F@H January TeAm record run?

[Endgame124]

Last year we set a new team record during the month of January for F@H points. Do we want to do that again?

Last years thread for reference:

F@H Holiday Race

I won't be setting up and running the F@H holiday Race this year. My wife has had a number of health issues this past year and she needs my help 24/7 so most of the time I'm exhausted by the end of the day. Here's the link for last years race if anyone is interested in running this years race...

forums.anandtech.com

Current team peaks:
Daily Peak (1/25/2024):
1,666,901,932 points

Week Peak (week of 1/21/24)
7,654,446,4125 points

Monthly Peak (Month of Jan 2024)
26,898,347,599 points

Peak month active Users (1/23/2024):
83
Peak Active Users (April 19th, 2020):
1481

 

[cellarnoise]

Time keeps on ticking and a year goes by fast!

I'm down for participating!

 

[Markfw]

I am currently at 150 million ppd to make up for the 10 days I lost during the PG race. I will not be at that level for all of Jan.

 

[drnickriviera]

I'm down, got a little more firepower than last year

 

[StefanR5R]

during the past four weeks, the TeAm moved up or down in the ranks of the following projects:

1↑ at PrimeGrid and Radioactive@home,
1↓ at NumberFields@home, RakeSearch, SRBase.

At that time I missed that we also dropped from 13th to 14th at Folding@Home at some point between August 18th...August 25. AFAICT we were overtaken by China Folding@Home Power.

A strong January 2025 may not get us back into 13th perhaps, but closer. PC Master Race - PCMR would be the team to overtake. Our distance to them has been steadily shrinking for quite some time now.

Monthly Peak (Month of Jan 2024)
26,898,347,599 points

That won't be easy to beat. Here are snippets from the weekly stats:

Spoiler: many large numbers

Weekly DC Stats - 07JAN2024 Folding@Home overall position - 15 TeAm total for the week - 4,576,356,327 TeAm rank for weekly production - 6 __Credit/week _ UserName 1_______965,915,902__Mark_F_Williams 2_______887,101,117__Pokey_TA 3_______858,069,450__crashtech 4_______407,718,430__biodoc 5_______219,081,128__cellarnoise2 6_______211,860,145__mmonnin 7_______175,754,439__Endgame124 8_______154,836,815__ChelseaOilman_GRC… 9_______142,703,582__10esseetony 10______59,571,421___ericlp 11______57,760,745___King_of_Efland 12______48,898,367___NeoteriX 13______43,886,185___Hostile 14______41,646,850___drnickriviera_GRC… 15______34,822,781___IEC 16______25,456,599___fhonyotski 17______20,572,377___xcaliber219 18______15,436,631___thecoolnessrune 19______15,270,087___[TA]Assimilator1 20______14,692,727___JohnnyWong1213 21______14,385,581___Dexter 22______13,601,238___salvorhardin 23______12,995,122___LOL_Wut_Axel 24______11,191,824___Daniel_Trudeau 25______10,509,534___creed3020 26______10,116,638___johnnevermind 27______9,191,826____slowbones 28______9,021,229____Paul_McAvinney 29______7,970,907____2StepBlack 30______7,147,966____Mihryazd […] 68______4,403________Dr.N0 69______4,158________Quedex 70______3,402________78Pickup

Weekly DC Stats - 14JAN2024 Folding@Home overall position - 15 TeAm total for the week - 5,426,947,374 TeAm rank for weekly production - 5 __Credit/week _ UserName 1_______1,235,687,180_____Mark_F_Williams 2_______1,207,617,096_____crashtech 3_______1,038,992,493_____Pokey_TA 4_______392,258,888_______biodoc 5_______271,172,573_______cellarnoise2 6_______226,618,098_______ChelseaOilman_GRC… 7_______221,547,571_______mmonnin 8_______214,235,833_______Endgame124 9_______143,746,631_______10esseetony 10______63,355,026___ericlp 11______53,593,323___IEC 12______49,292,579___Hostile 13______48,879,796___King_of_Efland 14______44,946,699___drnickriviera_GRC… 15______27,860,868___fhonyotski 16______26,809,258___[TA]Assimilator1 17______21,155,358___Dexter 18______20,923,931___thecoolnessrune 19______15,028,356___salvorhardin 20______14,770,553___xcaliber219 21______13,462,372___LOL_Wut_Axel 22______13,440,700___creed3020 23______11,143,769___JohnnyWong1213 24______10,678,899___Daniel_Trudeau 25______9,712,416____NeoteriX 26______9,191,287____Paul_McAvinney 27______8,224,692____slowbones 28______8,024,919____dennc66 29______7,497,749____Mihryazd 30______5,866,189____RK […] 69______19,383_______RL 70______16,639_______jakeg 71______8,099________monkeyspooge

