Dark theme replacement of the standard F@H web client.

[Joe_H]

It is online: https://foldingathome.org/beta/

Online does not equal publicly announced. It was posted there for invited testers to get the installer, that includes a few 3rd party app developers. There will be an announcement when a version is ready for public beta testing.

As I mentioned, this version has problems, fixes for 3rd party apps to work with it may become moot as those problems get fixed in a latter version.

[ajm]

What is online on a web page without protection is de facto public. I don't recall that it would need some official declaration, nor that there was any for 7.6.13, for example. It was announced only when it became the "New Release."

Besides, this link also was published tonight on the discord server of the FAH Translation team (which I'm a member of) by Anton Thynell, without any reserve.
https://discord.com/channels/7054708144 ... 7824557066

And Bruce told about it this morning on this forum: viewtopic.php?p=341355#p341355

As my main machine has had trouble with FAH for months, I couldn't help but try it. Now if this version is faulty, it must be taken down.

[Neil-B]

If a beta test was aborted every time a bug/fault was found there would be an infinite number of beta candidates ... Some bugs may no doubt have been found but actually leaving it accessible for invited testers with specific combinations of kit and/or for devs to check compatibility with parts which might not be bugged or likely to change is perfectly valid ... It will no doubt take it down as/when it ceases to be of use/needed.

[ajm]

Then it looks like 7.6.15 changes something in the description of GPUs, something https://folding.lar.systems/ are using for their database. Maybe they should be among the invited testers?

[bruce]

I'm running the *.15 beta and I don't see a number like that. Please post the segment of the log that you're looking at.

If we need to report an unusual behavior of 7.6.15 the people that need to fix it have to see what you're talking about.

[ajm]

I don't see anything unusual in the log:

Code: Select all

*********************** Log Started 2020-08-16T08:51:12Z ***********************
08:51:12:******************************* libFAH ********************************
08:51:12:           Date: Aug 6 2020
08:51:12:           Time: 12:41:41
08:51:12:       Revision: c8afa4a61c11532766fdd22f839ce8ab12abd88e
08:51:12:         Branch: master
08:51:12:       Compiler: Visual C++ 2015
08:51:12:        Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
08:51:12:       Platform: win32 10
08:51:12:           Bits: 32
08:51:12:           Mode: Release
08:51:12:****************************** FAHClient ******************************
08:51:12:        Version: 7.6.15
08:51:12:         Author: Joseph Coffland <joseph@cauldrondevelopment.com>
08:51:12:      Copyright: 2020 foldingathome.org
08:51:12:       Homepage: https://foldingathome.org/
08:51:12:           Date: Aug 14 2020
08:51:12:           Time: 17:01:16
08:51:12:       Revision: bbc81e8458541498156e28db7a1fa6af381e6787
08:51:12:         Branch: master
08:51:12:       Compiler: Visual C++ 2015
08:51:12:        Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
08:51:12:       Platform: win32 10
08:51:12:           Bits: 32
08:51:12:           Mode: Release
08:51:12:           Args: --open-web-control
08:51:12:         Config: C:\Users\ajmch\AppData\Roaming\FAHClient\config.xml
08:51:12:******************************** CBang ********************************
08:51:12:           Date: Aug 6 2020
08:51:12:           Time: 13:14:21
08:51:12:       Revision: eadd6f4368f7828f0e0a8433795546f2b1642c88
08:51:12:         Branch: master
08:51:12:       Compiler: Visual C++ 2015
08:51:12:        Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
08:51:12:       Platform: win32 10
08:51:12:           Bits: 32
08:51:12:           Mode: Release
08:51:12:******************************* System ********************************
08:51:12:            CPU: AMD Ryzen Threadripper 3970X 32-Core Processor
08:51:12:         CPU ID: AuthenticAMD Family 23 Model 49 Stepping 0
08:51:12:           CPUs: 32
08:51:12:         Memory: 127.87GiB
08:51:12:    Free Memory: 56.73GiB
08:51:12:        Threads: WINDOWS_THREADS
08:51:12:     OS Version: 6.2
08:51:12:    Has Battery: true
08:51:12:     On Battery: false
08:51:12:     UTC Offset: 2
08:51:12:            PID: 41028
08:51:12:            CWD: C:\Users\ajmch\AppData\Roaming\FAHClient
08:51:12:  Win32 Service: false
08:51:12:             OS: Windows 10 Enterprise
08:51:12:        OS Arch: AMD64
08:51:12:           GPUs: 2
08:51:12:          GPU 0: Bus:78 Slot:0 Func:0 NVIDIA:8 TU104 [GeForce RTX 2080 SUPER]
08:51:12:          GPU 1: Bus:1 Slot:0 Func:0 NVIDIA:8 TU104 [GeForce RTX 2080 SUPER]
08:51:12:  CUDA Device 0: Platform:0 Device:0 Bus:1 Slot:0 Compute:7.5 Driver:11.0
08:51:12:  CUDA Device 1: Platform:0 Device:1 Bus:78 Slot:0 Compute:7.5 Driver:11.0
08:51:12:OpenCL Device 0: Platform:0 Device:0 Bus:1 Slot:0 Compute:1.2 Driver:451.77
08:51:12:OpenCL Device 1: Platform:0 Device:1 Bus:78 Slot:0 Compute:1.2 Driver:451.77
08:51:12:***********************************************************************
08:51:12:<config>
08:51:12:  <!-- Folding Slot Configuration -->
08:51:12:  <client-type v='advanced'/>
08:51:12:
08:51:12:  <!-- HTTP Server -->
08:51:12:  <allow v='127.0.0.1 192.168.1.126-192.168.1.127'/>
08:51:12:
08:51:12:  <!-- Network -->
08:51:12:  <proxy v=':8080'/>
08:51:12:
08:51:12:  <!-- Remote Command Server -->
08:51:12:  <command-allow-no-pass v='127.0.0.1 192.168.1.126-192.168.1.127'/>
08:51:12:
08:51:12:  <!-- Slot Control -->
08:51:12:  <power v='FULL'/>
08:51:12:
08:51:12:  <!-- User Information -->
08:51:12:  <passkey v='*****'/>
08:51:12:  <team v='182116'/>
08:51:12:  <user v='ajm'/>
08:51:12:
08:51:12:  <!-- Folding Slots -->
08:51:12:  <slot id='1' type='GPU'>
08:51:12:    <pause-on-start v='true'/>
08:51:12:    <paused v='true'/>
08:51:12:  </slot>
08:51:12:  <slot id='0' type='GPU'>
08:51:12:    <paused v='true'/>
08:51:12:  </slot>
08:51:12:</config>
08:51:12:Trying to access database...
08:51:12:Successfully acquired database lock
08:51:12:FS01:Initialized folding slot 01: gpu:78:0 TU104 [GeForce RTX 2080 SUPER] - PAUSED by user
08:51:12:FS00:Initialized folding slot 00: gpu:1:0 TU104 [GeForce RTX 2080 SUPER] - PAUSED by user

