Xeon cpu is frying my hard disks

Message boards : Number crunching : Xeon cpu is frying my hard disks
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1842467 - Posted: 15 Jan 2017, 20:18:57 UTC

I have a used HP workstation running a Xeon W3665 at about 3.20 Ghz. I run it with a GTX 750Ti video card under Win7/Prof. The GPU continues to finish a WU every 10-20 minutes or so. But the motherboard processing isn't what it used to be.

I am running it with the TThrottle program because so far, this thing has killed 3 hard disks. Ever since the first hard disk died I have been running a Raid 1 (Duplex hard disks) so I don't have to rebuild the OS, the drivers, and Seti from scratch. The drives are lasting maybe a year or so (I seem to have to buy another drive a year). Since "heat kills" and everything I could see indicated it was running pretty hot.

So once I started TThrottling it down based on the temperature, the production of the 4 core/w/hyperthreading has slowed from 4-6 hours per WU to 1-2 days per WU. I have lately gotten a little more info which appears to mean I can up the temperature on the throttle to about 80 C and the CPU should be ok. That looks like the CPU's are running at near 30% which is an improvement.

So I am wondering if simply putting a floor fan in front of the case after I blow the dust out will let me approach full throttle again? Or am I really going to have to upgrade/reheat paste(?) the CPU cooling system to get the heat out of the box?

Tom
A proud member of the OFA (Old Farts Association).
ID: 1842467 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 1842470 - Posted: 15 Jan 2017, 20:33:55 UTC - in response to Message 1842467.  

Well, as you already say, you need cooling. Lots of it, as I can see.
Probably best to buy a new case that can have loads of fans in it, perhaps get water cooling for the CPU?
Although doing a quick search on others running a W3565 with water cooling shows that even then the CPU can run at 80C. However, since the water cooling has the fan on the outside or top of the case, the hot air should be blown out and away from your hard drives.

You can also just opt for not using the CPU, or not use it on all cores. That may have impact on your RAC, but is a cheaper solution than having to buy a new HDD every year.

Having said that, it may also depend on the HDD.
E.g. WD Green can withstand way less heat or writes than a WD Blue or Black can do.
Or you may want to look out for SSDs, since these have no moving parts and can withstand heat a lot better.

How much money are you willing to spend on it?
ID: 1842470 · Report as offensive
rob smith Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer moderator
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 22160
Credit: 416,307,556
RAC: 380
United Kingdom
Message 1842475 - Posted: 15 Jan 2017, 20:56:53 UTC

Make sure you clean out the dust from any HD fan, and make sure it is working - hard disks need cooling just as much as CPUs and GPUs.
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 1842475 · Report as offensive
Sidewinder Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 15 Nov 09
Posts: 100
Credit: 79,432,465
RAC: 0
United States
Message 1842517 - Posted: 16 Jan 2017, 1:44:26 UTC
Last modified: 16 Jan 2017, 1:46:14 UTC

You could also just replace the hard drives with an SSD. They have a larger range of operating temperature and produce less heat.

Edit: Didn't see that part in Ageless' response, but I'm in agreement.
ID: 1842517 · Report as offensive
Profile Brent Norman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 2786
Credit: 685,657,289
RAC: 835
Canada
Message 1842534 - Posted: 16 Jan 2017, 4:53:03 UTC - in response to Message 1842467.  
Last modified: 16 Jan 2017, 5:26:34 UTC

A case with 120/140mm fans and a 2x120mm CPU cooler should cost you about $100 at newegg, well worth the money to cool things off.

Edit: Also try app_config for throttling:

Example
0.5 gpu
2.0 cpu

Shutting down 3 or 4 cores is better than throttling, and they can still work just fine for OS work.

EDIT2: Intel doesn't list the Tcore for this CPU but it has a TDP of 130W (significant) W3565
ID: 1842534 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 1842566 - Posted: 16 Jan 2017, 10:09:11 UTC - in response to Message 1842534.  

EDIT2: Intel doesn't list the Tcore for this CPU but it has a TDP of 130W (significant) W3565

I couldn't find one official either, but CPU Data states it at 68C (154.4F).
The comparison between a W3550 and W3570 also shows a max of 67.9C (154.22F), so I wouldn't be surprised if that was the actual max for this CPU.

In any case, 80C is much too much. Perhaps that the OP wants to check into his thermal paste as well, and read this thread where Ahnilated tested Liquid Metal paste and got the temperature of his overclocked i7 4790K down by 17C with just air cooling.
ID: 1842566 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1842597 - Posted: 16 Jan 2017, 13:31:16 UTC - in response to Message 1842566.  
Last modified: 16 Jan 2017, 14:05:23 UTC

I couldn't find one official either, but CPU Data states it at 68C (154.4F).
The comparison between a W3550 and W3570 also shows a max of 67.9C (154.22F), so I wouldn't be surprised if that was the actual max for this CPU.

In any case, 80C is much too much. Perhaps that the OP wants to check into his thermal paste as well, and read this thread where Ahnilated tested Liquid Metal paste and got the temperature of his overclocked i7 4790K down by 17C with just air cooling.


1) I misremembered the 80C degrees. I had it set at 70, have backed it down to 68C for the moment.

2) Thank you for the Liquid Metal Paste pointer! I now have some on order.
ID: 1842597 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1842599 - Posted: 16 Jan 2017, 13:34:50 UTC - in response to Message 1842534.  

