Low level CPU job is too high !

[Neil-B]

The "folding power slider" was a quick, simple approach (using config settings for the most part already available I believe) to manage what was then a relatively simple situation, which is now a lot worse with the growth in the numbers of cores and faster/hotter/more tweaked consumer hardware … and it is just simplifying what can be done manually - a new variant of the slider could have more granularity/steps but for the most part it is just doing what Autourdupc didn't want which is restricting the number of cores (and pausing GPUcore on light iirc).

The very high boost clocks should be managed down by the laptop configuration if the load is sustained so there should be no need for a particular software package to have to adapt - sometimes manufacturers set boosts to the point that the cooling system can't respond quickly enough as they assume the laptop will only be used for short spikey bursts of usage … but if vendors have pushed their laptops CPUs and GPUs clocks that they are unstable at sustained high loads it is hardly fair to blame the software as this is a manufacturing design decision.

At the moment the interest from big corporations in allowing software such as FAH to run is at an all time high … Once the current situation is over many of the currently relaxed corporate security policies may well revert to pre Covid-19 levels which has less to do with anything performance wise and more to do with systems security postures … Any big corporation wishing to continue to support FAH has a significantly greater offering such as server farms and/or desktop workstations during off peak which are better designed to run heavy loads, rather than utilising hardware that for the most part hasn't been designed to cope with such loads.

[Autourdupc]

Take a look at VirtualBox, You can adjust CPU load from 0 to 100%... Slider in FaH should do the same…

[alxbelu]

The very high boost clocks should be managed down by the laptop configuration if the load is sustained so there should be no need for a particular software package to have to adapt - sometimes manufacturers set boosts to the point that the cooling system can't respond quickly enough as they assume the laptop will only be used for short spikey bursts of usage … but if vendors have pushed their laptops CPUs and GPUs clocks that they are unstable at sustained high loads it is hardly fair to blame the software as this is a manufacturing design decision.

If 99% of all software and use-cases are not causing the hardware to be running at max load continuously, then surely it makes sense to design according to actual use?

And it's not about it being unstable; running at, or close to, TjMax for extended periods of time is not ideal as it might also heat up other components that are not designed for such temperatures (e.g. causing cracks in PCBs) - changing a fan curve is thus simply not a solution if you're already hitting the throttling temperature.

As for interest from companies, yes, it's at an all time high and some companies may take greater steps to ensure that FAH can be used on their fleet of computers, including laptops, at least in terms of temporarily allowing it. As I've proven, the "load" is perfectly fine on laptops, as long as the load is kept in check it can generate significant processing power without being neither a nuisance nor a notable cost in terms of prematurely broken hardware or tech support. Thus if FAH wants to retain at least a fraction of this interest, it needs to have a functioning power slider/limiter.

I'm sorry, but there's simply no excuse nor argument not to attempt to solve this (cost/benefit is simply so low). That said, I'm obviously, again, not expecting this to be solved near-term while the team is still putting out fires (as the current cost/benefit is considerably higher).

[Neil-B]

Why? .. Virtual box developed by a large team of developers to deliver a configurable virtual machine … FAH developed by a much smaller team to deliver compute resources to Biochemists to get science done … How are they even comparable?

If the potential from laptops is as great as asserted then of course investment in resource (if it is available) should be investigated and a slider that delivers what each and every variant of user wants should be developed … but I fear the cost/benefit may not be as low as you believe and that there will be far greater gains to be had focusing development on other areas - I am however happy to be proved wrong.

[Autourdupc]

If FaH was able to manage CPU load, I could ask my friends to install it on their computers, but actually, I will not.
And I also removed FaH from my 5501... I only run it on 2 olds computers… I think there is a large quantity of CPU compute lost due to FaH not usuable for personal computers.

[PantherX]

Take a look at VirtualBox, You can adjust CPU load from 0 to 100%... Slider in FaH should do the same…

Please note that changing the execution cap may cause guest timing problems: https://www.virtualbox.org/manual/ch03. ... -processor

Also, there's the issue of context switching. Setting too low of an execution cap will cause BSOD: https://forums.virtualbox.org/viewtopic.php?f=1&t=92962

Given that VirtualBox by default has the execution cap set at 100% is a really good indicator of how it was meant to be run. I don't think F@H would use that approach because:
1) F@H shouldn't be causing BSOD which can happen if the execution cap is being implemented the same way as VirtualBox
2) F@H already has a more stable approach towards limiting CPU Usage, it's by modifying the CPU count
3) VirtualBox uses virtualization which would slow down the folding performance when compared to native
4) Given the dangers around BSOD, it could negatively impact the integrity of scientific data which is of paramount importance

[PantherX]

...If 99% of all software and use-cases are not causing the hardware to be running at max load continuously, then surely it makes sense to design according to actual use?

