FAHClient V7.1.38 released (4th Open-Beta)

[rhavern]

Following clients upgraded over top of existing 7.1.33 install:

Lab1-CPU HP dx2000 MT P4 2.8G 2G XP-32 SP3
Syslog-SMP Tyan Thunder K8S Pro 2x Opt 248 1G XP-32 SP3
Desktop HP 6530b P8700 @ 2.53G 3G W7-32

No installation issues encountered. Tried with FAHcontrol running and stopped, was correctly prompted to stop when left running. No problem with WU resuming. Nice work.

Estimated PPD is not correct immediately on start, it is infeasibly high:

4.3k on P4
44.7k on 2x Opt 248
1.3M on P8700

Will observe to see if it corrects with additional data points.

Edit: Added log snippets and current PPD, still abnormally high but not as ludicrous as initial. It appears that the Estimated TPF is quite incorrect even after a few completed percentage points. Note that all these bogus numbers are from WU that restarted from partially complete.

857 on P4
16.5k on 2x Opt 248
5.6k on P8700

Code: Select all

P4:
10:36:50:Unit 00:Completed 105000 out of 250000 steps  (42%)
10:36:51:Unit 00:Extra SSE boost OK.
10:52:17:Unit 00:Writing local files
10:52:17:Unit 00:Completed 107500 out of 250000 steps  (43%)
10:58:25:SocketDevice::read(): End of stream
10:58:25:Server connection id=1 ended
10:59:27:Server connection id=3 on 0.0.0.0:36330 from 154.8.22.51
10:59:27:Started thread 10 on PID 3060
11:07:50:Unit 00:Writing local files
11:07:50:Unit 00:Completed 110000 out of 250000 steps  (44%)
11:23:08:Unit 00:Writing local files
11:23:08:Unit 00:Completed 112500 out of 250000 steps  (45%)
11:38:24:Unit 00:Writing local files
11:38:25:Unit 00:Completed 115000 out of 250000 steps  (46%)
11:53:47:Unit 00:Writing local files
11:53:47:Unit 00:Completed 117500 out of 250000 steps  (47%)
12:09:06:Unit 00:Writing local files
12:09:06:Unit 00:Completed 120000 out of 250000 steps  (48%)
12:24:29:Unit 00:Writing local files
12:24:29:Unit 00:Completed 122500 out of 250000 steps  (49%)

2x Opt 248
10:52:52:Unit 01:Resuming from checkpoint
10:52:52:Unit 01:Verified 01/wudata_01.log
10:52:52:Unit 01:Verified 01/wudata_01.trr
10:52:52:Unit 01:Verified 01/wudata_01.xtc
10:52:52:Unit 01:Verified 01/wudata_01.edr
10:52:52:Unit 01:Completed 733970 out of 1000000 steps  (73%)
10:55:25:SocketDevice::read(): Receive error: 10054: An existing connection was forcibly closed by the remote host.
10:55:26:Server connection id=2 ended
10:58:26:SocketDevice::read(): End of stream
10:58:26:Server connection id=1 ended
10:59:28:Server connection id=3 on 0.0.0.0:36330 from 154.8.22.51
10:59:28:Started thread 10 on PID 2236
11:03:07:Unit 01:Completed 740000 out of 1000000 steps  (74%)
11:19:33:Unit 01:Completed 750000 out of 1000000 steps  (75%)
11:35:58:Unit 01:Completed 760000 out of 1000000 steps  (76%)
11:52:23:Unit 01:Completed 770000 out of 1000000 steps  (77%)
12:08:48:Unit 01:Completed 780000 out of 1000000 steps  (78%)
12:25:14:Unit 01:Completed 790000 out of 1000000 steps  (79%)