Weekly DC Stats - 21JAN2024 Folding@Home overall position - 14 TeAm total for the week - 5,739,648,315 TeAm rank for weekly production - 5 __Credit/week _ UserName 1_______1,352,806,000_____Mark_F_Williams 2_______1,311,945,978_____crashtech 3_______1,064,702,079_____Pokey_TA 4_______475,013,770_______biodoc 5_______264,043,528_______cellarnoise2 6_______241,301,172_______mmonnin 7_______212,006,403_______ChelseaOilman_GRC… 8_______173,217,002_______Endgame124 9_______166,318,054_______10esseetony 10______66,909,675___IEC 11______45,730,559___Hostile 12______37,735,075___King_of_Efland 13______33,010,208___Icecold 14______32,779,089___fhonyotski 15______26,345,452___[TA]Assimilator1 16______24,179,351___Dexter 17______21,352,072___Daniel_Trudeau 18______19,072,445___thecoolnessrune 19______18,052,314___xcaliber219 20______15,039,710___salvorhardin 21______13,852,398___LOL_Wut_Axel 22______13,286,644___slowbones 23______12,237,918___ericlp 24______10,720,031___dennc66 25______9,906,792____Gizmo23 26______9,747,626____Paul_McAvinney 27______9,510,179____creed3020 28______7,439,922____Mihryazd 29______4,762,352____drnickriviera_GRC… 30______4,657,093____CupCak3 […] 74______12,481_______jakeg 75______6,937________QuietDad 76______3,500________78Pickup

Weekly DC Stats - 28JAN2024 Folding@Home overall position - 13 TeAm total for the week - 7,654,446,412 TeAm rank for weekly production - 5 __Credit/week _ UserName 1_______1,439,303,663_____Mark_F_Williams 2_______1,382,765,412_____Skillz 3_______1,215,657,975_____crashtech 4_______1,014,428,870_____Pokey_TA 5_______399,949,200_______Icecold 6_______386,841,476_______biodoc 7_______363,372,352_______w_a_h 8_______266,809,914_______cellarnoise2 9_______244,047,727_______Endgame124 10______217,822,549_______mmonnin 11______200,077,206_______ChelseaOilman_GRC… 12______169,367,625_______10esseetony 13______56,004,767___King_of_Efland 14______45,393,152___Hostile 15______39,900,868___drnickriviera_GRC… 16______31,277,652___fhonyotski 17______26,343,688___ericlp 18______25,598,013___[TA]Assimilator1 19______20,669,229___xcaliber219 20______18,847,066___IEC 21______17,870,104___Dexter 22______16,458,799___Gizmo23 23______16,417,653___creed3020 24______14,744,205___Fardringle 25______14,642,012___salvorhardin 26______13,588,974___LOL_Wut_Axel 27______13,338,178___slowbones 28______11,330,487___Daniel_Trudeau 29______10,062,815___Paul_McAvinney 30______9,275,259____Ian-n-Steve_C. […] 80______9,261________39Th2TMX… 81______4,158________RL 82______4,158________Quedex

Weekly DC Stats - 04FEB2024 Folding@Home overall position - 13 TeAm total for the week - 5,117,082,731 TeAm rank for weekly production - 7 __Credit/week _ UserName 1_______1,223,070,240_____Mark_F_Williams 2_______966,013,797_______Skillz 3_______472,893,778_______Pokey_TA 4_______359,777,446_______biodoc 5_______345,708,356_______crashtech 6_______303,209,333_______Endgame124 7_______221,401,180_______ChelseaOilman_GRC… 8_______198,339,113_______w_a_h 9_______162,050,672_______cellarnoise2 10______114,572,087_______mmonnin 11______108,397,576_______10esseetony 12______68,717,565___drnickriviera_GRC… 13______57,724,394___King_of_Efland 14______57,484,056___IEC 15______44,303,392___Hostile 16______22,712,494___fhonyotski 17______19,318,460___xcaliber219 18______18,836,748___Gizmo23 19______16,916,984___Dexter 20______16,211,398___Icecold 21______15,700,755___[TA]Assimilator1 22______15,130,563___creed3020 23______14,780,682___Fardringle 24______13,469,855___LOL_Wut_Axel 25______13,283,956___slowbones 26______13,178,907___Oneto 27______10,661,154___Cliff_Taylor 28______10,560,326___dennc66 29______9,681,588____salvorhardin 30______7,975,905____Mihryazd […] 72______11,823_______QuietDad 73______4,802________ZipSpeed 74______3,402________Quedex

Yep, that's right, the TeAm popped up from 15th to 13th rank in January. There were more than 70 members showing up each week. But of the total of 28.5 B points of the five weeks which I pasted here, 15.6 B (55%) were made by three members. Four members managed to crack a speed barrier of 1 B per week.

