Number of clients able to be run on NC

Case number:699969-997221
Topic:General
Opened by:spmm
Status:Open
Type:Suggestion
Opened on:Monday, March 10, 2014 - 03:16
Last modified:Wednesday, April 11, 2018 - 14:48

Several people in veteran chat have just been kind enough to specify the number of clients they used to be able to run and the change they have experienced with NC. Note these numbers are for 'normal' puzzles.

In most cases the change is a big reduction, from 7 to 3, 4 to 1 and so on. This makes it very difficult for players on lower spec machines to experiment and play.

Even folders on better spec devices have experienced overheating and higher resource consumption so are reducing the number of clients they can run.

Some of the players have stated that they feel they won't be able to complete well in CASP X, CASP 9 required a number of templates to be examined and played, combined with a fast turnaround of puzzles.

There is concern from a number of players that they are damaging their machines and are unable to experiment and contribute as much as they would like, this is frustrating.

Design puzzles especially the highly constrained ones are almost unplayable on lower spec devices and reduce the number of usable clients on other devices.

Just passing the message on as it is concern for everyone.
We all know that a better device doesn't make you a better folder, but it is becoming much more difficult for standard devices to play foldit and even high spec consumer devices are feeling the heat.
The latest devprev update has made this worse.

(Mon, 03/10/2014 - 03:16  |  49 comments)


MurloW's picture
User offline. Last seen 8 weeks 1 day ago. Offline
Joined: 11/21/2012

I've been dreading every new update for a long while.
When I first started folding, I could easily run 5~7 clients and still run some light programs (like web browsing); now, I can barely run 3 clients if I still want to use my machine for anything other than folding.
And I don't use devprev releases. So.. I'll only be able to run 1 client and a browser in a few months? I wouldn't be surprised.
I must say though, Foldit has been acting weird (hogging and absorbing more memory than it ought to) since before Newchapter. Sometime around the summer months is when my number of clients went on the decline; and NC didn't change much in this respect, wiggle has slowed down tremendously though.

There have been feedbacks before, specifically about number of clients and memory hogging. (the search function is kinda useless, I couldn't find them)
Reminds me of a peculiar fact that has been reported before, by me as well; every newly opened client requires more memory than the previous one, even when sitting idle on the exact same puzzle as other clients.
This was worse before NC; sometimes up to 100MB more, exact same puzzle, view options, undo graph, etc.
At the moment, it seems like every next client wants 20~40MB more than the one before.

Joined: 09/21/2011

I set up this pc in Jan. 1013. It ran 4 clients bone silent. I could run 3 regardless of type ( mutate dimer, etc.) bone quiet. I think it was around Sept. 2013 that my fans started kicking and aften reaching full speed even with just 2 clients running. Didn't notice a slowdown, but I was clearly using way more juice. NC gives me the same fan speeds, but now I get slightly slower speeds even with 2 clients.A 3rd client and it slows considerably.

MurloW's picture
User offline. Last seen 8 weeks 1 day ago. Offline
Joined: 11/21/2012

.
Something changed between June and Oct. 2013 that was never addressed or probably even acknowledged by devs that made the clients heavier to run. Newchapter probably inflated this issue into slower speeds as well, resulting in more people noticing.

How many more complaints about needing high end specs to fold do you guys need before you will try to do something about it; instead of working on s*** like Leap and the remote control stuff?

I understand that Newchapter was essential and a priority to keep up with CASP and Rosetta; but it won't help much if most players won't even be able to play the casp puzzles comfortably. We won't fare so well if veterans can only run half the clients they are used to; regardless of how accurate wiggle may be.

Joined: 04/15/2012
Groups: Beta Folders

"...instead of working on s*** like Leap and the remote control stuff"

Like in almost any game development team, I'm fairly certain they have different teams working on different things.

MurloW's picture
User offline. Last seen 8 weeks 1 day ago. Offline
Joined: 11/21/2012

I should have taken a breather before posting that, might have found a better choice of words.
Please excuse my language in the above post, it was uncalled for; even though I feel like scalding the devs in every single thing I post...

