Number of active folders going down?

[7im]

I can fully understand the Linux count going down, Its where BA units reside and a lot of heavy hitters have already shut down

Funny thing is the Linux TFLOPS have not gone down, nor has the active client count. The first set of numbers is from November 1, 2013. The second set of numbers was taken last week. Same TFLOPS, active clients have gone up, active cores have gone down. Decide for yourself.

Code: Select all

Native x86 CPUs Cores Total CPUs
14	14	4512	20226	45204
14	14	4586	17426	48451

[EXT64]

Yeah, something is wrong with Active Cores vs Active CPUs for the GPUs. I have 2 AMD GPUs in one box, so I know the two should not be equal. That, or for fermi the cores should equal cpus.

[bruce]

There has been some mis-reporting of ATI vs. NV GPUs. That, too, has apparently been corrected.

As far as over-counting clients, remember that if you uninstall and reinstall, that same machine is counted twice for up to 50 days. The servers are not notified that client XXXXXXXXXXXXXXXX has been uninstalled, only that client YYYYYYYYYYYYYYYY has just returned a result.

[netblazer]

There has been some mis-reporting of ATI vs. NV GPUs. That, too, has apparently been corrected.

As far as over-counting clients, remember that if you uninstall and reinstall, that same machine is counted twice for up to 50 days. The servers are not notified that client XXXXXXXXXXXXXXXX has been uninstalled, only that client YYYYYYYYYYYYYYYY has just returned a result.

Fine, but how many people actually uninstalled 7.3.6 and tried both betas and kept posting wus during all that time??? That can't possibly be more than 5000?? people out of 50K?? active members. On the lot? a rather statistically insignificant amount (no offense intended).

[7im]

No one is trying to explain the change with a single explanation, just that any small change in numbers, up or down, has multiple causes. Not all of them because of a perceived sky falling. More often than not it's just a pine cone.

[P5-133XL]

Today, I noticed that Linux TFLOPS had dropped from 11 to 9. That is the first indication of bigadv abandoning folding that I've seen.

[Joe_H]

Maybe, or some of the ATI stats were being misapplied to the Linux category. I did notice that a similar drop started on the OS X TFLOPS figure shortly after the PG announcement of the stats fix. It had risen to 19 or 20 and dropped to 13 as the ATI number rose.

[Rattledagger]

Funny thing is the Linux TFLOPS have not gone down, nor has the active client count. The first set of numbers is from November 1, 2013. The second set of numbers was taken last week. Same TFLOPS, active clients have gone up, active cores have gone down. Decide for yourself.

Code: Select all

Native x86 CPUs Cores Total CPUs
14	14	4512	20226	45204
14	14	4586	17426	48451

AFAIK multiple teams has some kind of FAH-Christmas-race, having an extra spike in "normal" 2-core to 12-core computers at the same time some other users shuts-down their BA-computers are a possibility that can explain the small increase in active cpu's but a large decrease in #cores. Since where's a 50-day delay from shutting-down until this has any effect on the stats-page, the numbers for 10.02.2014 was around 22.12.2013 so likely during any kind of Christmas-race. Todays numbers on the other hand means around 31.12.2013 and depending on race-end this can be afterwards. For the last two stats-updates, the numbers was:

Code: Select all

Linux 	9 	9 	2878	10030	46946
Linux 	9 	9 	2875	10019	46943

Well, to me a halving of active cores since 01.11.2013 and a 36% drop in active cpu's is a marked decrease for Linux.

The possible reasons for this drop are:
1: There's a problem with the Google-worksheet so the old numbers are incorrect.
2: The stats for 10.02.2014 was incorrect meaning can't really compare to old numbers.
3: The bug-fix for Amd-GPU's broke the stats for Linux (and from the look of things Windows & Mac).
4: Users are turning-off their FAH cpu-clients.

Seeing how the "Total CPUs" for Linux decreased with 3 between the two latest stats-updates, the stats are atleast partially broken but this doesn't neccessarily means the other numbers also are broken.

For the other stats, Amd has a marked increase after the bug-fix. The two Nvidia-types also has a marked increase, maybe these also has been under-counted, or it's possible all GPU-numbers are basically "pre-fix" + "clients returned anything after the fix", so can continue to increase until hit the 50-day-limit and would afterwards get a large drop for the Nvidia-types (No idea if GPU also uses a 50-day limit or not, previous it was shorter). Amd-numbers was so low before the fix so any drop due to a few clients being counted twice will be minimal.

For CPU-clients on the other hand it looks more like "pre-fix" + "fails to count clients returned anything after the fix" meaning these stats are broken. Or, users are choosing #4 instead, and is turning-off their FAH cpu-clients.

[7im]

... Snip

Or, users are choosing #4 instead, and is turning-off their FAH cpu-clients.

I, like many others, have chosen to take a 2000 points per day Dual core Computer, turned off the CPU client, and added a GPU or two, making between 100K and 200K points per day.

[Nathan_P]

20-30 4p machines and around 40-50 2p machines gone that I am aware of. Not sure of the exact dates but the effect should start coming through in the next week or so since they started pulling the plug around xmas/new year

[7im]

The revised BA announcement was Jan 15th. Let's see where we are on March 7th (+51 days).

[Rattledagger]

I, like many others, have chosen to take a 2000 points per day Dual core Computer, turned off the CPU client, and added a GPU or two, making between 100K and 200K points per day.

