7 replies [Last post]
jeff101's picture
User offline. Last seen 4 hours 39 min ago. Offline
Joined: 04/20/2012
Groups: Go Science

I am curious what would happen if you compared the rankings
of Foldit players on each platform (Windows vs Mac vs Linux).
Do players on one platform tend to find better solutions than others?
If there is a bias, could it be just because there are more players
on one platform than on another? Do you even know the distribution
of platforms players use?


Joined: 10/23/2014
Groups: Contenders

Could you clarify what you are trying to advance? I would think the underlying computer algorithms on the three platforms are essentially the same and there are insignificant differences in the final foldit application on each platform.

There are many aspects to better solutions unrelated to the platform. How long the player has played foldit, how much time is spent on a puzzle, how many designs are tried, how many instances are run, how many computers are used, the speed of the computer(s), is there a history of similar reference puzzles, the sharing of solutions and advice by your group, and so on. I would think the platform is one of the less important aspects on finding better solutions.

jeff101's picture
User offline. Last seen 4 hours 39 min ago. Offline
Joined: 04/20/2012
Groups: Go Science
Why ask?

The question just came to me the other day,
and I don't recall anyone else discussing it.

If there did seem to be an advantage to using
a certain platform for folding, I think more
players would want to use that platform.
Maybe one platform has more/faster CPU's for
the same price. Maybe one platform usually
comes with higher resolution and faster
graphics capabilities. Maybe one platform
usually has mice with 2 buttons and a finger
wheel. Maybe one platform is more likely to
stay in contact with the server when the
server misbehaves. Maybe one platform is
more responsive when you want to hand-fold.
Maybe one platform crashes less often in the
middle of a long GAB run or when running a
recipe you wanted to run all night. Crash
recovery and restarting clients can be very
distracting and time-consuming.

There are also some stereotypes about users
of the different platforms. For example,
I would expect Linux users to have more
programming experience and Mac users
to be more artistic.

Joined: 09/24/2012
Groups: Go Science
Some comparison here

We compared (only computers) some time ago here "all other thinks equal":

Joined: 09/24/2012
Groups: Go Science
It appears that:

-the OS or platform does not seem to be determinant;
-the Processor plays a role (as well of the exact date version of it - e.g. there are many different "i7" processors)
-the number of cores plays a role for the number of clients we can run in parallel
-the RAM is very important for the number of clients we can run in parallel (8 Gb is much better than 4 Gb); if you get too close to the physical memory limit, your computer dramatically slows down
-the video RAM or video card seems to play a role, but minimizing the windows is much more important (20 % power gained at least)
-the fan (refreshment) seem to play a role (laptops tend to allow less parallel clients than desktops)

Personally, on a Mac I didn't find how to optimize the number of clients ("activity monitor" does not say explicitly if we are over of bellow the limit, in Windows, you can see that you use x % of the CPU). On a Mac, I know that I'm over the limit when everything slows down (even the mouse on the screen doesn't react normally anymore).

My best "Foldit" computer is still a 4 years old (family) intel i7 Windows tower designed for video games (with a good graphic card, 8 Gb ram etc). It's almost 2 times faster and it can run 8/3 more clients than a recent (3 years old) intel i7 Mac laptop (that I paid twice the price of the family one ...)
Acer Predator G3600
Processor: Intel Core i7-2600 CPU@3.40GHz 3.40GHz
64 bits
RAM 8 Go
Graphics: NVIDIA geforce GT440
OS: Windows 7 familial edition, Service Pack 1

Joined: 01/18/2017
Groups: None
I accept it.

maybe you can find but at the end the foldit will win only. It's the matter of quality and consistency. IF you find someone doing this then it can create some errors for you otherwise you should take care of everything in order to bring the saturation.

Joined: 03/05/2015
Groups: Gargleblasters
Who is this user?

??? Suspicious post ???

Joined: 03/05/2015
Groups: Gargleblasters
In puzzles that don't break due to other issues...even, based on

speed and configuration of their computer... so I'd say it's pretty even as Bruno points out in his reference to the test recipe. My tests were done with a lot of clients open (I wanted to test it the way I usually work).

But beyond the speed and configuration of the machine, the human variable (and those game bug-induced variables) seems like it would make it tough to get an accurate read on this.

What recipes does each player use? How much time do they spend per day playing? How much related science does the player understand and how does that scientific knowledge affect what they do? Why do they play the game in the first place? How long have they been playing? Do they play only solo or group? Etc. etc...

It gets complicated very, very quickly.


Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, Boehringer Ingelheim, RosettaCommons