 

[TennesseeTony]

I am nearing my PG goals, and have already started shifting to F@H , so I am in !

 

[minendo]

Decided to get back in to DC so count me in.

 

[Pokey]

Pokey is planning to be there, but I doubt I will exceed my production from last year. I just hope the word gets out to our other Folders.

 

[TennesseeTony]

16525 is showing 37M ppd on Win11, RTX 4090. Mtb-RNAP-Dynamics project
18228 32M ppd-cancer

Has Win11 overcome the Linux advantage?

 

[Pokey]

Has Win11 overcome the Linux advantage?

Someone else will have to answer that. I haven't run Windows for several years.

So…….. I had a couple of rigs with more than one Folding setup and I had gotten them so FUBAR’d I had to clean them up and re-install just one version. Now they are all uniformly set up and running. I assume we start the record run on Jan 1, correct?

 

[Endgame124]

Someone else will have to answer that. I haven't run Windows for several years.

So…….. I had a couple of rigs with more than one Folding setup and I had gotten them so FUBAR’d I had to clean them up and re-install just one version. Now they are all uniformly set up and running. I assume we start the record run on Jan 1, correct?

Yes, starting Jan 1. I’m just using EOC stats and they break up the numbers by month and week

 

[StefanR5R]

16525 is showing 37M ppd on Win11, RTX 4090. Mtb-RNAP-Dynamics project
18228 32M ppd-cancer

Has Win11 overcome the Linux advantage?

Hard to say because:
During the run of a workunit, performance is fluctuating a lot. That's in terms of power draw, GPU core clock, SM utilization, memory utilization, and consequently in FAHClient's momentary PPD estimation.

I ran several workunits so far and got four 18228 but no 16525 yet. The 18228s' credit estimations wiggled around in the 30M…31M range most of the time. This is in a dual GPU PC, each GPUs attached with PCIe v3 x8 to the CPU.

Instead of watching FAHClient's continuous PPD estimation, one could filter the client log for respective lines which show when workunits are received and sent and what the "Final credit estimate" is and compute PPD from there. But this figure includes the times needed for download and upload, i.e. includes internet connection performance [edit: and CPU and disk performance] alongside GPU performance.

 

[StefanR5R]

But wait, here are log contents which show almost only GPU performance:
Filter the log for lines with :Project: and :Average performance: and you get a computation performance figure in ns/day. The latter line is logged when the computation is done and before the result files are saved and uploaded.

From my four 18228s:

08:48:34:WU00:FS01:0x23:Project: 18228 (Run 47, Clone 8, Gen 5)
08:48:34:WU00:FS01:0x23:  cuda-device 1 specified
09:37:51:WU00:FS01:0x23:Average performance: 151.579 ns/day

09:38:08:WU02:FS01:0x23:Project: 18228 (Run 12, Clone 7, Gen 2)
09:38:08:WU02:FS01:0x23:  cuda-device 1 specified
10:27:26:WU02:FS01:0x23:Average performance: 148.966 ns/day

09:46:57:WU00:FS00:0x23:Project: 18228 (Run 471, Clone 7, Gen 78)
09:46:57:WU00:FS00:0x23:  cuda-device 0 specified
10:36:03:WU00:FS00:0x23:Average performance: 148.966 ns/day

10:27:42:WU01:FS01:0x23:Project: 18228 (Run 168, Clone 7, Gen 68)
10:27:42:WU01:FS01:0x23:  cuda-device 1 specified
11:16:57:WU01:FS01:0x23:Average performance: 151.579 ns/day

It seems as if there were two differently sized simulations in these four workunits. Alas I don't see an easy way to look up the base credit of each workunit, or something else which can tell such differences between workunits.

(For the above log snippets, I also grep'ed for cuda-device . specified on the suspicion that my two 4090s in this computer might perform differently. But as you can see, 148.966 ns/day at least was logged on either of the two GPUs.)

 

[Markfw]

But wait, here are log contents which show almost only GPU performance:
Filter the log for lines with :Project: and :Average performance: and you get a computation performance figure in ns/day. The latter line is logged when the computation is done and before the result files are saved and uploaded.

