Unusual activity detected in Task Manager during Prime95 execution on an Intel Core i7-870 processor
Unusual activity detected in Task Manager during Prime95 execution on an Intel Core i7-870 processor
The discussion involves considerations about airflow and fan placement for a PC setup. The user mentions using a specific case and notes that direct airflow to the motherboard's top isn't ideal. They also ask about installing an additional fan above the motherboard and whether it should be part of the CPU cooler's fans, suggesting a potential trade-off in power management.
I understand your position.
Additionally, the temperatures shown by the motherboard sensors are averages; each core might be operating significantly hotter. You can check individual core temperatures using Speedfan or CPUID HWMonitor.
Increasing speeds is only beneficial for short bursts, as continuous operation will lead to overheating unless you enhance the heatsinking design.
Thank you for the recommendation of HWMonitor. I'll look into it.
My heatsink performance seems adequate, as it's rated for up to 210W and supports a maximum TDP of 320W. Would adding a modest boost clock around 200-300MHz from the CPU's base frequency of 2.93GHz increase its TDP significantly enough to cause throttling? I'm curious—given that the CPU doesn't even show thermal throttling in Prime95 and has a Tj-Max of 40°C, do you think it would really overheat?
The suggestion of using HWMonitor was appreciated, and I plan to investigate further. My heatsink performance seems adequate, as it supports up to 210W and a maximum TDP of 320W. A modest increase in clock speed—around 200-300 MHz from the CPU’s base frequency of 2.93GHz—would raise the CPU’s TDP significantly. This could potentially cause throttling, especially if the thermal paste or pad isn’t optimal, though the CPU itself doesn’t show thermal throttling in Prime95 and maintains a Tj-Max of 40°C. It’s unlikely the processor will overheat under normal conditions.
I chose to inspect the BIOS settings and attempted to disable the "CPU TM Function." The result was consistent—CPU usage remained a steady 100% at 3.2GHz during the 32-minute Prime95 Small FFT stress test. That’s impressive.
To investigate potential VRM problems, I positioned my infrared thermometer near several key areas: the large passive VRM heatsink directly behind the CPU, adjacent to the rear case fan and motherboard I/O, and at the top section of the motherboard above the CPU, between the RAM slots and the CPU power connector. This is where the situation became intriguing.
I proceeded to overclock the CPU to three different speeds: 2.93 GHz (24x clock speed; 123 BCLK), 3.2GHz (24x with 133 BCLK), and 3.6GHz (24x with 150 BCLK). Throughout all overclocks, the VRM heatsink maintained a stable temperature around 40°C. Not great, but acceptable.
Nevertheless, the top section of the motherboard displayed significant temperature fluctuations depending on the clock speed. At "stock" settings, temperatures rose from roughly 30°C to nearly 70°C when moving from the RAM slots to the CPU power connector. At 3.2GHz, the rise reached 30°C to 80°C, and at 3.6GHz, it jumped from 40°C to 90°C. The most concerning aspect was that after several minutes of Prime95 testing, the system abruptly shut down—likely due to overheating components on the motherboard.
In summary, my conclusion is that the CPU throttling isn’t caused by the CPU itself or the VRMs (especially those under the large passive heatsink), but rather by the overall temperatures of other motherboard parts. If this theory holds, I’d appreciate confirmation from someone familiar with similar setups.
This also raises several questions: Where are these temperatures coming from? Why isn’t the motherboard displaying such sensors to the operating system? If it’s not receiving data from sensors, where does it get the information to decide throttling? And if, under "CPU TM Function" enabled, the CPU would naturally slow down at a specific clock, wouldn’t it be prudent to run the CPU at higher frequencies for extended periods—without risking damage to the board or other components?
Thank you for your help in clarifying this.
the_ultra_code :
Bub, HWMonitor displayed identical results to HWInfo. It seems HWMonitor functions as a streamlined alternative to HWInfo.
OK, I've heard about stability problems with HWinfo; I definitely faced crashes there last year.
Regarding the temperature readings, could there be an issue with airflow in that zone? Occasionally, two fans might counteract each other, forming a stagnant area where temperatures rise significantly.
The cooler zones on either side may not receive any airflow at all.
I'm using a Fractal Design Focus G Mini as my PC case, meaning the top of the front fan is where the CPU cooler's fans are mounted. So, no direct airflow reaches the top part of the motherboard. Perhaps adding another fan above the board could help, but that would require connecting it to the CPU cooler's fans and possibly turning off one header. Not ideal.
Regarding your questions, do you have any possible answers you could share?