P8700
10:59:41:Unit 01:- Digital signature verified
10:59:41:Unit 01:
10:59:41:Unit 01:Project: 11051 (Run 0, Clone 21, Gen 22)
10:59:41:Unit 01:
10:59:41:Unit 01:Entering M.D.
10:59:47:Unit 01:Using Gromacs checkpoints
10:59:47:Unit 01:Mapping NT from 2 to 2
10:59:47:Unit 01:Resuming from checkpoint
10:59:47:Unit 01:Verified 01/wudata_01.log
10:59:47:Unit 01:Verified 01/wudata_01.trr
10:59:47:Unit 01:Verified 01/wudata_01.xtc
10:59:47:Unit 01:Verified 01/wudata_01.edr
10:59:47:Unit 01:Completed 39910 out of 1000000 steps  (3%)
10:59:58:Unit 01:Completed 40000 out of 1000000 steps  (4%)
11:16:23:Unit 01:Completed 50000 out of 1000000 steps  (5%)
11:31:18:Unit 01:Completed 60000 out of 1000000 steps  (6%)
11:46:07:Unit 01:Completed 70000 out of 1000000 steps  (7%)
12:01:40:Unit 01:Completed 80000 out of 1000000 steps  (8%)
12:17:06:Unit 01:Completed 90000 out of 1000000 steps  (9%)
12:32:11:Unit 01:Completed 100000 out of 1000000 steps  (10%)



[Treefrog]

Installed and working great, even removing the WUs that fail, super.

2 issues. 1st is that the About Project area is just a big grey box, no info what so ever. 2nd can the installer resume the installation after an uninstall from the previous version, i hate having to restart the installer.

[Jesse_V]

2nd can the installer resume the installation after an uninstall from the previous version, i hate having to restart the installer.

I believe this is a ticket currently set for the Public Release milestone: https://fah-web.stanford.edu/projects/F ... ticket/726

[codysluder]

Estimated PPD is not correct immediately on start, it is infeasibly high:

4.3k on P4
44.7k on 2x Opt 248
1.3M on P8700

Will observe to see if it corrects with additional data points.

Edit: Added log snippets and current PPD, still abnormally high but not as ludicrous as initial. It appears that the Estimated TPF is quite incorrect even after a few completed percentage points. Note that all these bogus numbers are from WU that restarted from partially complete.

857 on P4
16.5k on 2x Opt 248
5.6k on P8700

Since PPD and TPF and ETA are all directly related, there are at least two open tickets on this particular subject. One has been around for quite a while and is still unresolved https://fah-web.stanford.edu/projects/F ... ticket/395 and a new one specifically about after resuming work https://fah-web.stanford.edu/projects/F ... ticket/730 I think this will be resolved in a future upgrade.

[^w^ing]

I would like to mention that in the documentation (https://fah-web.stanford.edu/projects/F ... tUserGuide), the "pause-on-start" option is written as "start-paused", which the client won't accept.

[7im]

I would like to mention that in the documentation (https://fah-web.stanford.edu/projects/F ... tUserGuide), the "pause-on-start" option is written as "start-paused", which the client won't accept.

Fixed. Thanks for the heads up.


With all the fixes and improvements, I'm sure there will be more doc changes needed. Keep the reports coming...

[derrickmcc]

Installed v7.1.38 over v7.1.33 on Windows XP 32 bit with 2 GTX 460 GPUs.

Custom install, kept data, install for all users, start as a service, start FAHControl on login.

Went without a hitch and FAHControl started up correctly, no problems this time (upgrade from v7.1.21 to v7.1.33 resulted in greyed out screen which I had to kill and then restart FAHControl, as reported by other testers).

Configuration correctly set from previous installation.

Just checked preferences and noted that Local Client autostart is not ticked.

Also noted that checking the Show Project Info in Advanced box (Preferences, User Interface) displays a big blank box (headed Project 7621) and obscures the Work Queue window.

Congratulations on picking up the pace on releases, and the progress being made

[7im]

Not all projects have descriptions available to display, hence the blank box.

Local Client autostart is one of the check boxes the installer prompts for in the custom setup screen.

The rest sounds great.

[jcoffland]

If "Laptop" is selected, automatically check the "pause on battery" option. If "Desktop" do nothing and just continue the installation. If the user for some reason wanted v7 to run while on battery, this would likely be unusual and so they ought to look under the documentation and the Advanced setting as to how to uncheck that again. This wasn't that big of a deal in v6 as the setting was right there. Thanks.

