Folding Forum

P5-133XL wrote:Actually, 59-60C for a machine at idle is quite high. 30C - 35C should be a normal idle temp.

A simple google search produced this: Review of Dell Vostro 3460 Notebook. Scroll down and Look at the temperature section to see normal.

Thanks for this. Well, I think it might be partially due to the fact that the temperatures I have stated above were not measured while idle (I had Excel, Firefox, Outlook and some other programs opened). Also, my configuration is slightly different - I have i7 (the tested was i5), 8GB RAM (against 4GB RAM), SSD HDD (not sure if this makes any difference comparing to a 'classic' HDD) and an integrated graphics (against NVIDIA) - which I guess is the reason for slight overheating of my motherboard.

The differences in equipment would be responsible for some increase in temperature, but not that much. The integrated video should actually reduce the temperature compared to the tested discrete nVidia graphics. Would not be the first Dell laptop with overheating issues from an improperly installed heat pipe. Of the 2 dozen plus we got the last time for where I work, 4 or 5 had to be put in for warranty service due to overheating within the first few months.

Just because there is opened software running, does not mean that the machine is not idle. Most software uses very little CPU on a continuous basis (unlike folding or BOINC) rather they are just sitting there using ram and waiting for some user input. A much better test would be to run the task manager and monitor the %CPU being used. if the %CPU being used is less than 5% the machine can be considered totally idle.

P5-133XL wrote:Just because there is opened software running, does not mean that the machine is not idle. Most software uses very little CPU on a continuous basis (unlike folding or BOINC) rather they are just sitting there using ram and waiting for some user input. A much better test would be to run the task manager and monitor the %CPU being used. if the %CPU being used is less than 5% the machine can be considered totally idle.

Let's restate the same facts in another way. With less that 5% CPU, the temperature should be 30C - 35C rather than 59-60C. Both FAH and BOINC are designed to use the CPU continuously and that will use more power and will certainly produce a lot more heat. Almost all desktop machines and many laptops can manage that much heat; some cannot.

Ok, I think that in terms of the next steps I will just properly backup everything on my laptop and will see how it behaves for a couple of days/weeks. As I said, it is a fairly new PC (only 4 months old) and with the 3-year next business day warranty. If I get a feeling that the laptop might not be able to sustain the continuous loading/heat, I will call the customer service and raise the issue with overheating. I am just a bit cautious about calling them immediately, because the last time I used this type of service for my old Dell laptop, they did the repair so quickly and not really properly, that I had to call them again in a couple of days... And it went on and on - the on-site maintenance can get quite frustrating!

All

I am sorry for posting another (and a bit off-topic) message, but I have now realised that I might not be able to finish the work unit before its timeout... The project code is 8579 (I guess we cannot choose a project) and the ETA is 7.56 days. However, the Timeout shows 03/11, which is in four days. The estimated TPF is 2h 30 min. I guess I might be able to meet the deadline if I increase the CPU/core usage, but that might overheat my laptop completely.

I therefore wonder, whether there are any means for controlling which tasks to take, so that I would make sure that I will meet the deadlines with the minimum configuration. Or, if not, then maybe FAH is really not for my laptop at all...

Thanks for your responses!

As I said before in not so many words, if the hardware cannot run a minimum number of hours per day with a minimum amount of processing power without overheating, then the client won't make the deadline and you shouldn't run fah on this light weight laptop.

Right. You cannot choose projects -- and even if you could it wouldn't help. Projects which require more processing are given longer deadlines and earn more points but conversely, projects which happen to need less processing are given fewer points and shorter deadlines. They're roughly normalized and if you can/cannot meet the requirements of one project, you probably can/cannot meet the requirements of another project. There's no formal minimum stated, but that doesn't mean there isn't one and apparently your preferences or your hardware is insufficient.

Thanks. I indeed have the PC turned on cca 18 hours per day (and can keep it running over the night as well), but if I am unable to meet the deadlines, then it is just a waste of hardware, I guess. Could you therefore please advise, which other platform should I try? I would really like to contribute somehow, because I have my laptop turned on anyway and most of the time I don't need its full power at all.

Should I give BOINC another try? Are the BOINC projects also limited in terms of time/minimum computing power?

Ivkosky wrote:Should I give BOINC another try? Are the BOINC projects also limited in terms of time/minimum computing power?

I believe so, but definitely not as much as F@h. You are welcome to try BOINC if you like. There are biomedical projects out there under that platform that you can join.

Can't answer about BOINC, don't use it and that is not the purpose of this forum. However, is the estimated completion still the same after running long enough to finish several percent? The client's estimates can be off if the WU is recently restarted and until 2 or 3% has been processed.

Joe_H wrote:Can't answer about BOINC, don't use it and that is not the purpose of this forum. However, is the estimated completion still the same after running long enough to finish several percent? The client's estimates can be off if the WU is recently restarted and until 2 or 3% has been processed.

I know it's not a purpose of this forum, but I thought somebody would have an experience with both clients and, while also reading this thread, can advise what to do in my situation without me having to explain the whole thing from the scratch on another forum. And, also, I guess BOINC is not really a "competitor", as we don't play on competition in science (or do we? ).

The estimates vary quite a lot. I have had my laptop turned on and computing for almost 48 hours and sometimes it shows 7.65 days, sometimes 7.52 days, but it goes up and down almost randomly. I guess this has something to do with me using the PC or not... However, as I mentioned before, I am running it only on 2 cores out of 4 and with 30% CPU power, so my other work on the laptop (which is mostly Word, Outlook, Excel - but not heavy computing at this moment, or Firefox) should not have any effect...

Again, anything other than fah and fah support is considered off topic here. This is a fah forum, not a forum for other projects. Sorry. And you don't want to know my experience with DOINC.

Ivkosky wrote:However, as I mentioned before, I am running it only on 2 cores out of 4 and with 30% CPU power,

By this do you meant you have moved the CPU slider in the client from 100% to 30%? You had not mentioned that earlier, that setting is really for just single-core processing. That would explain the very slow running of the WU, with just one core limited to 30%, the other is always waiting to synch data between the two threads.

Joe_H wrote:

Ivkosky wrote:However, as I mentioned before, I am running it only on 2 cores out of 4 and with 30% CPU power,

By this do you meant you have moved the CPU slider in the client from 100% to 30%? You had not mentioned that earlier, that setting is really for just single-core processing. That would explain the very slow running of the WU, with just one core limited to 30%, the other is always waiting to synch data between the two threads.

Yes. Well, I have moved the slider to as low as 10% in the Advanced tab in the settings and limited the number of cores to two in the Slots tab. What should I do then?