Getting the duct repaired Wednesday. Hope that clears up the heat problem.

---

Intel is reportedly changing it's "lidding" process (How the metal cpu lid is fitted to the chip underneath) on the next generation HEDT chips.

Should be Good News for those who can afford them...uh, Not Here..:(

Lt

Hi Michel. I'm a Network Administrator with 15 year experience with computer hardware. I suspect the problem is that you are overheating because the I7 core runs very hot.

I'd recommend something like my rig:

Specs:

AuthenticAMD
AMD FX(tm)-8150 Eight-Core Processor [Family 21 Model 1 Stepping 2]
(8 processors) Overclocked to 3.87GHZ with Liquid Cooling

NVIDIA GeForce GTX 460 (1024MB) driver: 337.88 OpenCL: 1.1 Microsoft Windows 7
Ultimate x64 Edition, Service Pack 1, (06.01.7601.00)
32 GB 1333MHZ RAM

I also pulled these parts out of my old Sonata II case which offered really good cooling, but not enough. I would run 78 to 98 degrees Celsius, which is near max thermal for an AMD CPU.

I put my gear into a case that has top venting (and is full size), so heat naturally rises from one chassis fan that sits besides the liquid cooling fan assembly. Together, the cooling is excellent.

I've been running my 8 cores overclocked at 100% for the past two hours and I don't peak 48 degrees C.

Your rig looks pretty good. I would just get a cool liquid cooler and replace your current CPU Fan. It will change how stressed the cores get.

Good luck.

-Steve

I suspect the problem is that you are overheating because the I7 core runs very hot.

???

AMD have the most power hungry & hottest CPUs ever since the introduction of Intel's Core architecture.
The current Haswell CPUs aren't as well cooled as the previous series, but they are still use much less power, and produce much less heat, than the equivalent AMD CPUs.

---

Grant
Darwin NT

Hello,
I've recently got a new rig and try running seti@home on it, using the latest boinc currently available. I am getting high CPU temperatures where each core is reaching 100 degrees C after a maximum of half an hour (even if I lower max cpu to 70%).

I first suspected my stock fan, but Intel Extreme Tuning Utility stress test for half an hour didn't bring the cpu temps above 80 degrees C. So I'm starting to suspect that there's something wrong with the application itself.

Is there anyone else facing this issue? any ideas?

Specs:
- i7 4770k, stock fan, stock configuration
- MSI Z87 MPOWER MAX AC
- MSI GeForce Gaming N780 TF 3GD5/OC
- ThermalTake VN300M1W2N Chaser MK-I (3 default fans, back, top, front)
- Kingston HyperX Beast 16GB(2x8GB)
- windows 8.1 professional

NEVER EVER EVER EVER RUN A STOCK INTEL THERMAL SOLUTION!!!

They are made for office computing situations with very little utilization of the CPU.

Find out what your clearance is in your case for an aftermarket cooler, and buy the biggest one you can afford. Also, there are a lot of internal watercooling kits that are now all-in-one, and easy to install, and cheap (even the corsair 240mm one is under a hundred dollars).

---

Delidded i7-4790K (CLU/CLU) at 4.7GHz @ 1.310Vcore 24/7, 32GB DDR3-2400, Corsair H100i v2, Gigabyte Z97X-Gaming G1 WIFI-BK, MSI Radeon RX 480 Gaming 4G, HX-650 PSU, Corsair 750D

NEVER EVER EVER EVER RUN A STOCK INTEL THERMAL SOLUTION!!!

Got a C2D that's been crunching Seti 24/7/365 for almost 8 years in 35°c+ temperatures with no issues.
My i7 ran with stock cooling 24/7 for over 3 years before I put on an after market cooler to bring the temperatures down.

As long as the ambient temperature is mostly below 30°c and you've got a good case, there's no problem with the stock cooler even with 24/7/365 operation.
However if you have to deal with extended ambient temperatures above 30°c then it's worth using an after market cooler. For overclocking it's essential.