Yes this definitely is a possibility, but it's still a little strange the users responsible for shutting-down 24210 cpu-clients between 22.12.2013 and 31.12.2013 all waited until 10.02.2014 before adding 14964 GPU-clients (not counting Amd) and the majorify of these was added the last 48 hours.

[7im]

I, like many others, have chosen to take a 2000 points per day Dual core Computer, turned off the CPU client, and added a GPU or two, making between 100K and 200K points per day.

Yes this definitely is a possibility, but it's still a little strange the users responsible for shutting-down 24210 cpu-clients between 22.12.2013 and 31.12.2013 all waited until 10.02.2014 before adding 14964 GPU-clients (not counting Amd) and the majorify of these was added the last 48 hours.

With all the stats reporting adjustments in the last few years because of client changes, and client type changes, and fahcore changes, I can pick another two dates and show just the opposite viewpoint.

Stuff is changing. It will change again. Let's see what shakes out in the next few months.

[Risme]

Funny thing is the Linux TFLOPS have not gone down, nor has the active client count. The first set of numbers is from November 1, 2013. The second set of numbers was taken last week. Same TFLOPS, active clients have gone up, active cores have gone down. Decide for yourself.

Code: Select all

Native x86 CPUs Cores Total CPUs
14	14	4512	20226	45204
14	14	4586	17426	48451

AFAIK multiple teams has some kind of FAH-Christmas-race, having an extra spike in "normal" 2-core to 12-core computers at the same time some other users shuts-down their BA-computers are a possibility that can explain the small increase in active cpu's but a large decrease in #cores. Since where's a 50-day delay from shutting-down until this has any effect on the stats-page, the numbers for 10.02.2014 was around 22.12.2013 so likely during any kind of Christmas-race. Todays numbers on the other hand means around 31.12.2013 and depending on race-end this can be afterwards. For the last two stats-updates, the numbers was:

Code: Select all

Linux 	9 	9 	2878	10030	46946
Linux 	9 	9 	2875	10019	46943

Well, to me a halving of active cores since 01.11.2013 and a 36% drop in active cpu's is a marked decrease for Linux.

The possible reasons for this drop are:
1: There's a problem with the Google-worksheet so the old numbers are incorrect.
2: The stats for 10.02.2014 was incorrect meaning can't really compare to old numbers.
3: The bug-fix for Amd-GPU's broke the stats for Linux (and from the look of things Windows & Mac).
4: Users are turning-off their FAH cpu-clients.

Seeing how the "Total CPUs" for Linux decreased with 3 between the two latest stats-updates, the stats are atleast partially broken but this doesn't neccessarily means the other numbers also are broken.

For the other stats, Amd has a marked increase after the bug-fix. The two Nvidia-types also has a marked increase, maybe these also has been under-counted, or it's possible all GPU-numbers are basically "pre-fix" + "clients returned anything after the fix", so can continue to increase until hit the 50-day-limit and would afterwards get a large drop for the Nvidia-types (No idea if GPU also uses a 50-day limit or not, previous it was shorter). Amd-numbers was so low before the fix so any drop due to a few clients being counted twice will be minimal.

For CPU-clients on the other hand it looks more like "pre-fix" + "fails to count clients returned anything after the fix" meaning these stats are broken. Or, users are choosing #4 instead, and is turning-off their FAH cpu-clients.

1. What is the problem more specifically that you're referring to?
2. How are/were the stats in the spreadsheet incorrect for 10.02.2014? I checked the data from an older revision of the spreadsheet and saw no changes in values for that row. If there's a problem with something in the spreadsheet, please leave a comment in the note column to specify the problem. Please note that whatever osstats2 reports is out of my control; the spreadsheet is only meant to be used as a record of the data osstats2 has reported historically.

I'd like to know since i maintain that document, although i encourage people to help me maintain it by checking if data has been entered each day since i forget to do that sometimes. I think i'll start taking a screenshot of http://fah-web.stanford.edu/cgi-bin/mai ... e=osstats2 for each entry starting next month, so there's some evidence to verify that the data entries match what osstats2 is reporting during the time the data was extracted.

[Rattledagger]

1. What is the problem more specifically that you're referring to?
2. How are/were the stats in the spreadsheet incorrect for 10.02.2014?

I'm not aware where being any problem with the spreadsheet, I just included this option since it's one of the possibilities what would explain the current numbers of active cpu-clients are lower and I wanted to include all possibilities, even some of them is very unlikely to have happened.

The FAH-stats the spreadsheet is based upon on the other hand looks suspicious, taking a look the current numbers are:

Code: Select all

OS Type 	Native TFLOPS*	x86 TFLOPS* 	Active CPUs	Active Cores	Total CPUs
Windows 	484	484	111272	187163	5283146
Mac OS X 	12	12	4387	24522	34507
Linux 	8 	8 	2626	9271	46716
ATI GPU 	374 	789 	2633	2633	363856
NVIDIA GPU 	1650 	3482 	8728	8728	304037
NVIDIA Fermi GPU 	10360 	21860 	30561	67596	347366
Total 	12888	26635	160207	299913	6379628


1723798 people have non-anonymously contributed to Folding@home.
Table last updated at Thu, 20 Feb 2014 01:17:17 

Linux is continuing losing "Total CPUs", it's lost an additional 1735 CPUs and things are similar for Windows & Mac. As for GPU's, the so-called "Active Cores" is clearly bogus, either for Fermi or for Amd, since EXT64 has a dual-AMD-box but this is not showing-up in the stats.

BTW, the stats-page should also include whatever time-zone it's using.