Aside from that, ofcourse they have different teams for different aspects of the software; that is beside the point. The point is that there are resources being spent on aspects of Foldit that are completely irrelevant for a massive majority of players. If they can afford to spend those resources anywhere, they can afford to (and should) spend them on playability.

Joined: 04/19/2009

I had 2 clients running overnight, and when I checked this morning my CPU fan was under 1k rpm. I opened a third client on the 853 puzzle (not a mutate which always seems to crank the fan), and now have the CPU fan over 2300 rpm.

I used to run 4 clients easily with the fan only cranking up if one was a mutate puzzle (occasionally 5) - now I can only obviously run 2 without stressing my not quite 2 years old, then top of the line mac desktop :(

I had wanted to experiment with the Auto wiggle in the DP update, but there is no way for me to do so if I can only run 2 clients without stressing my computer.

And just a note - the mutate puzzles with the filters and whatall are the worst offenders since all the conditions have been introduced - even before NC, but worse now.

bertro's picture
User offline. Last seen 7 weeks 4 days ago. Offline
Joined: 05/02/2011
Groups: Beta Folders

For information: running all 6 clients using latest devprev.

More memory is used than before, about same CPU usage and temperature 65C or lower (as usual).

bertro's picture
User offline. Last seen 7 weeks 4 days ago. Offline
Joined: 05/02/2011
Groups: Beta Folders

and now with 7 clients running: 1x853, 2x854, 1x855 and 3x856.

Joined: 09/19/2011
Groups: Beta Folders

I'm thinking you are running a desktop. Auntdeen and I are running laptops. Mine is a workstation class and I have a lot of overheating issues.

bertro's picture
User offline. Last seen 7 weeks 4 days ago. Offline
Joined: 05/02/2011
Groups: Beta Folders

Yes a desktop.

spmm's picture
User offline. Last seen 41 weeks 2 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

Well we probably need something like these new Intel E7-8880L chips for:

'Customers who want fifteen cores running at a low clock speed and emitting less heat and a midrange price can get the E7-8880L chip.'

Which are x86 based, introduced 18 Feb 2014 - $5729 each; especially as they can address 1536 GB of physical (RAM) memory (not included in the price, nor is anything to put them in :D)

http://www.cpu-world.com/CPUs/Xeon/Intel-Xeon%20E7-8880L%20v2.html

jflat06's picture
User offline. Last seen 1 hour 11 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

While it's entirely possible that Foldit is taking up more resources, keeping the computer cool is the job of the operating system and hardware - not of softare.

Computer processors are designed to stay stable under full load (100% CPU usage for sustained periods of time). If your computer is overheating under full load, it is due to inadequate cooling.

Programs do not know your CPU temperature, so they can't know how much CPU is going to cause your computer to overheat. Nor does it make sense for every program to try and throttle itself to prevent overheating. That is why computers are built with cooling capable of supporting the given processor under full load.

You can test overheating issues independent from Foldit with programs like superpi (http://www.superpi.net/). You can run superpi for extended periods of time to verify your system's overheating and stability issues.

Some reasons for overheating/stability issues might be dust that has built upside the computer that block the airflow. Clearing out this dust can drastically improve your computer's cooling. Some poorly designed computers just have inadequate cooling from the start (more common in laptops).

gitwut's picture
User offline. Last seen 41 weeks 6 days ago. Offline
Joined: 05/18/2012
Groups: Contenders

jflat06,

I believe most everyone knows that the OS/hardware are responsible for keeping the computer cooled properly. I would hazard a guess that most computers being marketed today weren't designed with the idea of running multiple Foldit clients efficiently. However, it is our experience that since NC, we are having overheating problems ONLY when Foldit is running AND it is happening more frequently with fewer clients running.

My desktop computer at home is complaining about overheating with only two clients running. One is 857 (handfolding) and the other is 860 (Monomer design). I used to be able to run 4 or 5 clients before getting warnings. This is a quad core i7 with 9GB RAM and a video card with dedicated memory.