---

Grant
Darwin NT

NEVER EVER EVER EVER RUN A STOCK INTEL THERMAL SOLUTION!!!

Got a C2D that's been crunching Seti 24/7/365 for almost 8 years in 35°c+ temperatures with no issues.
My i7 ran with stock cooling 24/7 for over 3 years before I put on an after market cooler to bring the temperatures down.

As long as the ambient temperature is mostly below 30°c and you've got a good case, there's no problem with the stock cooler even with 24/7/365 operation.
However if you have to deal with extended ambient temperatures above 30°c then it's worth using an after market cooler. For overclocking it's essential.

Tried running my i7/4770K with Intel's stock cooler with ambient temps below 30c when it first came online 3 months ago, and it immediately went to high 80+c as soon as BOINC/SETI started running. When not crunching it fell to mid 50c. Didn't crunch for a month until I could get a Corsair H75 LCS and the temps now average in mid 60c with spikes to 82c every now & then. It can reach an average of mid 70c when ambient temps reach mid 30c. All of this using a Cooler Master HAF 32 full tower case. I've never been able to crunch without a LCS to keep the CPU temps below mid 60c, even in the winter on the east coast. Some profess fans as being great chilling the CPU, but I find them to be too bulky and I don't need the noise, so they are out of the question. If I could afford it, I'd put everything in 2 closed loop LCS, 1 for the CPU and the other for dual GPUs with large rads.

---

I don't buy computers, I build them!!

Tried running my i7/4770K with Intel's stock cooler with ambient temps below 30c when it first came online 3 months ago, and it immediately went to high 80+c as soon as BOINC/SETI started running. When not crunching it fell to mid 50c.

Even with a stock cooler, correctly fitted, that shouldn't happen. I'd expect low 40°s at idle, low to mid 70°s at most when under 100% load- especially with the ambient temps below 30°c.

It can reach an average of mid 70c when ambient temps reach mid 30c.

Way too high for a liquid cooled setup IMHO.
When I finally replaced my stock cooler with an Artic i30, with the ambient in the mid to high 30°s the CPU temp dropped from low 80°s down to around 70°c. I run my systems without the side case on- that drops temperatures by about 5°c.
The hottest it's been would have been around 38°c, and at that time the CPU temperature stayed below 75°c.

If I could afford it, I'd put everything in 2 closed loop LCS, 1 for the CPU and the other for dual GPUs with large rads.

That's the almost ideal setup. Not much point using coolant that's already been heated up significantly to cool the following devices in the loop.
Perfect one- one cooler & radiator for each device being cooled.

---

Grant
Darwin NT

Turning HT off lowered my temps by 10-15 C.

---

Mine is still running 24/7 on 100% with default configuration (x37 multiplier and HT on). Room temperature is now in the lower 30's (compared to 20's when this started). CPU temperature is usually in the 60's, with peaks in the mid 70's.

For those who turned off HT, what was the impact on performance/RAC?

I am considering some upgrades to my setup, mainly adding two graphics cards. I currently have a GTX 780, so I will be adding another for SLI. Which leaves me with a 3rd slot that I can fill, but not in SLI (motherboard doesn't support 3 way SLI, already regretting my choice).

The motherboard manual says that it has 3 x PCIe 3.0 x16 slots (support x16, x8/x8, x8/x4/x4 modes). Nvidia cards requires x8/x8 to work in SLI, does that mean if I add a 3rd one (not in SLI), it will affect the 2 in SLI? if so, then I will be limited to 2 780's.

If I am wrong, which card would you recommend as a 3rd one? It will be used for crunching only. Power is not an issue, I've got a 1275W PSU. Cooling might be an issue and price, since it is only for crunching.

In both cases, should I consider any other bottlenecks? i.e. CPU (it has a i7 4770K)?

