Folding@home-Team Linux Mint-76140

[tinca]

Hi all,

I think that I have found out what the problem is/was.

I have spent all this morning playing with the computer. I have uninstalled Fahclient a couple of times and downloaded a new one, as this program seems to be the cause of the heating problems?. The results are the same.

We had come back from the shopping and I restarted the computer again. I started Fahclient and it had been running about 2 minutes when I heard the fans in the computer slow right down to a whisper. I was looking at Psensors and saw the temperatures drop from 84 down to 30 in 2 seconds. I watched and it carried on at the same normal temperature. So what caused that to happen.

I believe that the pump in the Corsair Hi 100i had stopped working. This allowed the small amount of now static water in the pump to get very hot, very quickly, and the fans sped up in an attempt to lower the temperatures. For some reason the pump started to work again, the water started to flow through the radiator and the temperature fell rapidly. The fans now sensing the lowered temperatures slowed right down.

The computer has been running for about 5 hours now and is still running as normal. I have no confidence in it at the moment and I think that I will have to purchase a new cooler and fit it. Not a job I would look forward to as my eyesight is not as good as it once was, and due to my age, my fingers and hand coordination is a bit of a worry. I do not know whether to let it run overnight or to shut it down.

Still we are up and running again.

Best regards Keith

[BG405]

I believe that the pump in the Corsair Hi 100i had stopped working. This allowed the small amount of now static water in the pump to get very hot, very quickly, and the fans sped up in an attempt to lower the temperatures. For some reason the pump started to work again, the water started to flow through the radiator and the temperature fell rapidly.

I'm no guru on water cooling (beyond my limited budget) but is it plain water or do you use any anti-scale/corrosion additive, like that used in central heating systems? Perhaps it just needs cleaning out, especially if it's been running for a while. Distilled water may also be a good idea, just my thoughts on this.

As for FaH, I expressed an interest early on in this thread but didn't follow it up. I'm wondering if it's worth setting it up on the machine used for watching videos on the TV as it's left on all the time; not exactly high-end kit but if enough of us added such machines, would the contributions make a worthwhile difference?
Specs:

Code: Select all

brian@BG-Aspire-ES1-111M ~ $ inxi --full
System:    Host: BG-Aspire-ES1-111M Kernel: 4.15.0-15-generic x86_64 bits: 64 Desktop: N/A 
           Distro: Linux Mint 18.2 Sonya 
Machine:   Type: Laptop System: Acer product: Aspire ES1-111M v: V1.10 
           serial: <root required> 
           Mobo: Acer model: R2 v: Type2 - A01 Board Version serial: <root required> 
           UEFI: Insyde v: 1.10 date: 08/20/2014 
Battery:   ID-1: BAT1 charge: N/A condition: 30.3/29.9 Wh (101%) 
CPU:       Topology: Dual Core model: Intel Celeron N2840 bits: 64 type: MCP 
           L2 cache: 1024 KiB 
           Speed: 500 MHz min/max: 500/2582 MHz Core speeds (MHz): 1: 500 2: 500 
Graphics:  Device-1: Intel Atom Processor Z36xxx/Z37xxx Series Graphics & Display 
           driver: i915 v: kernel 
           Display: server: X.Org 1.17.1 driver: intel unloaded: fbdev,modesetting,vesa 
           resolution: 1024x768~75Hz, 1024x768~75Hz 
           OpenGL: renderer: llvmpipe (LLVM 6.0 128 bits) v: 3.3 Mesa 18.0.5 
Audio:     Device-1: Intel Atom Processor Z36xxx/Z37xxx Series High Definition Audio 
           driver: snd_hda_intel 
           Sound Server: ALSA v: k4.15.0-15-generic 
Network:   Device-1: Qualcomm Atheros QCA9565 / AR9565 Wireless Network Adapter 
           driver: ath9k 
           IF: wlp2s0 state: down mac: 10:08:b1:c5:6e:e1 
           Device-2: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet driver: r8169 
           IF: enp3s0 state: up speed: 1000 Mbps duplex: full mac: c4:54:44:e7:92:e5 
Drives:    Local Storage: total: 29.12 GiB used: 16.28 GiB (55.9%) 
           ID-1: /dev/mmcblk0 model: HBG4e size: 29.12 GiB 
Partition: ID-1: / size: 9.72 GiB used: 9.06 GiB (93.3%) fs: ext4 dev: /dev/mmcblk0p3 
           ID-2: /boot size: 495.8 MiB used: 328.1 MiB (66.2%) fs: ext2 dev: /dev/mmcblk0p2 
           ID-3: /home size: 15.62 GiB used: 5.51 GiB (35.3%) fs: ext4 dev: /dev/mmcblk0p4 
           ID-4: swap-1 size: 2.50 GiB used: 1.38 GiB (55.2%) fs: swap dev: /dev/mmcblk0p5 
Sensors:   System Temperatures: cpu: 39.0 C mobo: N/A 
           Fan Speeds (RPM): N/A 
Info:      Processes: 215 Uptime: 106d 20h 58m Memory: 1.83 GiB used: 1.45 GiB (79.3%) 
           Shell: bash inxi: 3.0.24

                                                                                                                            ...
Logged in via ssh -X hence no DE listed; it's KDE fwiw.

