Question CPU Test with Prime 95 - Intermittent system performance (Please, HELP!)

Sep 25, 2022
5
0
10
Recently I've been facing a really annoying problem in my computer. When I play a more demanding game like Call of Duty the avarage fps is around 90 in max settings but suddenly the frame rate drops to 5~8 for a few seconds. This makes the game unplayable so I decided to search for the real cause of this mistery.

Unfortunately it happens in all games not only COD.


My PC specs:

CPU: i5 11400 (Normal clock)
CPU Cooler: DEEPCOOL GAMMAXX 400 V2
GPU: RX 6600 (Normal clock)
RAM: 2x 8GB 2400MHz
MOBO: Biostar H510MH 2.0
PSU: MWE 600W COOLER MASTER
SSD: 480GB Kingstone
DRIVER: 22.5.1 Adrenalin Edition
OS: Windows 10 Pro

I ran a test for the CPU with Prime 95 and these are the results that I could notice:

1 - In Idle CPU Boost is ON (4.3 Ghz) but in LOAD is not at its max value.
2 - I don't know what TMPIN0 is but I realized that when it hits 92°c the whole CPU performance goes down drastically
3 - All temps except TMPIN0 are normal.


What should I do to solve this problem?

IDLE
idle2.png



LOAD
LOAD2.png
 
What should I do to solve this problem?
Run ThrottleStop.
https://www.techpowerup.com/download/techpowerup-throttlestop/

Open up the Limit Reasons window.

When your CPU slows down to a crawl, you are likely going to see something light up red under the CORE column of Limit Reasons. This will tell you why your CPU is throttling. As already mentioned, this throttling problem is likely being caused by some under designed or poorly cooled voltage regulators. Common throttling reasons are VR CURRENT, VR THERM when they are too hot or maybe you might see BD PROCHOT light up red when throttling.

Post lots of ThrottleStop screenshots including the FIVR and TPL windows when throttling if you need help.
 
  • Like
Reactions: LuqDeCastro
Sep 25, 2022
5
0
10
Run ThrottleStop.
https://www.techpowerup.com/download/techpowerup-throttlestop/

Open up the Limit Reasons window.

When your CPU slows down to a crawl, you are likely going to see something light up red under the CORE column of Limit Reasons. This will tell you why your CPU is throttling. As already mentioned, this throttling problem is likely being caused by some under designed or poorly cooled voltage regulators. Common throttling reasons are VR CURRENT, VR THERM when they are too hot or maybe you might see BD PROCHOT light up red when throttling.

Post lots of ThrottleStop screenshots including the FIVR and TPL windows when throttling if you need help.

I've done just like as you suggested here are the screenshots:

ThrottleStop window in idle:

ts-idle.png


ThrottleStop window in load before throttling:

ts-before-bd-TPL.png

ts-before-bd.png


ThrottleStop window in load while throttling:

ts-while-bd.png

ts-while2-bd-TPL.png


As you said, BD PROCHOT lights up when throttling.
Probably the VR is heating up because of the warm air that comes from the CPU cooler. Can I solve this issue using Throttlestop?


What do you think about installing heat sinks? Will the the VR temp cool?


s-l500.jpg
 
@LuqDeCastro

If the voltage regulators are causing a problem, you would usually see VR CURRENT or VR THERM lighting up red in Limit Reasons.

Try clearing the BD PROCHOT box on the main screen of ThrottleStop and see if this fixes your throttling problem. Your CPU temperature looks great. Your CPU will still be able to thermal throttle if it ever gets too hot whether BD PROCHOT is enabled or not. BD PROCHOT and PROCHOT (processor hot) are two different things.

To fix the power limit throttling problem go into the TPL window and check the MMIO Lock box near the top right.

Clear the Disable Controls box and set PL1 and PL2 to at least 129. Set the turbo time limit to the default 28 second value.

Do some more testing to check for any throttling. Watch the TMPIN0 sensor temperature while testing. You can try putting some heatsinks on the voltage regulators to see if that makes a difference or use an external fan to direct more air flow towards this area. Tower coolers are great for the CPU but do not direct enough airflow towards the motherboard like the cheap Intel OEM coolers do.
 
Last edited:
  • Like
Reactions: LuqDeCastro
Sep 25, 2022
5
0
10
It seems that the throttling stopped but the temp kept going higher so I decided to stop the tests.
Is there any risk to burn the mobo or will it shut down when it hits a specific temp?


After changing the settings:

TPL.png


Right before stopping the stress test:
TMPIN0 - 102° c


disabled.png
 
Is there any risk to burn the mobo
If that sensor is accurate and something on your motherboard is running at more than 100°C when you disable the throttling, then yes, it is possible that you could seriously damage or kill your motherboard.

I would plug in a small fan that you can move around. When the temps from that sensor get high, move the fan around and point it at different parts of the motherboard. See if this helps lower the temperature from that sensor. If you can find something that you can put a heatsink on, maybe that will help. You might have to permanently install a secondary fan.

You might have to try going back to the Intel OEM heatsink and fan. It is designed to direct airflow around the CPU socket. It would be interesting to see how hot this sensor gets when using the Intel heatsink and fan. Your CPU will run a lot hotter but maybe you will get better overall performance without having to worry about frying your motherboard.

I avoid running Prime95. It has become an excessive torture test. How hot does that sensor get when gaming when you are using ThrottleStop to disable the throttling?
 

Karadjgne

Titan
Ambassador
Everything in the pc connected to the motherboard has a physical address on a variety of buss systems. It's how the cpu keeps track of what is where, by that address.

Tmpin0-6 are reading those addresses. But have exactly no clue as to what each address belongs to, and wouldn't matter if it did, because they are different to another motherboard. And that changes with each generation, upgrades to pcie and usb, vendor etc.

So Tmpin0 could be the super i/o for you, but could be the PCH for me.

Best way to figure it out is with a IR camera. At 100°C it'll be very evident which component is cooking.

Most of the silicon components like chipsets and mosfets were generally rated at 125-150°C, although recent advances have put that upto 175°C and there are a few that are rated 200°C .

Doesn't mean typical design parameters like anything over 100°C.