And both GPUS have been folding with 7.6.15 without any issue since the installation. But the fact is that https://folding.lar.systems/ Chrome extension has been in trouble since then, too.

EDIT: Ah, those lines here:

Code: Select all

08:51:12:FS01:Initialized folding slot 01: gpu:78:0 TU104 [GeForce RTX 2080 SUPER] - PAUSED by user
08:51:12:FS00:Initialized folding slot 00: gpu:1:0 TU104 [GeForce RTX 2080 SUPER] - PAUSED by user

And later on, FAH keeps calling them "0 TU104":

Code: Select all

08:57:56:WU03:FS01:Requesting new work unit for slot 01: gpu:78:0 TU104 [GeForce RTX 2080 SUPER] - RUNNING from 18.188.125.154
(...)
10:39:05:WU01:FS00:Requesting new work unit for slot 00: gpu:1:0 TU104 [GeForce RTX 2080 SUPER] - RUNNING from 143.89.243.111
(...)
11:08:06:WU00:FS01:Requesting new work unit for slot 01: gpu:78:0 TU104 [GeForce RTX 2080 SUPER] - RUNNING from 143.89.243.111
(...)
11:43:08:WU02:FS00:Requesting new work unit for slot 00: gpu:1:0 TU104 [GeForce RTX 2080 SUPER] - RUNNING from 18.188.125.154

and so on...

The whole log since the 7.6.15 update and until now: https://web.tresorit.com/l/sXFi9#Lq2SJDt_J9jXBzBD3JXeDg
The previous log up to the update: https://web.tresorit.com/l/g8htt#g_HqYnM-0I4v1uOVdQW7fw

Finishing all slots now.

[bruce]

Any beta version of anything carries risk. Whether you heard about it from Anton or Bruce or you next-door neighbor you cannot expect it to be bug-free. When the first bug is discovered, that version is NOT withdrawn. It's your responsibility to report bugs to the developer so the next version doesn't have any of the ones that were reported (or it should carry a disclaimer "I know about the xxx bug and a fix for it was not included yet.")

[ajm]

Of course, Bruce, no problem. And the cards did fold quite well, and I got the beloved points for it. Everything's fine. But I'm afraid that Joe is right and that this version will have to stay a beta...

[bruce]

Good plan (@joe). We still are obligated to report unreported bugs so 7.6.16+ will be cleaner.

[marknd59]

And both GPUS have been folding with 7.6.15 without any issue since the installation. But the fact is that https://folding.lar.systems/ Chrome extension has been in trouble since then, too.

The database has had a quick patch and should be working again for your cards, but as this seems to be a problem with the beta client there is not much more to do.

[ajm]

Yes, thank you! But I have uninstalled that beta and this bug should not make it to the release.

[marknd59]

Yes, thank you! But I have uninstalled that beta and this bug should not make it to the release.

Glad to help out. If you have any questions, problems or would like to make suggestions to improve the extention feel free to pop over to the extention thread at the LTT forum.

https://linustechtips.com/main/topic/12 ... by-gpu-wu/

[marknd59]

Yes, thank you! But I have uninstalled that beta and this bug should not make it to the release.

Thanks for coming over to LTT to keep us upto date with whats going on. No more weird GUP types have have shown up in the database so it looks like the patch is working and as you said this problem should not make to the release so it will go away.

[ajm]

Yes, thank you! But I have uninstalled that beta and this bug should not make it to the release.

Thanks for coming over to LTT to keep us upto date with whats going on. No more weird GUP types have have shown up in the database so it looks like the patch is working and as you said this problem should not make to the release so it will go away.

Well, no, I don't think so. Two other betas came out since then and both had that same issue. Another software affected, HFM.NET, has been adapted to cope with it. And if the database hasn't seen any new weird GPUs, it's because the extension now just ignores them.

[LAR_Systems]

I have accounted for the changes to GPU description format in the latest build that shows the cause and atoms with the GPU slot.

I am going to try and keep an eye on the change logs on this forum more often to spot such issues going forward and have made some changes to my services and extension code to allow hotfixes to be deployed faster when needing to adapt to changes in the main F@H client.