I rarely run more than 2 clients on any of the four computers I use now. I run 7 clients now, whereas I used to run about 14. I start evos later (less than 2 days remaining) than I used to (as soon as one was posted) and run alternate solos to the maximum 7 clients. When it's time to start an evo, I close the poorest performing solo to keep my max at 7.

Joined: 04/19/2009

Foldit IS taking up more resources, period.

All of us didn't have our computers go dusty or obsolete the day that NC was introduced ;)

And what gitwut said.

jflat06's picture
User offline. Last seen 1 hour 11 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

I'm not arguing against that. Rosetta definitely uses more resources since we merged. I'm just pointing out that we cannot control overheating issues - that is a hardware problem, not a software problem.

Joined: 04/19/2009

Then all of us but the dozen or so with the top of the line, newest computers need to accept that the number of clients we can play now is the best we can do?

Sigh.

There are less of us, the logged in vet has dropped from 100-125 before NC to 65 - 85. Many of us have had to cut down by half the number of clients we can run. As a very loose estimate, let's assume that players were running an average of 4 clients before NC, so minimal 400 clients were running. Let's assume that most have had to drop down to half the number they used to run - so now there's 80 x 2 (average) = 160 clients running. Depending on the real averages, that's one third to one half the clients that used to run before NC.

IMHO, this is the single worst "bug" affecting us now. Less clients means less options, both for testing and learning now - less clients means that many solutions will be left on the table during CASP.

Unfortunately then, it seems that the devs have vastly overestimated the players resources in terms of computing power, or the finances we have to sink into foldit.

:(

MurloW's picture
User offline. Last seen 8 weeks 1 day ago. Offline
Joined: 11/21/2012

" IMHO, this is the single worst "bug" affecting us now. Less clients means less options, both for testing and learning now - less clients means that many solutions will be left on the table during CASP. "

This. A thousand hail AuntD's.

Players complaining about overheating issues are not saying that the software has changed into a demon burning up their hardware from the inside; they are complaining because their systems should not be overheating (aka being thís stressed) while running what they are, be that 2 or 47 clients.

You may not be able to control overheating issues, ofcourse you can't; but you should be putting effort into slimming the software down. Is it still not clear that this is not only due to NC? At some point foldit became heavier to run, resulting in less clients for those who were used to running a bunch of them; then NC came and this issue became even more dramatic, causing loads of players to just kinda forfeit. I say forfeit because actually enjoying this game has become somewhat of an uphill battle; resulting in the massive drop of active players.

Joined: 04/19/2009

And if this is the case - then it's time for foldit to post minimum system requirements.

frood66's picture
User offline. Last seen 5 hours 37 min ago. Offline
Joined: 09/20/2011
Groups: Marvin's bunch

I've just read the above. Jflat...those are about the most ignorant, self serving statements I have heard put forward by the devs to date. such a shame.

Joined: 09/21/2011
Groups: Void Crushers

This is about the worst feedback I have read since I joined Foldit.
Temperature of core is the business of the operating system. Period. Exclamtion point.

Joined: 09/19/2011
Groups: Beta Folders

If game "upgrades" make it too difficult for the average person to play it becomes the business of the developers if they want players.

I run a workstation class laptop and am now down to only two clients due to overheating. My CPU and RAM usage is only 28% and 30%, but, my temperature is still in the high 70's. Whatever changes the developers are making to cause this is something they need to look at closely. Foldit is the only application that makes my computer run hot. And, it never used to run hot before the introduction of the filters. Since new chapter it's taken it higher again in temp.

Higher temps are clearly due to a change in the foldit client since this wasn't an issue before the introduction of filters.

New instructions? Optimization for speed? Whatever they have done, the majority of players are having a problem. This feedback is making it obvious.

Joined: 04/15/2012
Groups: Beta Folders

I'd like to again repeat that both sides are arguing different things, and aren't necessarily disagreeing.

YES, the client has become heavier, and this should be addressed.

NO, the Foldit client alone can't cause your computer to overheat, no matter how heavy it is.