Please no additional prompting! The install needs to stay simple. Either autodetect it, or default to on (1). Desktops should never be affected, and laptops can turn it off if they so choose.

Let me propose a compromise. Change the default to Pause on Battery=true. I'll bet that Desktops run fine with that setting and that they never happen to get the signal that they're running on battery.

There's more to this issue than meets the eye and we've discussed it internally a few times before as well. The reason pause-on-battery is not defaulted to true is that we are concerned that laptop users will install F@H and then wonder why it's not working. A possible solution to this would be to display a clear message that the client is paused because it's on battery and some indication of how to override this but it's not as simple as just changing the default. Also, I agree with 7im that we should not add any more prompts to the installers. I just created ticket #734.

[bruce]

Also noted that checking the Show Project Info in Advanced box (Preferences, User Interface) displays a big blank box (headed Project 7621) and obscures the Work Queue window.

Show Project Info is supposed to obscure the Work Queue, and it's supposed to show something similar to this information. It was originally restricted to the Novice status screen, where the Work Queue doesn't add any useful information. Uncheck the box if you want to see the queue.

It shouldn't be blank, though. In a recent log, I see the following:

Code: Select all

07:10:48:Unit 01:Completed 280000 out of 1000000 steps  (28%)
07:50:25:Unit 01:Completed 290000 out of 1000000 steps  (29%)
08:03:18:Downloading project 7505 description
08:03:18:Connecting to fah-web.stanford.edu:80
08:27:27:Unit 01:Completed 300000 out of 1000000 steps  (30%)
09:06:12:Unit 01:Completed 310000 out of 1000000 steps  (31%)

Do you have something similar specifying "project 7821 description"

[Jesse_V]

I just created ticket #734.

And this is exactly why I am a complete fan of how your software development is going. Someone puts something forward, someone opposes parts of it, others try to find common ground, and then the developer gets in there and explains the real reason behind things and marks it for follow-up. It's very impressive I think.

Anyway, I just installed v7. Installation went without any issues on my Windows 7 64-bit machine, and the SMP+GPU was all set up right and started off fine. Well done! I've been exploring the software, and looks very impressive and nice. I've also noticed how intuitive the control labels are.

I do have several suggestions, and they're pretty minor. When I switched the GUI over to Advanced/Expert, there're a vertical divider separating the Folding Slots + Work Queue and the Work Unit information on the right hand side. Where it was positioned by default covered up the Credit and PRCG info in the Work Queue area. As the Work Unit area doesn't have as much text going across as the Work Queue does, is it possible to get the text boundary of the PRCG info and set the vertical slider to the end? Because I had to move the divider over myself. Clearly its a pretty minor and low priority issue, but I felt like I should point it out because it would be nice if that was done automatically as other software does. Perhaps the vertical divider is set at its default position so that things work no matter what people's screen resolution is, at least for me I'd rather have the end of the Unit ID cut off than not be able to see the PRCG.

My second suggestion is to add some mouse-over text to those icons on the right hand side in FAHViewer. Just so I see instantly what they do without being slightly confused as to what action they are going to take. They are very appropriate icons yes, but mouse-over text that says "Information" or "Help" or "About" would be nice. On the same note I was very happily surprised to notice how you can drag the protein around with your mouse and scroll in with your mouse wheel. Pretty awesome!

Neither of these are a big deal, and overall I'm very impressed.

[Jesse_V]

Also noted that checking the Show Project Info in Advanced box (Preferences, User Interface) displays a big blank box (headed Project 7621) and obscures the Work Queue window.

Show Project Info is supposed to obscure the Work Queue, and it's supposed to show something similar to this information. It was originally restricted to the Novice status screen, where the Work Queue doesn't add any useful information. Uncheck the box if you want to see the queue.

It shouldn't be blank, though.

I actually experienced the same issue. While I think that the Project Info should not obscure the Work Queue in Advanced/Expert, and its relatively easy to drag the dividers around to get everything straightened out again, I'm not seeing anything for my project description either.