---

Mine is still running 24/7 on 100% with default configuration (x37 multiplier and HT on). Room temperature is now in the lower 30's (compared to 20's when this started). CPU temperature is usually in the 60's, with peaks in the mid 70's.

For those who turned off HT, what was the impact on performance/RAC?

I am considering some upgrades to my setup, mainly adding two graphics cards. I currently have a GTX 780, so I will be adding another for SLI. Which leaves me with a 3rd slot that I can fill, but not in SLI (motherboard doesn't support 3 way SLI, already regretting my choice).

The motherboard manual says that it has 3 x PCIe 3.0 x16 slots (support x16, x8/x8, x8/x4/x4 modes). Nvidia cards requires x8/x8 to work in SLI, does that mean if I add a 3rd one (not in SLI), it will affect the 2 in SLI? if so, then I will be limited to 2 780's.

If I am wrong, which card would you recommend as a 3rd one? It will be used for crunching only. Power is not an issue, I've got a 1275W PSU. Cooling might be an issue and price, since it is only for crunching.

In both cases, should I consider any other bottlenecks? i.e. CPU (it has a i7 4770K)?

If you are a gamer just use 2 cards in SLI as the 3rd card will pretty much screw the SLI configuration up by the way your PCI-e lanes are shared.

If you just want to crunch then just install the 3 cards and forget about the SLI bridges.

If you go with 3 cards and all Nvidia then make sure that the 3rd 1 isn't a lot slower than the other 2 or you could run into elapse time issues.

Cheers.

If you are a gamer just use 2 cards in SLI as the 3rd card will pretty much screw the SLI configuration up by the way your PCI-e lanes are shared.

If you just want to crunch then just install the 3 cards and forget about the SLI bridges.

If you go with 3 cards and all Nvidia then make sure that the 3rd 1 isn't a lot slower than the other 2 or you could run into elapse time issues.

Cheers.

I am a gamer and this is mainly a gaming rig. I'll get another 780 and leave it at that... I missed this point when I considered this motherboard, lesson learned for the future.

---

The temperature is jumping a lot (and quickly) back and forth between 50 and 70 degrees C. The processors' load is also behaving the same way, between 15 & 60%. I am limiting SETI to 50% currently to keep the temps below 80 while it finishes the current tasks, then I will use it only for gaming till the problem is solved.

BOINC recommend lowering "use at most X% CPU time" to lower CPU temps (if you run resource monitor you will see all cores being turned off/parked for a % of time - causing "spikes" in all cores).

I set it to 50% (to get a temp within Tcase specs) but I found it completely unsatisfactory - like you mention the temps jump up and down (and the fan speed goes up and down with the temp driving me crazy).

I solved it by lowering the "on multiprocessor systems - use at most X%".

I set it to 50% (resource monitor shows half cores turned off - constant load on remaining cores - constant temperature, with no fan speeding up, then slowing down). Yes it runs half the cores at 100% but it keeps the temps/fan speeds steady ...

PLUS

It freed my other cores for GPU processing - doubling to tripling my GPU processing output (which more than makes up for halving my CPU usage when it comes to RAC).

Ps: Why does POEM@home use 1 CPU and 1 ATI GPU on my machine (giving me a max RAC of over 30,000) while SETI@Home uses 0.318 CPU and 1 ATI GPU (I only manage around 8,000-9,000 RAC). Half the CPU is available for GPU processing - so what gives ?

---

BOINC actually recommend TThrottle for controlling CPU temperatures on Windows.

Mine is still running 24/7 on 100% with default configuration (x37 multiplier and HT on). Room temperature is now in the lower 30's (compared to 20's when this started). CPU temperature is usually in the 60's, with peaks in the mid 70's.

I noticed that your GTX780 is similar to my GTX750Ti FTX in one respect, and that is it exhausts air into the case. Up to a few days ago I was pulling air from inside of the case getting the same temps that you are now getting. I reversed the fan direction and the CPU temps immediately dropped 10+c.