From my four 18228s:

08:48:34:WU00:FS01:0x23:Project: 18228 (Run 47, Clone 8, Gen 5)
08:48:34:WU00:FS01:0x23:  cuda-device 1 specified
09:37:51:WU00:FS01:0x23:Average performance: 151.579 ns/day

09:38:08:WU02:FS01:0x23:Project: 18228 (Run 12, Clone 7, Gen 2)
09:38:08:WU02:FS01:0x23:  cuda-device 1 specified
10:27:26:WU02:FS01:0x23:Average performance: 148.966 ns/day

09:46:57:WU00:FS00:0x23:Project: 18228 (Run 471, Clone 7, Gen 78)
09:46:57:WU00:FS00:0x23:  cuda-device 0 specified
10:36:03:WU00:FS00:0x23:Average performance: 148.966 ns/day

10:27:42:WU01:FS01:0x23:Project: 18228 (Run 168, Clone 7, Gen 68)
10:27:42:WU01:FS01:0x23:  cuda-device 1 specified
11:16:57:WU01:FS01:0x23:Average performance: 151.579 ns/day

It seems as if there were two differently sized simulations in these four workunits. Alas I don't see an easy way to look up the base credit of each workunit, or something else which can tell such differences between workunits.

(For the above log snippets, I also grep'ed for cuda-device . specified on the suspicion that my two 4090s in this computer might perform differently. But as you can see, 148.966 ns/day at least was logged on either of the two GPUs.)

Where is the log ? I can view it in fahcontrol, but that scrolls too fast to read it. I have 4080 and 4090's. What is the ppd you get credited for ? Right now I have 3 running, estimates at 18k, 19k and 21k. These are usually very close.

 

[mmonnin03]

In windows there is log.txt. The start menu folder for Folding@Home will have a Data Directory link

C:\ProgramData\FAHClient

 

[Markfw]

In windows there is log.txt. The start menu folder for Folding@Home will have a Data Directory link

C:\ProgramData\FAHClient

I found it. The path is c:\users\{username}\appdata\roaming\fahclient\logs\log-{date and time}.log

 

[Skillz]

Yes, starting Jan 1. I’m just using EOC stats and they break up the numbers by month and week

I thought the EOC stats was Sunday to Saturday. Thus wouldn't it be smarter to start Saturday night for a week tally? I mean you could just keep track of each individual day for 7 days, but seems EOC would make it easier if you started when their week starts.

 

[Endgame124]

I thought the EOC stats was Sunday to Saturday. Thus wouldn't it be smarter to start Saturday night for a week tally? I mean you could just keep track of each individual day for 7 days, but seems EOC would make it easier if you started when their week starts.

Their next week will start 12/29. My plan was to just deduct 12/29 though 12/31 in the first week numbers, and then just use the EoC weekly number for the remaining weeks. The goal of the challenge is to break the Jan points total, so it doesn’t really bother me if everyone wants 12/29 though 12/31 to count for a weekly record if people really want to get rolling now.

 

[StefanR5R]

c:\users\{username}\appdata\roaming\fahclient\logs\log-{date and time}.log

Yes, according to the F@H FAQ, the client's data directory is C:\Users\{username}\AppData\Roaming\FAHClient . In there, the log file to which the client actively writes to should be called log.txt, and periodically archived log files should be in logs\log-{date and time}.txt .

On Linux, the client data directory with logs et cetera is usually called /var/lib/fahclient/.

 

[mmonnin03]

Looks like Ubuntu 20.04 is too old for core24. After 12 min of FahCore returned: FAILED_2 (1 = 0x1) another core23 WU downloaded and continued.

FAH Core 24 Fails Xubuntu 18.04 extended - Folding Forum

 

[Ken g6]

Looks like Ubuntu 20.04 is too old for core24. After 12 min of FahCore returned: FAILED_2 (1 = 0x1) another core23 WU downloaded and continued.

FAH Core 24 Fails Xubuntu 18.04 extended - Folding Forum

On the latest Ubuntu or Mint, Python 2 is removed. The Advanced Control client was written in Python 2. So I'm using the V8 FAH client with the web control right now, and looking into how one might control it otherwise with websockets.

 

[TennesseeTony]

Just spent the last 3 hours dual-booting Mint 22 Cinnamon on 4 machines. The 2080Ti was detected and ran fine. ONE of 3 4090's was detected and ran fine, the other two were disabled. NONE of the 4 would start BOINC. Back to Windows.

 

[Ken g6]

Hey, @geecee, I just passed you in F@H! Are you going to do anything about it?

 

[Ken g6]

On the latest Ubuntu or Mint, Python 2 is removed. The Advanced Control client was written in Python 2. So I'm using the V8 FAH client with the web control right now, and looking into how one might control it otherwise with websockets.

The FAH V8 client is working well enough for me, for now, except that I missed estimates for individual WU points. So I made a user script to add them!

Folding@Home V8.3 show WU points

This script calculates and displays a F@H V8 WU's points.

greasyfork.org

 

[TennesseeTony]

The greasyfork screenshot looks identical to the standard V8 webpage, no additional information that I can see?

And I don't mean to rain on your parade, but this has a lot of fun info: https://webclient.lar.systems/8_3_5/

Edit: Breath-Weapon might need a tictac?