And it's not about it being unstable; running at, or close to, TjMax for extended periods of time is not ideal as it might also heat up other components that are not designed for such temperatures (e.g. causing cracks in PCBs) - changing a fan curve is thus simply not a solution if you're already hitting the throttling temperature...

I could say that 99% of software used by content creators and multimedia producers regularity run software that uses 100% of their system resources... it's very subject and a slippery slope

As long as a CPU and GPU is running within the technical specifications that it was designed to operate in, it shouldn't matter as they are guaranteed to be functional within those parameters.

The issue about heating up components is two-fold:
1) Chassis design - IT should accommodate adequate cooling system which handles all components being stressed out. There's a reason burn-in tests are run by reputable system builders: https://www.pugetsystems.com/included.php?inc=burnin
2) Use case - When you buy a system, it is generally catered for web-based activity, light usage, office work, gaming, rendering, etc. It would not be fair to perform rendering/gaming tasks on a web-focused system.

Do note that I am only focusing on pre-built systems. Building your own system is a completely different aspect.

F@H started with desktops and then with the rise of laptops, they attempted to make a "light version" of F@H which was deployed via the Chrome browser (viewforum.php?f=95) which was successful IMO but after few years, Chrome deprecated that support hence F@H had no choice but to stop it. Having said that, who knows what technology can be used in the future to capture all the portable devices

[ipkh]

Unless you buy a workstation class laptop, your typical consumer pc/ laptop isn't designed for high utilization.
You can blame Intel for this one, they designed the chips to use max watts on all workloads from Single core on up. That's why your all core speed if 2.5Ghz it so and Single core is 4.2Ghz or so. For Windows use the power plan settings to adjust max clock speed. This is the best way to limit power/heat use.

On a macbook, blame Apple for prioritizing looks over heat dissipation.

[uyaem]

And it's not about it being unstable; running at, or close to, TjMax for extended periods of time is not ideal as it might also heat up other components that are not designed for such temperatures (e.g. causing cracks in PCBs) - changing a fan curve is thus simply not a solution if you're already hitting the throttling temperature.

Thermal stress is caused by change of temperature, not constantly high temperature.
Therefore you'd rather have a constant load (whether that's half or full...)

[Autourdupc]

Using VirtualBox, it works like a charm !

But what is FaH looking for ?
Heavy CPU, but not so many as possible
or
Plenty of CPU with some low power ?

[bruce]

A well designed computer will dissipate the heat generated by its internal components. If I take my car out to a safe off-road location and it is damaged by overheating when I see how fast it will go, can I expect the manufacturer to warranty replacement of my engine? I guess it would be an interesting lawsuit and trial. The manufacturer never tested it at full throttle but they also didn't warn me that it couldn't tolerate opratiing it at full throttle.

[alxbelu]

I could say that 99% of software used by content creators and multimedia producers regularity run software that uses 100% of their system resources... it's very subject and a slippery slope

As long as a CPU and GPU is running within the technical specifications that it was designed to operate in, it shouldn't matter as they are guaranteed to be functional within those parameters.

Right, and content creators never actually use their laptops, they only have them for 100% rendering and encoding. Also clearly the majority of laptop users are content creators.

Thermal stress is caused by change of temperature, not constantly high temperature.
Therefore you'd rather have a constant load (whether that's half or full...)

Yes, and laptops are usually allowed to maintain folding while unplugged in people's bags.

A well designed computer will dissipate the heat generated by its internal components. If I take my car out to a safe off-road location and it is damaged by overheating when I see how fast it will go, can I expect the manufacturer to warranty replacement of my engine? I guess it would be an interesting lawsuit and trial. The manufacturer never tested it at full throttle but they also didn't warn me that it couldn't tolerate opratiing it at full throttle.