Just like that other log, mine shows that it was downloaded as well:

Code: Select all

21:52:52:News: Welcome to Folding@Home
21:52:52:Assigned to work server 129.74.85.15
21:52:52:Requesting new work unit for slot 01: READY smp:4 from 129.74.85.15
21:52:52:Connecting to 129.74.85.15:8080
21:52:52:Slot 01: Downloading 50.21KiB
21:52:52:Slot 00: Downloading 61.87KiB
21:52:52:Slot 00: Download complete
21:52:52:Received Unit: id:00 state:DOWNLOAD error:OK project:10504 run:61 clone:0 gen:563 core:0x11 unit:0x0000059d6652eda54b75aecb0000241b
21:52:52:Downloading core from http://www.stanford.edu/~pande/Win32/AMD64/NVIDIA/G80/Core_11.fah
21:52:52:Connecting to www.stanford.edu:80
21:52:52:Slot 01: Download complete
21:52:52:Received Unit: id:01 state:DOWNLOAD error:OK project:7013 run:3 clone:90 gen:47 core:0xa4 unit:0x0000009c0001329c4dfb9abb032b3c83
21:52:52:Downloading core from http://www.stanford.edu/~pande/Win32/AMD64/Core_a4.fah
21:52:52:Connecting to www.stanford.edu:80
21:52:52:FahCore 11: Downloading 648.82KiB
21:52:52:FahCore a4: Downloading 2.89MiB
21:52:53:Server connection id=1 on 0.0.0.0:36330 from 127.0.0.1
21:52:53:FahCore 11: Download complete
21:52:53:Valid core signature
21:52:53:Unpacked 1.82MiB to cores/www.stanford.edu/~pande/Win32/AMD64/NVIDIA/G80/Core_11.fah/FahCore_11.exe
21:52:53:Starting Unit 00
21:52:53:Running core: C:/Users/Jesse/AppData/Roaming/FAHClient/cores/www.stanford.edu/~pande/Win32/AMD64/NVIDIA/G80/Core_11.fah/FahCore_11.exe -dir 00 -suffix 01 -lifeline 5000 -version 701 -checkpoint 15 -gpu 0
21:52:54:Started core on PID 3540
21:52:54:FahCore 0x11 started
21:52:54:Unit 00:
21:52:54:Unit 00:*------------------------------*
21:52:54:Unit 00:Folding@Home GPU Core
21:52:54:Unit 00:Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
21:52:54:Unit 00:
21:52:54:Unit 00:Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
21:52:54:Unit 00:Build host: amoeba
21:52:54:Unit 00:Board Type: Nvidia
21:52:54:Unit 00:Core      :
21:52:54:Unit 00:Preparing to commence simulation
21:52:54:Unit 00:- Looking at optimizations...
21:52:54:Unit 00:DeleteFrameFiles: successfully deleted file=00/wudata_01.ckp
21:52:54:Unit 00:- Created dyn
21:52:54:Unit 00:- Files status OK
21:52:54:Unit 00:- Expanded 62845 -> 336799 (decompressed 535.9 percent)
21:52:54:Unit 00:Called DecompressByteArray: compressed_data_size=62845 data_size=336799, decompressed_data_size=336799 diff=0
21:52:54:Unit 00:- Digital signature verified
21:52:54:Unit 00:
21:52:54:Unit 00:Project: 10504 (Run 61, Clone 0, Gen 563)
21:52:54:Unit 00:
21:52:54:Unit 00:Assembly optimizations on if available.
21:52:54:Unit 00:Entering M.D.
21:52:56:FahCore a4: Download complete
21:52:56:Valid core signature
21:52:56:Unpacked 9.59MiB to cores/www.stanford.edu/~pande/Win32/AMD64/Core_a4.fah/FahCore_a4.exe
21:52:57:Starting Unit 01
21:52:57:Running core: C:/Users/Jesse/AppData/Roaming/FAHClient/cores/www.stanford.edu/~pande/Win32/AMD64/Core_a4.fah/FahCore_a4.exe -dir 01 -suffix 01 -lifeline 5000 -version 701 -checkpoint 15 -np 4
21:52:58:Started core on PID 4928
21:52:58:FahCore 0xa4 started
21:52:58:Unit 01:
21:52:58:Unit 01:*------------------------------*
21:52:58:Unit 01:Folding@Home Gromacs GB Core
21:52:58:Unit 01:Version 2.27 (Dec. 15, 2010)