Joined: 09/19/2011
Groups: Beta Folders

You are wrong asdawnbreaks... the foldit client is the only thing that makes my computer overheat. And, it does it for several different computers. And, I'm not the only one experiencing this.

spmm's picture
User offline. Last seen 41 weeks 2 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

Would you like to put some expertise or internationally recognised qualifications behind the statement that foldit can't consume resources to the extent that it will cause overheating or increased cooling resource expenditure? Or perhaps you don't actually fold very much?

Joined: 04/15/2012
Groups: Beta Folders

NOT what I'm saying. I'm saying that Foldit IS consuming so much resources that it can cause the computer to overheat. But as Jeff said, if your firmware and OS are correct (and you have your fans cleaned out) you shouldn't have a problem with overheating, as it should take care of the issue before it does.

spmm's picture
User offline. Last seen 41 weeks 2 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

Frood we are all ostensibly playing a game to help science, and by derivation the scientists who provide the puzzles, self serving is no less than can be expected, hardly expect scientists to not push the envelope, it is up to the folders to let the scientists know if there are problems which is what is happening in this thread.

Angus's picture
User offline. Last seen 2 hours 12 min ago. Offline
Joined: 06/04/2008
Groups: Beta Folders

I expect this will become a desktop-only application. Most laptops can't run flat out all the time, and even the top of the line workstation-class ones have trouble with cooling.

spmm's picture
User offline. Last seen 41 weeks 2 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

Agree Angus, I almost burnt out a high end laptop during CASP 9, with added mutating design puzzles thrown in. Entirely my own fault of course.
I agree that foldit is not responsible for looking after my hardware although I fully expect not to get any malicious code.

I assumed the highend laptop would protect itself, which it now does by shutting down, it survived and improved for nearly 9 months after one professional clean and re-application of thermal paste. The second clean and re-paste did nothing and it slowed right down again, still ok for email and browsing.

At least these posts will indicate to folders who may not be aware of it that prolonged heat over many days will seriously damage their computers especially laptops.

If you cruise the Overclockers folding forums, players who run 24x7 are all running water cooled desktops or workstations because you get far more grunt and associated cooling for a lot less money than a laptop.

spmm's picture
User offline. Last seen 41 weeks 2 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

Software can easily damage hardware:

This was a recent firmware attack where they set fire to a Mac OS X: http://www.networkworld.com/news/2014/022614-rsa-apple-attack-279212.html

This is slightly older but still relevant:
http://www.serverwatch.com/tutorials/article.php/3936741/Stop-Software-Attacks-From-Destroying-Your-Servers.htm
From the article:
Some simple ways that malicious software can damage your server hardware include:
Overclocking
Overvolting
Overusing (mechanical)
Overusing (electronic)
Overcycling

Stuxnet was software

jflat06's picture
User offline. Last seen 1 hour 11 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

Firmware is different from software. It generally requires elevated privileges to access.

I am not arguing that software cannot damage hardware. There are security holes in the operating system that can make that possible. I'm saying that the responsibility for keeping the system cool does not belong to the client programs. We're just using the CPU. We don't know what will overheat your machine. It is up to the hardware manufactures and operating system.

Joined: 04/19/2009

So the bottom line here is that the devs are not even working on slimming the client down so that the players can run more clients without overheating, or considering working on it?

We've all made it crystal clear that the program is now too bloated for most of our computers, starting when you introduced the filters, and dramatically increasing with NC.

Apple has this to say about CPU usage: "If any process is taking more than 70% of the CPU … it is considered to be putting a significant load on the system." Atm, I have two clients running on the Handfold puzzle, running a simple GUI script on each… one is at 165.9% and the other at 151.8%.

The majority of us are in no position to go out and buy a hot new top of the line computer for foldit.

I'm a little perplexed at why the devs do not see a problem here. Not only are you reducing the number of players, you have effectively reduced the number of clients those remaining players can run. You have also raised the bar so high that new vets will be almost unattainable.

What happened to the "human factor"? Now you only want "human factor" if it comes with the top of the line computer??

Please don't kid yourselves that those of us who can't run many clients will be content to fuss around by hand with a puzzle without trying to see if we can score with it, too. We are all a little unusual here, but part of our enjoyment is the competition.