I have learned the hard way that HP hardware is "wonky" and doesn't completely follow ATX-type standards. For instance, you can't just drop in a replacement power supply. It either has to be an HP brand or you have to figure out how to jump some of the wiring. :(

So a new case with fans probably won't cut it.

I have ordered the fan that is supposed to go in the "front" of the case.

Thanks,
Tom
ID: 1842599 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1842601 - Posted: 16 Jan 2017, 13:40:05 UTC - in response to Message 1842470.  

Well, as you already say, you need cooling. Lots of it, as I can see.
Probably best to buy a new case that can have loads of fans in it, perhaps get water cooling for the CPU?
Although doing a quick search on others running a W3565 with water cooling shows that even then the CPU can run at 80C. However, since the water cooling has the fan on the outside or top of the case, the hot air should be blown out and away from your hard drives.

You can also just opt for not using the CPU, or not use it on all cores. That may have impact on your RAC, but is a cheaper solution than having to buy a new HDD every year.

Having said that, it may also depend on the HDD.
E.g. WD Green can withstand way less heat or writes than a WD Blue or Black can do.
Or you may want to look out for SSDs, since these have no moving parts and can withstand heat a lot better.

How much money are you willing to spend on it?


I have seen an HP Z400 specific water cooler (used) for about $150. So that might make sense. Thank you for the comment about hard drives that tolerate heat more. I will look into that. That is a good point about reducing the number of CPUs I am using. On the other hand, it doesn't do much but Seti and a little file server work.

Thank you,
tom
A proud member of the OFA (Old Farts Association).
ID: 1842601 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1842604 - Posted: 16 Jan 2017, 14:03:36 UTC - in response to Message 1842470.  
Last modified: 16 Jan 2017, 14:04:07 UTC

You can also just opt for not using the CPU, or not use it on all cores. That may have an impact on your RAC, but is a cheaper solution than having to buy a new HDD every year.


I could shut down the CPU completely. I have just reset the local preferences to 50% of CPUs at 100% but the Win7 task manager is reporting significant activity across all CPU's. I could disable hyperthreading which would drop it to just 4 hardware CPU cores and see if it heats up less.

Oh, well, by my top of the head calculation(s) when the motherboard CPU's running flat out, provides about 1/3 the work unit processing of my GTX 750ti. So a good argument could be made for shutting down all the MB CPUs and upgrading the power supply (HP power supplies are non-standard) and adding a second GTX750ti. <groan>

I am going with the liquid metal thermal paste and the optional inside the box fan for the moment. I will keep the water cooler idea in the back of my head. I am too leery of the power supply upgrade to want to try it again. (Sent back 1 PS that should have worked except it wasn't an HP PS).

Thank you all for your input. Now if I can just figure out where I laid down my screwdriver.....

Tom
'
ID: 1842604 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1842706 - Posted: 17 Jan 2017, 3:07:18 UTC - in response to Message 1842534.  

Edit: Also try app_config for throttling:

Example
0.5 gpu
2.0 cpu

Shutting down 3 or 4 cores is better than throttling, and they can still work just fine for OS work.


Hi, I have two questions on this.

1) Where should the file be located? In the main Bonic data dir or in a Seti data dir?

2) What is the exact syntax for reducing the total # of cpus that Seti/Bonic can use?

Thanks,
Tom
A proud member of the OFA (Old Farts Association).
ID: 1842706 · Report as offensive
Profile Brent Norman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 2786
Credit: 685,657,289
RAC: 835
Canada
Message 1842717 - Posted: 17 Jan 2017, 3:53:48 UTC - in response to Message 1842706.  
Last modified: 17 Jan 2017, 3:54:31 UTC

File name is app_config.xml and is located in the setiathome.berkeley.edu folder.
Use a plain text editor i.e. Notepad to create it.

P.S. Yes, I believe you should shutdown multi threading.

This is an example.
<app_config>
	<app>
	<name>setiathome_v8</name>
	<gpu_versions>
		<gpu_usage>1.0</gpu_usage>
		<cpu_usage>2.0</cpu_usage>
	</gpu_versions>
	</app>

	<app>
	<name>astropulse_v7</name>
	<gpu_versions>
		<gpu_usage>1.0</gpu_usage>
		<cpu_usage>2.0</cpu_usage>
	</gpu_versions>
	</app>
</app_config>
ID: 1842717 · Report as offensive
Ianab
Volunteer tester

Send message
Joined: 11 Jun 08
Posts: 732
Credit: 20,635,586
RAC: 5
New Zealand
Message 1843413 - Posted: 21 Jan 2017, 4:21:40 UTC

I have ordered the fan that is supposed to go in the "front" of the case.


That was my thought, the hard disks "should" be sitting in the path of the fresh air coming INTO the case, and what happens behind that shouldn't really matter to them. They should be breathing fresh cool air no matter what.

The hard disks themselves do generally needs some airflow, or need to be bolted to some solid case metal to get rid of the heat they make themselves, I've seen problems with a stack of hard disks in a poorly vented external enclosure "cooking" themselves, when they were basically the only source of heat. Air was flowing into the case, then out the fans, but not actually circulating over the hot disks.

Problem was solved by removing the front blanking plates and spacing the individual disks so there was a free slot between them. Now the disks had free air flow around them, and where much happier.
ID: 1843413 · Report as offensive

Message boards : Number crunching : Xeon cpu is frying my hard disks


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.