21:52:58:Unit 01:
21:52:58:Unit 01:Preparing to commence simulation
21:52:58:Unit 01:- Looking at optimizations...
21:52:58:Unit 01:- Created dyn
21:52:59:Unit 01:- Files status OK
21:52:59:Unit 01:- Expanded 50906 -> 198568 (decompressed 390.0 percent)
21:52:59:Unit 01:Called DecompressByteArray: compressed_data_size=50906 data_size=198568, decompressed_data_size=198568 diff=0
21:52:59:Unit 01:- Digital signature verified
21:52:59:Unit 01:
21:52:59:Unit 01:Project: 7013 (Run 3, Clone 90, Gen 47)
21:52:59:Unit 01:
21:52:59:Unit 01:Assembly optimizations on if available.
21:52:59:Unit 01:Entering M.D.
21:53:00:Unit 00:Tpr hash 00/wudata_01.tpr:  1725232751 2483324523 963396390 613730075 2194324458
21:53:00:Unit 00:
21:53:00:Unit 00:Calling fah_main args: 14 usage=100
21:53:00:Unit 00:
21:53:01:Unit 00:Working on Protein
21:53:02:Unit 00:Client config unavailable.
21:53:02:Unit 00:Starting GUI Server
21:53:04:Unit 01:Mapping NT from 4 to 4
21:53:04:Unit 01:Completed 0 out of 10000000 steps  (0%)
21:54:49:Server connection id=2 on 0.0.0.0:36330 from 127.0.0.1
21:55:00:Server connection id=2 ended
21:55:02:Server connection id=3 on 0.0.0.0:36330 from 127.0.0.1
21:55:46:Server connection id=3 ended
21:57:46:Unit 00:Completed 1%
21:57:51:Downloading project 7013 description
21:57:51:Connecting to fah-web.stanford.edu:80
21:57:52:Downloading project 10504 description
21:57:52:Connecting to fah-web.stanford.edu:80
21:59:16:Unit 01:Completed 100000 out of 10000000 steps  (1%)
22:02:26:Unit 00:Completed 2%
22:05:02:Unit 01:Completed 200000 out of 10000000 steps  (2%)
22:05:43:Server connection id=4 on 0.0.0.0:36330 from 127.0.0.1
22:06:28:Server connection id=4 ended
22:06:37:Server connection id=5 on 0.0.0.0:36330 from 127.0.0.1
22:07:05:Server connection id=5 ended
22:07:16:Unit 00:Completed 3%
22:08:09:Server connection id=6 on 0.0.0.0:36330 from 127.0.0.1
22:08:13:Server connection id=6 ended
22:10:10:Server connection id=7 on 0.0.0.0:36330 from 127.0.0.1
22:10:19:Server connection id=7 ended
22:11:13:Unit 01:Completed 300000 out of 10000000 steps  (3%)
22:11:57:Unit 00:Completed 4%
22:16:32:Unit 00:Completed 5%
22:17:11:Unit 01:Completed 400000 out of 10000000 steps  (4%)


[Grandpa_01]

Well I did not have to use purge fahclient this time. I un installed .33 using software manager and installed .38 over the top and it started right up and continued the WU with no problems this time. Thanks for fixing the .deb package. It now appears to work on Ubuntu 10.10

[^w^ing]

I had to stop using the v7 client when it was in 2nd open beta because my fahcontrol kept getting completely grayed out and the fahclient had very slow transfer rates. I am happy to say that with this build fahcontrol is working nicely in that regard, but unfortunately I have to say, from what I have seen so far, the transfer rates of the fahclient are still painfully slow compared to the v6 client with the same WUs and therefore the same servers.

[Tagus]

2 issues. 1st is that the About Project area is just a big grey box, no info what so ever.

Not all projects have descriptions available to display, hence the blank box.

I also have a blank About Project area. Client is doing Project 7809, which has a description.