And it personally leaves me wondering whether it's even worth it anymore. Four clients running was adequate for CASP for me. Two simply isn't - and I am unsure if the frustration of playing only 2 puzzles at a time in the CASP madness is a reasonable way for me to spend my time and computer resources.

spmm's picture
User offline. Last seen 41 weeks 2 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

Elevated privileges means running as admin on your computer - of course no one does that IRL on a Windows home computer.

Susume's picture
User offline. Last seen 1 week 6 days ago. Offline
Joined: 10/02/2011

At some time in the distant past, the biologists and game designers addressed the question, "How do we slim Rosetta down enough to run in real time on the desktop, so it can be interactive?" For several years this slimming effort was successful. During that time, veteran players discovered that they wanted to use the scripting capabilities to explore multiple solutions simultaneously, or multiple puzzles simultaneously. The game developers have never explicitly supported multiple clients, but the biologists have benefited from it because more human-generated solutions were explored, and the players have come to depend on it for their fun and motivation. Now the players are giving you the feedback that the game is no longer slim enough to run multiple clients on the desktop. This reduces the machine-hours and person-hours spent on each puzzle, and reduces the number of starting models that can be explored during CASP. It also reduces the fun factor and motivation of many experienced players.

The devs may be seeing that the number of players to open each puzzle has not gone down, but the players are seeing that the number of people all day and evening in global and vet chat is way down, which means less time is being spent on each puzzle by the people who open it. One person logged into chat may represent one client open or four clients open, or more. When that person has zero chat logins rather than one, it may be a loss of one client pushing one solution to the best score it can reach, or of several clients pushing several solutions to their best scores. Many more solutions are remaining "unpushed," meaning the improved scoring algorithm will not have a chance to select them even if they are the best ones.

Rather than saying "that is not our problem," perhaps you could take the approach of "We don't have a solution for this right now, but we will discuss the possibility of slimming down the client again and see if it is feasible." That at least would give us some hope, and let us feel heard. Or you could come right out and say "We are not planning to work on slimming down the client," if that is the case. Just please don't say "not our problem."

Joined: 04/15/2012
Groups: Beta Folders

I think we all need to cool down:

Jflat was only addressing the issue of overheating. NOT the fact that the clients are hogging extra resources. Is that still an issue? Yep. Should that still be addressed? Sure. Has that already been addressed? Nope.

So:

Players-Stop flaming Jflat about something he never touched on.
Devs: Please work on getting our clients to a state where they no longer hog as much resources as they do now (meaning we're not just seeing a small increase, but seemingly about double).

Joined: 04/15/2012
Groups: Beta Folders

Oh, yeah:

Flame wars will get nobody anywhere. That's what actual communication and working to get the problem fixed is for.

spmm's picture
User offline. Last seen 41 weeks 2 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

Adb - players are not flaming jflat, we are concerned about the noticeable effects on our computers and trying to make that very clear to the developers, testers and the scientists. If we don't tell them they don't know.

Folders from most active groups have contributed to this discussion, so please don't trivialise their comments, which are genuine.

Jflat is quite correct about dust etc, also consumer devices wear out, laptops 3yrs, desktops 5yrs, is a standard industry replacement cycle for businesses because repairs are not worth the loss of productivity or the cost of repair. Mostly these devices only run during business hours.

I think it is very important that the folders who don't know much about computers (I include myself in that category) understand that, depending on the way they play, they can damage their device, and shorten the device life span, especially any form factor which is going to be challenged by heat, all in ones and laptops in particular.

The way folders play has changed since LUA scripts were introduced. Folders grind scripts 24x7 for days, most consumer computers are not designed for that continuous load, especially laptops and all in ones because they haven't been designed to effectively dissipate that level of heat as desktops and towers can. Consumer devices are not designed for continuous workstation, server loads.

Running several clients at the same time just adds to the load of course. Script writers are not necessarily concerned about or aware of the load they place on devices, and they don't know anything about the device as jflat says.

Windows 7 (probably the most common OS) does not warn you that you are possibly doing damage. At the end of the day the effort should be going to getting the right pose, unfortunately we play for score, because it is a game, and it is the only clue we have.