Another thing that I noticed is that your running you AP tasks using the r1316 version. I've joined the following thread that you might find interesting.

http://setiathome.berkeley.edu/forum_thread.php?id=75309

I've made the suggested changes for my AP tasks and couldn't be more satisfied, you might want to take a look at it.

---

I don't buy computers, I build them!!

I reversed the fan direction and the CPU temps immediately dropped 10+c.

Which fan? Side fan, front case fan, rear case fan, top case fan?

---

Grant
Darwin NT

BOINC actually recommend TThrottle for controlling CPU temperatures on Windows.

That's not what the official BOINC Help homepage says http://boinc.berkeley.edu/wiki/BOINC_Help

"BOINC may cause your computer to run hot. You can reduce this by limiting the percentage of time that BOINC is allowed to compute. This is one of many preferences that you can control."

So as far as I can see this is still the "official" BOINC advice/position re CPU overheating, perhaps it needs to be updated - because this isn't a satisfactory method for controlling CPU temperature for many users (as I said in a previous post I prefer to limit the number of cores, than percentage of CPU time).

Also, I don't like installing third party software to control basic functions of my PC (I'm not familiar with the program you linked to, perhaps you could provide a link to an official BOINC page advising this program as the recommended and "officially supported/recommended" way to control CPU temperature).

---

I've found the best way to control temperature is decent system cooling. No need for software.

---

Grant
Darwin NT

I've found the best way to control temperature is decent system cooling. No need for software.

Mate, you live in the tropics (less than 1400 Kilometres from the equator) you need decent system cooling just to browse the web ;-)

I'm in Brisbane (3000 kilometres from the equator) and I can tell you BOINC could have fried my CPU if I didn't install ASUS PC Probe (it ran at almost 90 degrees Celsius with stock cooling).

I think BOINC really needs to rethink their policy of enabling 100% CPU usage, on all cores, 24/7, as the default installation settings (the average PC user isn't aware of their CPU temperatures, runs stock cooling/fans, and installs BOINC with default settings - this could potentially damage the CPUs of these people).

At the very least, I think BOINC should inform the user during setup that using BOINC using default settings may pose a risk of increasing CPU temperatures above manufacturer maximum specifications (which could damage their CPU), then it should allow people to select how much CPU time they want BOINC to use, finally it should place a shortcut on the desktop that could include general information on how to monitor your CPU temperature.

I think BOINC has an ethical responsibility to make this information easily available, easy enough to understand by the average user, BEFORE installation is made possible. It should also provide options to change default CPU settings DURING installation of the BOINC client.

Oh, and by the way, when I installed a CPU monitoring program and found the temperatures excessive I right clicked the system tray icon and clicked exit (to stop BOINC and reduce the CPU temperature) my Windows 7 Resource monitor showed that it was still being used 100% - I had to start BOINC again, click "snooze" (to stop all running tasks) THEN exit BOINC before my CPU usage went down from 100% to less than 5%.

It seems that just exiting BOINC isn't enough to stop some projects from happily crunching away anyway - how bizarre !!!

I think BOINC should seriously address this CPU overheating issue - it's beyond a joke as far as I am concerned.

Ps: this post is probably more appropriate in a BOINC forum, not a SETI forum (because it's not SETI's fault).

Still, people should be aware that CPU overheating is a common, and potentially very damaging, problem with the BOINC client (one that I think BOINC should address as a priority).

---

I reversed the fan direction and the CPU temps immediately dropped 10+c.

Which fan? Side fan, front case fan, rear case fan, top case fan?

The mentioned fans are connected to the H75 LCS in a intake push/pull configuration. All other fans (3) are case - 230mm with the front & side as intakes and the top as exhaust. I did have a 120mm on the bottom that was used as an additional input supplying air for the GTX660SC, but I don't need it anymore.

@CZ - Have you thought of installing an LCS on your machine, it can do wonders for CPU cooling even where you're located.

---

I don't buy computers, I build them!!