[powerwagon75]

...My guess is that you had to update the driver with the new install. I lost at least 10% when I installed mint 19 and nvidia 370/375 was no longer available.

Thanks, pretty much what I suspected, and you are correct, it gave me 384.130 this time around. Wonder if it would be worth the headache or not to try and go back to 375. And you're spot on with the 10% hit too.. 475K down to 427K is 10%, and once it gets settled in, I'll be lucky if that is my "new" average.

Were you already implementing intel-microcode prior to the driver change?

[greerd]

Were you already implementing intel-microcode prior to the driver change?

Yes.
I happened to be part of a thread where the OP used the old Edgers ppa which was the defacto nvidia driver ppa before graphic-driver (launchpad) ppa was here and the OP said that 375 was the last available. Since I have timeshift setup I might just try and install 375 this weekend. It was around 384, 389 that nvidia patched their drivers for meltdown/sceptre and it seems to offer the same 10% performance hit that the cpu patches offer.

[tinca]

BG405,

to the best of my knowledge the coolant used in the Corsair H100i is mostly water, Propylene Glycol, and then trace amounts of anti-corrosives and biocides.

The system is a sealed closed loop one, so changing the fluid is not a practical thing. I have looked at when I built this machine and find that it was 2014, some 5 years ago. The working life of such a unit is 5 years. It has been working non stop for all of that time so it is reasonable to say it is close to dying.

As to your own computer and your question as to is it worth it, well as you say it is left on all the while, then why not let it run folding. I do not know how it will run, but as the old saying goes "every little helps". It is for 2 really good causes. A quest to help Stanford find out as much as possible about the major diseases and it is also good for the name of Linux Mint.

Give it a go and let us know how you do.

Best regards Keith

[battlenut]

F@H is a great way to use idle resources if you can afford to. plus on the side it gets more points for our team.

[battlenut]

well we are at 109 by team.

[tinca]

Code: Select all

battlenut wrote
well we are at 109 by team.

thanks to the great push by darkteheme

Best regards Keith

[tinca]

Hi all,

since the computer threw a fit the other day I have tried to come to a conclusion as to what could have happened.

I was running folding as it always does, and for some reason I looked a Psensors which is always on the screen. I was stunned to see that all 4 cores of the CPU were in the high 80's. I know that folding will be using 1 core, but this was all 4. I paused the folding and just sat and waited to see if the temperatures would come down. Well they did, but very very slowly.

I shut the computer down and left it for about an hour. When I started it up again, Psensors still showed the CPU cores were all in the high 60's, and I would have thought that they would be lower. As folding was involved I purged the folding program from the computer and just let it run. The temperatures slowly started to rise again, and that was when it was doing nothing at all.

I reinstalled folding and as I looked at Psensors I saw the temperature rise from 60 to 98 and the computer shut down in a couple of seconds. I had forgotten that as soon as folding starts it will be using a single CPU unit. I tried several times that day and several more times the following day and the only thing that I can think of is that the water pump had stopped. This left the small amount of water in the pump to get hot rapidly. I checked both of the pipes and they were barely warm. They both felt the same. If the pump was running one of the pipes would be warmer than the other, this would also explain the 4 cores going so high. It would be like running a CPU without any cooling at all.

Suddenly the temperatures dropped back to 30ish and I set up folding to use a GPU unit and away we went. It has been running OK since.

My problem now is do I purchase and fit a new cooler or do I wait?. Anyone have any other ideas please.

Best regards Keith

[darktheme]

From what I've read so far, I would start looking for a new cpu cooler. I have 2 AIO coolers but 1 has 2 years service and the other 1.5 years. 5 years non stop service is very good for one of these I think. But you should consider something new and you could save by going to a good air cooler but that's the decision you'll need to make.
Better safe than sorry by risking taking out the mobo or gpu with water leakage or temp overload.
Always leave at least 1 cpu core available for folding on the gpu.

[BG405]

Better safe than sorry by risking taking out the mobo or gpu

+1. In your case leakage might not be likely, but those high temps could well damage the CPU & other components.

the old saying goes "every little helps"

F@H is a great way to use idle resources if you can afford to.

My thoughts too!