There has been an increasing compute load. This is undisputed, even by folders with high spec devices. This is a problem for folders, also undisputed.

Additionally, experimentation with puzzle set ups, and the reintroduction of old bugs, endless changes to wiggle, filters and rebuild makes it difficult for folders to effectively analyse impacts as there are too many variables and we all have different rigs.

So please don't try to oversimplify or emotionalise a complex conversation which is being carried out with goodwill between the folders and the foldit team.

Even if some of the folders are concerned about the possible growth of tails :D

Joined: 04/15/2012
Groups: Beta Folders

Thank you, you are helping to put this into better perspective for me, and I admit I may have been looking at the comments wrong. Just file my comments away somewhere.

Joined: 09/21/2011

"(meaning we're not just seeing a small increase, but seemingly about double)."
I think slimming the program is one objective. What was overlooked as people talked past each other is this:

When we see these kinds of changes in performance from a program, we can't help but question if there is something not performing correctly in the program.

As gamers, we grow concerned because we are running our own equipment at a steep price. So, while overheating may sound like a problem at our end, it may also be a problem at your end by default.
If you make me a bike that only performs in 10th gear, and I go, "oh crap, I blew out a knee!" and you say, "wow, maybe you should take better care of your knees, or operate your bike more efficiently." It can make a person want to flip you the bird. Not me of course, but another person ;P
Also, when my knee blows out, I can't help wonder, "maybe it wasn't just the 10th gear thing (slimming down)". I want to inspect the whole bike for flaws. Maybe (or maybe not) I'll find the chain had one less link than it should have for this bike (program performing incorrectly).

Joined: 09/24/2012
Groups: Go Science

I did certainly not experiment such a big decrease. On a Mac (laptop), I used 3, now, 2 to 3 depending on the size of the proteins.
On a windows (desktop), I used 3 (to 5 with crashes), now 3 to 5 with fewer crashes.

I might be wrong, but I suppose a computer is able to manage its heating and decreasing CPU performance if heating is too high?

I observed that the number of clients depends heavily on the size of the protein, and the type of commands (scripts) that is running (including wiggle power).

With small proteins with no filter, it's the same as before NC. I suspect that bigger proteins with more filters are released today. This would be the problem, not NC.

Joined: 04/19/2009

Bruno - please see the discussion in these feedbacks, there is proof that something is wrong for many computers.

http://fold.it/portal/node/997299
http://fold.it/portal/node/997291

And please add your CPU chip number to Timo's feedback with a note that you are one who has no problems, it will help diagnose faster!

spmm's picture
User offline. Last seen 41 weeks 2 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers
Status: Open » Closed

Closing this it can be accessed via search and reopened if necessary.

Joined: 09/24/2012
Groups: Go Science

Since the last update (Begin 2018), I observed a reduction of the number of possible clients on the same computer. This could be linked to the size of the proteins (aflatoxin, symmetric ...) and/or the filters.

-reduction from 7 to 2-3 on a Linux platform (I suspect bad memory management problem - the number decreases over time).
-small reduction on a Windows platform
-no apparent reduction on a Mac platform

(this is no true science).

Suggestion for symmetric h Bond filtered puzzles: add a few days (I had no CPU time enough to finish the 1499c one) and/or give us some more days for the "share to scientists" (and some more number of puzzles - I had to delete 2 candidates).

Hanto's picture
User offline. Last seen 1 week 5 days ago. Offline
Joined: 05/10/2008
Groups: None

I suspect 1503 could reignite this issue, much slower for me then prior afavatoxin puzzles on multiple computers. May have to reduce clients by half again, below current ( 1 client/physical core ) and yes, I keep my computers clean. I hear new major windows 10 updates in 2 days or so, could really make for some error prone machines well before 1503 finishes.

LociOiling's picture
User offline. Last seen 15 min 24 sec ago. Offline
Joined: 12/27/2012
Groups: Beta Folders

We should have probably opened a new feedback for the big 2018 update, but here goes.

Using Process Explorer on Windows, I'm seeing Foldit consistently using more than one logical processor.