Pro-tip, do not drive your normal car offroad and then try to make a warranty claim - I guess under common law jurisdictions you might have a case unless there's any fine print in the manual (which there definitely is). As for overheating, a modern car would most likely not do this even under special circumstances (e.g. limiting ventilation/cooling), just as laptops in general won't (both would generally have hard-off thresholds). This doesn't mean there's no wear or risk for either, whether that is discovered/noticed during warranty or not.

I'm also absolutely amazed at the replies and defence of a broken (valuable) feature in this thread - clearly you want to enable more folders to join.

[Neil-B]

There is no defence of a broken feature going on … the slider is what is has always been … technology has moved on and bypassed its current usefulness, but it still does what it was designed to do … people just want more - and that is not a bad thing but the current slider just doesn't do it … and actually some of the things people want (in this thread and others) aren't in the remit of FAH to sort - they will require hardware manufacturers and OS developers to make changes.

Yes, there are groups of the folding community who would like a very much cleverer "Slider v2" (although it might actually be 3 or 4 I can't remember) but to be honest a fair bit of what is being asked for might be better delivered with a different approach … when resource can be found and as future rollouts of clients and cores occur I am sure these views will have been heard and taken into account.

Your final sarcastic statement does both the FAH team and yourself a disservice … You see things one way, from one perspective and that is your right - but the FAH team and the many volunteers who actually help as well as they can to keep FAH running (and growing as fast as they can make it) can only do so much and may have a much clearer view of a much greater picture … telling them that they clearly don't care about enabling more folders to join is in my view a silly, insensitive and patently untrue statement (but that is just my opinion).

If you really think that stopping doing any of the other valuable work that people are maxed out on at the moment and sorting out this one group of features that impact a relatively small part of the community would actually enable more people to fold than all the other work that is going on then you are absolutely right to state that … you may however be wrong in your assertion.

[alxbelu]

If you really think that stopping doing any of the other valuable work that people are maxed out on at the moment and sorting out this one group of features that impact a relatively small part of the community would actually enable more people to fold than all the other work that is going on then you are absolutely right to state that … you may however be wrong in your assertion.

I have on multiple occasions clearly stated that I understand that this is not something that should be prioritized short-term, but should be implemented long-term to enable the massive potential of currently unsuited laptops and mobile devices. You can argue all you want that these devices should "just get better cooling solutions" and that there's "nothing wrong" with FAH (which is plainly the sentiment conveyed by many here), but it won't change the fact that slim/ultraportable devices with inferior cooling but snappy responsiveness in bursts sells more than draggable heavy laptops with appropriate cooling for long-term loads.

I've understood FAH as having the intention of allowing nearly everyone to participate in the folding, especially if you have fairly recent hardware, yet the sentiment here on the forums is very often that anything not designed to run 100% load all of the time is poorly designed and shouldn't participate - which arguably means anything that's somewhat portable, and that indeed not everyone should join but only those with appropriate technical skills or with appropriate machines. As nothing about limitations are mentioned on neither the FAH landing page, nor on the download page, I can only assume the latter forum sentiment is not really something that should be associated with FAH.

Either this forum sentiment or the FAH official description will need to change.

[Neil-B]

There are many, many instances of forum members helping people with myriad issues ... some relating to heat/cooling/noise ... some relating to drivers/setups/configs ... others relating to confusion about what FAH actually is and how it works ... and there are significant numbers of threads where people are helping each other through technical issues and spread knowledge as they can ... for every thread I spot that might be considered elitist I see many more where people are just trying to help, and I see the occasional one where someone simply wants to say FAH is S&*t and why aren't you all mining bitcoins as that is the sensible thing to do !!

Yes there are a core of people who strive to have the best, fastest, kit who believe that WUs should be prioritised to the kit that turns around the Science the fastest - feel free to read my posts (and those of numerous other people) that speak out against that and don't mentioned to me on the "relay team" metaphor people use to support these views ... On the flip side of the coin there are people who want (for very understandable reasons) to fold on kit that which may be old or underpowered in someway, or don't have the capabilities to utilise the current folding cores, and are simply not able to turn around the science at the speed currently necessary to return WUs before the expiry deadlines ... But there are people on this forum who try really hard to help people get to the point where they can do this, or explain as simply as can be done why their kit isn't able to fold.

