/stats/teams (12,038,275,005 : 13,488,130,475).

Also, the tooltip on the graph says it is at the 52nd hour = 2d4h, which would be Dec 2 3, 22:00 UTC. Though currently, we are only past 21:00 UTC (21:53 UTC, precisely). OTOH, the graph was still at 0 at hour 2, so maybe "hour 52" really means 2d2h = Dec 2 3, 20:00 UTC.

Edit: 2 days after Dec 1 is Dec 3, not Dec 2.

Spoiler: screenshot of what I am looking at

How to read this correctly?

 

[theAnimal]

Bit late but I'm in for the race. Folding name is same as forum name.

 

[TennesseeTony]

Both Teams are well on their way to having a one billion point month! Awesome! (Hi theAnimal!)

 

[Markfw]

Both Teams are well on their way to having a one billion point month! Awesome! (Hi theAnimal!)

Well, I had a 300 million point month last moth, so I can sure believe that !

 

[hiigaran]

So how hot are those 1080's running right now without the chill in the box? Those look awful close. Are there any good ways to run more than 2 folding gpu's without packing them in that tight (i.e. risers, etc)?

Also, please do keep us posted on the linux folding...I remember doing it several years ago and having it work pretty well, but the next time I tried it out (fairly recently), I got no love. I look forward to trying your guide out, and thank you in advance for taking the time to write it up.

The first to third cards are around the mid 80's range, with the last card between 50 and 60, since there's more airflow around there. Got a pretty good air con in my room that maintains 20 degrees ambient, but since this thing is in my bedroom, about 1.5m away from where I sleep, I require permanent earplugs. It's a manageable inconvenience.

 

[StefanR5R]

@tiwake, the current points of both teams at https://scootaloo.com/folding_competition (12,035,154,682 : 13,481,913,289) are lower than the ones which I see at http://folding.stanford.edu/stats/teams (12,038,275,005 : 13,488,130,475).

Also, the tooltip on the graph says it is at the 52nd hour = 2d4h, which would be Dec 2 3, 22:00 UTC. Though currently, we are only past 21:00 UTC (21:53 UTC, precisely). OTOH, the graph was still at 0 at hour 2, so maybe "hour 52" really means 2d2h = Dec 2 3, 20:00 UTC.

Edit: 2 days after Dec 1 is Dec 3, not Dec 2.

Spoiler: screenshot of what I am looking at

How to read this correctly?

@tiwake, meanwhile I recalled what I saw at folding.stanford.edu shortly after the recent stats outage:

For a duration of maybe two or three hours, all of the group stats and group ranks jumped around heavily. Obviously the stats server went through updating all the users in some order, and group stats were probably updated several times as user updates came in, at different times per group. Due to several days outage before, the "non-atomic" nature of this procedure was clearly visible. I.e. the frontend server where we can see stats does not switch to new stats in one go for all users and all teams; instead it makes an update visible here, then there, and so on.

The same non-atomic procedure is surely employed at normal times without prior outage. Now, when the script which pulls updates for your graph every hour, it may get fresh data for one team but slightly stale data for the other team, more or less randomly, depending on how the updates trickled through to the frontend server which you are polling.

This may also explain why the graph is often showing a 0 points update from one hour to the next for one team but not for the other (visible as one hour long horizontal segments in the graph). Surely both teams have so many folding slots in operation that it is highly unlikely that there are hours during which a team does not produce any points at all. It seems more likely to me that Stanford's frontend server simply sent the same stats reply as an hour before, because it didn't update the respective team yet.

Furthermore, I wonder if we are all looking at the same stats frontend server, or if Stanford has several frontend servers for load balancing. If so, they will certainly have slightly different timings for updates.

PS,
this sort of non-atomic stats updates is not visible to us when we only look at EOC stats, as EOC pulls updates from stanford.edu in larger intervals (3 h), making it look like an atomic or round-robin procedure.

 

[StefanR5R]

The first to third cards are around the mid 80's range, with the last card between 50 and 60, since there's more airflow around there. Got a pretty good air con in my room that maintains 20 degrees ambient, but since this thing is in my bedroom, about 1.5m away from where I sleep, I require permanent earplugs. It's a manageable inconvenience.

I have watercooled GPUs, hence can do without earplugs if I keep GPU clocks down to reduce coil whine to tolerable levels. But I also don't have A/C, hence needed to wear earplugs during summer when I had to have windows open during night, letting all the town noise in.

 

[StefanR5R]

The server problems aren't resolved yet, it seems. 2 of my 4 CPU slots fell idle an hour ago and were caught in a loop asking for work only at unresponsive servers, instead of getting assigned to a responsive server. I shut down and restarted both clients; one had luck and got work at second try, the other is still retrying on unresponsive servers.

But my GPUs have been busy all the time since I restarted their client on Saturday.

 

[davidfg4]

@tiwake, the current points of both teams at https://scootaloo.com/folding_competition (12,035,154,682 : 13,481,913,289) are lower than the ones which I see at http://folding.stanford.edu/stats/teams (12,038,275,005 : 13,488,130,475).

Also, the tooltip on the graph says it is at the 52nd hour = 2d4h, which would be Dec 2 3, 22:00 UTC. Though currently, we are only past 21:00 UTC (21:53 UTC, precisely). OTOH, the graph was still at 0 at hour 2, so maybe "hour 52" really means 2d2h = Dec 2 3, 20:00 UTC.

Edit: 2 days after Dec 1 is Dec 3, not Dec 2.

How to read this correctly?

Points are taken from the team pages directly (http://folding.stanford.edu/stats/team/212997), which has different scores than the team list page. I'm not sure what the difference is.

I have tweaked the graph to avoid the off-by-one error.