For example, on my laptop, which has four logical processors, Foldit is often using 30% or more of total CPU processing power, so over 100% of one processor. With three clients running, the laptop is showing 100% CPU most of the time. I'm seeing clients taking nearly 50% of total CPU for at least brief periods.

All three puzzles are relatively large. Even before the big update, large puzzles used more CPU. With smaller puzzles, one thread is normally doing most of the working, taking around 100% of one logical processor. A second thread was active once in a while.

With larger puzzles, the second thread is much busier. I'm currently seeing a third thread using a significant amount of CPU on at least some clients.

The good news is that even with three clients running and Process Explorer showing 100% CPU, I'm still able to use the laptop for other things, and everything is mostly fine. Some things are of course a bit slower.

But I'll probably scale back to two clients on the laptop, as long as we have large puzzles running. Previously, three clients was the rule for this laptop.

I'm also running the aflatoxin puzzle on an older laptop with two logical processors. Several days in, running EDRW on low wiggle power, and no problems. I'll post more about that later.

See Number of processors (Windows) for more on Process Explorer and how to determine the number of logical processors your system has.

"Threads" are mini-programs that are started by a main program. Each Foldit client typically has around 15 threads running. Each thread can run on only one logical processor at a time. When running a recipe or even just a wiggle, there's normally one Foldit thread using about 100% of one logical processor. The problem with the current puzzle mix is that there are two or more busy threads, so one Foldit client is using over 25% of total processing power on my laptop.

jflat06's picture
User offline. Last seen 1 hour 11 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

Just for a recap of the changes that happened in the latest update from the Rosetta software -

There are 2 big changes that have impacted performance:

1) The score function of the game got slower, as the more accurate computations unfortunately also take more time. But the work they do should be more accurate and valuable.

2) We also split up the computation that happens during a Wiggle and some other tools so that the meat of the work is being done in its own thread, separate from the code that copies those updates into the UI. This means that the 'worker' thread is able to work a lot more efficiently. These threads will automatically get split up into your available logical processors.

The net effect of the update is that the computations are slower because of (1), but you should still see good performance and responsiveness because of change (2). However, if you run multiple clients, the advantage of change (2) becomes diminished, since those other clients will already be using up the other cores' processing power.

Hanto's picture
User offline. Last seen 1 week 5 days ago. Offline
Joined: 05/10/2008
Groups: None

Your explanation jflat06 would seem a good reason for removing tracks from the client since it is highly unlikely that the average size of puzzles will be getting smaller. Loci, I use Process Lasso and ALL clients are run at below normal priorities. Been that way virtually from my roots in Foldit and about the only complaints Ive ever had was the increasing folder sizes several years back. I strongly support client diversity as each client creates its own potential. The more the merrier, I'm just a monitor and I like to rag the devs about the over-complexity of the client that actually started with tracks, a looong time before new chapter.

Joined: 09/24/2012
Groups: Go Science

It's a good strategy. Thanks jflat for the explanation.

It's not very important to be able to run more clients, because parallel folding on the same puzzle is exploratory and ultimately a waste of resource (real double blinds are done by different players).

jeff101's picture
User offline. Last seen 3 hours 51 min ago. Offline
Joined: 04/20/2012
Groups: Go Science

I don't think parallel folding on the same puzzle is wasteful.
For example, on different clients you can try different recipes
(or the same recipe with different input parameters) from the same
starting structure, then when one client gets far ahead of the others,
you can stop the laggards and copy the best structure from the leader
to the rest. Then you can start all clients from this new structure
with different recipes on each client. Kind of like having your own
team of evolvers.

Some recipes use random numbers, so having 2 clients run the exact
same recipe from the exact same starting structure can give very
different results.

Another example is for puzzles known to contain disulfide bonds,
one can have each client try a different combination of cysteine
pairs. Usually some combinations work better than others.

bertro's picture
User offline. Last seen 7 weeks 4 days ago. Offline
Joined: 05/02/2011
Groups: Beta Folders

Totally agree here! More clients trying different thing equals more exploration of the solutions landscape.

Sitemap

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, RosettaCommons