Give it a go and let us know how you do.

I'm on it! Installed and has been running for an hour & five minutes @ time of post. Seems to be working OK; using 1 CPU core and am happy with the temps (fanless system).

[battlenut]

@tinca,
If you are going to buy a new cooler check out Raijintek stuff, they have both water and air. I have one of there GPU coolers (Morpheus II) and one of there CPU Coolers (TISIS). The GPU cooler with BeQuiet fans on low it hangs around mid 50s whiles folding. The CPU cooler hangs in the low 60s under full load. just something to check out. I highly recommend this company.

[tinca]

Hi all,

thanks for your comments.

I ordered a new cooler early Saturday morning as the computer is making me very nervous just waiting for the pump to stop again.

battlenut I ordered the cooler before I read your latest post. I have never heard of the company Raijintek. I had a look on the overclockers.UK website and they do look very interesting, and I will spend a bit of time learning more about them.

Best regards Keith

[tinca]

BG405,
welcome to the folding team. glad to have you aboard.

Our team is still going up and at the moment we are 108. Not too far to go to get that top 100 place.

Best regards Keith

p.s. I hope to put in the new cooler today?

[darktheme]

Good luck Keith and also a big welcome to BG405!

Apologies for my lack of production the last 2 days. My crunching team is in the middle of a competition and although it's CPU only, I needed to do some testing and figure the best way to go after it. I've settled on just running 1 GPU until further notice but should get about 800k-ish ppd.

Onward and upward!

[BG405]

Thanks for the welcomes tinca and darktheme! I didn't really expect to see anything on the stats pages yet and certainly didn't expect to not be at the bottom of the list, permanently .. a pleasant surprise and nice to see it's working & actually doing something useful.

Next job, getting the server back up and running again .. another fatal crash yesterday morning so decided "time for the long-postponed reinstall with the replacement HDD" only to find the installer crashes at the end each time. Tried a few workarounds with no luck but that's not for this thread. I'll probably just do the install on another machine, maybe this one. Disk is already populated with data & mostly up-to-date; can do some of that tonight as have the old & new HDDs with me. Server will likely get an instance of F@H running on it at least part time, after some testing with temps & system load. It could potentially more than double my output.

[tinca]

Hi all,

back and running again. A relatively easy job turned into a nightmare.

I purchased a Corsair H100i Pro to replace my Corsair H100i Extreme water cooler. I removed the old cooler and installed the new one, so far no problems. I fired up the computer and as soon as Psensors was visible on the Desktop I saw that all 4 cores were in the high 80’s and low 90’s. I quickly shut the computer down and had a good look at what I had done. Nothing apparent, but I did notice that although the backplate and pump were fitted correctly it wobbled.

I fired up the wife’s computer and looked to see if anybody else had seen the same situation. Well they had and there was lots of post about the cooler wobbling about on its back plate. The answer was to fit a washer(s) on the back plates movable location pins. When you think about it unless all motherboards are exactly the same thickness it would allow the pins to move even after the stand offs were fitted and screwed down if the pins protruded too far above the motherboard.

Problem solved I thought. I fitted a paper motherboard stand off washer to each of the 4 pins and tried to install the backplate again. As I pushed the backplate onto the back of the motherboard one of the rubber pin locators pinged off, dropped onto the computer table and promptly rolled off the end. My wife and myself spent the nest 4 hours looking for it. We gave up nearing bed time and would try again in the morning.

Still could not find it so I decided to use the old back plate as it fitted before on the old cooler. I decided to make sure and fitted a washer onto each of the location pins. Have you ever tried to do this job on your own. Hold the plate on with your left hand and reach through the other side of the computer to try and locate the holes to screw the stand offs into, and hold a torch. I gave up in frustration and asked my wife for an extra pair of hands.

Finally finished, plugged in all of the cables and fired up the computer. I needed to get to the Desktop as soon as possible to check the CPU temperatures. The sorry story continued as the keyboard did not work and I could not log into the system. Found another keyboard and saw that the temperatures were OK. Thank heaven. A simple job jumped up and bit me.

Back up running and folding.

Best regards Keith

[battlenut]

Glad you figured it out keith, good to hear all your temps are back to normal. next week or two one of my machines will possibly be down for maintenance for a few days. got a different case that I could install 3 X 360mm Radiators in. So I have to transfer my whole loop over.


Frank

[tinca]

battlenut wrote

Code: Select all

 got a different case that I could install 3 X 360mm Radiators in.

I did not know that you could get 3 radiators into a single case. Just out of curiousity, what is the case??

Best regards Keith

[battlenut]

Its in the link except mine is black. Tons of room a little on the big side,but exactly what i need.

https://www.newegg.com/Product/Product. ... lsrc=aw.ds