You are absolutely right that slim/ultraportable device outsell draggable heavy laptops - no-one has disputed this as far as I am aware ... but laptops are simply one type of kit that folds an expanding market for sure and one that does bring its challenges (heat, noise, not always on) and these needs to be looked at, and if possible addressed ... and FAH has many, many, many issues with all types of other kit, and the infrastructure behind it, and the science that needs to be done, and the development and support structure, and the recent amazing swell of support and new folders which has rapidly overwhelmed the old ways of working which had been build on decades of enthusiast support ... this is what the forums have always been, an outlet for enthusiasts to try to guide people new to folding through some of the mess from their own experiences of having been through the same learning curve ... with limited resources and a not huge pool of folders FAH was what it was - an enthusiasts platform donating compute resource for scientists who didn't have access to anything better ... and there are many lessons to be learnt about how the current surge of new folders has been addressed - no one would have planned it this way - but it has happened.

People who help on this forum are not trained support staff, they simply try to help and I guess sometimes the way they try may not come over right to other people ... The nature of the beast that has evolved is that some things are fairly complex technically (sometimes for no better reason than it evolved that way) and explaining these in terms that make sense to others who may not have either the technical background or the experience with their kit to necessarily fully grasp why "odd" decisions may have been made in the past doesn't help one iota ... Some peoples "style" verges on the abrupt - others wax lyrical (or try to) - some people are blunt, others soften the blow and I guess that is simply where we are today ... Are the forums the right way to support FAH? - I don't know, but at the moment it is what we seem to have and I try to help as I can.

Personally (and I am really glad it wasn't my call) I might have actually put in place some form of "qualifications for entry" - not to be elitist, far from it actually - simply to try and protect potential folders from some of the dangers/pitfalls/costs/issues that enthusiasm and a desire to help and be part of "Fighting Covid-19" might have left them unaware of ... but it wasn't my call and I am fairly sure it wouldn't have worked, and it would have had to be sorted damn quick, but hey it wasn't ... instead a more than significant number of new folders have been able to install/fold/return WUs with little/no problems ... yes, an equally significant number may well have installed the software (or tried to) and then decided it wasn't for them and removed it - a nicer tooling and better web/social media presence may have retained a good few of these - and their loss is a loss to the science without doubt - but the folders who have been able to get on with it have been a massive gain.

Forums are always "tricky", people tend to come to them when things aren't the way they want them to be or when they need help ... and these can be emotive, stressful times for both the people needing assistance and for those trying to help ... For the most part I have seen far, far less abuse/insults/negativity on these forums than others I have frequented, but is it perfect - absolutely not.

The potential of laptops (and lower powered desktops, and Intel GPUs, and Phones) needs to be properly explored ... There will always be hard decisions for the FAH team to make as a balance needs to be found, or a way of balancing needs to be found that works the scientific approach (currently) a need for fast turn around of WUs (at least at the moment) with the use of masses of machines that are simply "slower" for whatever reason ... Decisions need to be made by those who have to find the balance and who have control of the resources ... I hope that going forward that if enough support for FAH sticks around that resource will be found to develop a much more polished and coherent experience than has been able to be developed to date.

Communications and "induction" is way overdue for an overhaul to help potential folders have a fuller understanding of the challenges that folding can bring - much of it exists and is scattered around the forums in the various threads helping people but a forum structure that "forces" people to at least be aware of a clearer up front statement might help - as would people actually reading any of the existing guidance and searching existing threads?

You have your view on the sentiment of this forum and you feel passionately about it which is good - I have mine, I see a lot of people trying really hard to get through a wholly unique and difficult situation getting told that FAH in a surprisingly helpful and well mannered way ... Will FAH change, yes I am sure it will ... Will it address your concerns over the current sentiment, I really hope it does because it needs to ... Will FAH be able to provide instantly working solutions for every type of kit, whilst I will not hold my breathe for this it would be awesome if it did.

I apologise if anything I personally have said it this post, in other my other posts on this topic, or anything other post I have made in these forums has added to you views about the sentiment within the forums and FAH - This has sincerely not been my intention in any way.

Please stick around, challenge as you have done where you feel FAH and/or these forums are not addressing points/views/issues you feel are valid - because that is how things will change.

Thank You for giving me the opportunity to write this ... and a final apology that this missive is so long.