Percentage of F@H participants who are 'overclockers'

[rocketmama]

Hi everyone,

I have been researching F@H as part of my PhD on online citizen science projects. Some of you very kindly took the time to complete my survey last year and answered some follow-up questions. A recent estimate of the number of F@H participants is somewhere in the region of 100 000. I would be really interested in getting an idea of what proportion of participants are overclockers / hardware enthusiasts. I realise that it's going to be difficult to know this with any degree of certainty, but I wondered if anybody out there had any idea, or was willing to make an estimate.

Many thanks for any help,

Vickie Curtis
The Open University, UK

[P5-133XL]

A very large number

[PantherX]

Welcome to the F@H Forum rocketmama,

To get a rough idea, if a donor is member of a Team which is related to Hardware Reviewing, there is a good chance that they have overclocked their system. You can get the list of Teams from this site (https://folding.stanford.edu/home/teams-stats/).

Moreover, reputed motherboard companies may have a BIOS feature called "Automatic Overclocking" which can be easily enabled by the user even though they may not know how to manually overclock their system. Thus, the barrier to overclock their system has greatly been reduced.

[billford]

And some GPUs are factory-overclocked.

In my case, this was a significant factor in the choice.

[rocketmama]

Just over 400 people responded to my survey.

Based on their comments and team affiliation, I estimated that 70-80% of the survey sample were overclockers. However, enthusiasts were more likely to respond to the survey because of where it was posted and how it was communicated, so I doubt that the percentage of overclockers in the total F@H population is as high.

I'm wondering if it is more like 5-10% of participants (5-10K individuals).

Rocketmama

[bruce]

I doubt there's a good answer to your question. There's no doubt that there are a number of non-enthusiast/non-overclocker who fold, but almost by definition, you won't be able to contact them because they didn't see your survey and they don't visit this site. It's like the government trying to figure out how many (perhaps homeless) people were not counted in the last official census.

I wish you luck.

[rocketmama]

I suspect you're probably right Bruce.
I thought those within the various OC communities might have a sense of their numbers.

[Grandpa_01]

I suspect you're probably right Bruce.
I thought those within the various OC communities might have a sense of their numbers.

Just curious but where do you get the 100,000 number from, The 3rd party stats sites show in any given 24 hr period there are between 10,000 and 13,000 individuals truing in completed WU's and in any given 7 day period there are between 33,000 and 34,000 individuals turning in completed work. The 180,000 number the home page shows is CPU's 180000 / 33000 = 5.45 CPU's per active member in a 50 day period of time. When you take into consideration that every GPU and individual computer / fah client = 1 that number is very reasonable.

I think the % of OCed and enthusiast is actually pretty high as far as the private donor world goes. KaKao stats shows there are 34,537 active members in the last 7 days, http://kakaostats.com/ if you look at that list of teams and visit there web sites you will soon realize a large % of the private folding community are over clockers and run multiple clients. The enthusiast community most likely provides the lions share of work done by the private individual donors. Now if you are talking about corporate donors that would be a totally different story. I believe Vijay recently refereed to a corporate donor who had just fished up a run and it was 40,000 cpu's by itself which is also counted in the 180,000 number although we do not know how many clients those 40,000 cpu's were running (I would imagine quite a few) so that is an unknown as far as how many it would have added to the 180,000 number goes.

[mdk777]

First, there are many definitions and gradations of OC.
Many MB and Graphics cards OC automatically, or with a simple one time setting.

If I build a computer for my grandmother, setting with a very mild OC...Is she considered an OC ?

Others run some tuning adjustments on buying or building a computer, and then "set it and Forget it."

Does the custom memory settings and SSD optimizations on my LENOVO corporate Laptop workstation(W530) count as OC?

Finally, there are those who actively adjust and benchmark on a regular basis, or change settings in response to a specific WU.

I would say that the first two categories might be the 80% number you mention, and the last closer to the 5-10% number.

However, given the complexity of definition, and the extreme variations in participation rates reported overall...I don't think there is any way to accurate characterize the big picture from your limited survey sample.

[Jesse_V]

This is not directly related to your question, but on a similar note there has been a shift with F@h wherein the WUs and cores have been getting more and more sensitive to overclocks. We've seen this particularly with FahCore 17. I can't speak for the future of course, but it does seem that those F@h overclockers will need to really ensure that their systems are stable.