Does this video even make sense?

[Valour549]

So searching for a go-to way of setting up TS for the i7-8750H comes up with this very popular video.

At the 9:15 mark he reduces the max Turbo of the processor from 4.1 to 3.4GHz.

If my computer game bottlenecks on the CPU, surely doing this will decrease FPS rather than increase FPS? While going with the settings in the video does result in a cooler temp, I imagine even with thermal throttling on the default turbo one would still get better FPS than using his settings?

 

[TheoneandonlyMrK]

So searching for a go-to way of setting up TS for the i7-8750H comes up with this very popular video.

At the 9:15 mark he reduces the max Turbo of the processor from 4.1 to 3.4GHz.

If my computer game bottlenecks on the CPU, surely doing this will decrease FPS rather than increase FPS? While going with the settings in the video does result in a cooler temp, I imagine even with thermal throttling on the default turbo one would still get better FPS than using his settings?

So click on a random link, you're going to struggle for biters.

 

[Valour549]

So click on a random link, you're going to struggle for biters.

I don't even know what that means, but it's cool I'll wait for someone that actually answers the questions.

 

[Athlonite]

next time you want to add a video to your post just copy the URL and use the insert media button and past the link

 

[unclewebb]

@Valour549 - If you need help setting up ThrottleStop, post some pictures of how you have the program setup now. If you look through the ThrottleStop thread here on TechPowerUp, you should be able to find a user or two with the same CPU that you have. I keep answering the same questions over and over again so you do not have to look too far to find some good advice and see some examples of appropriate settings for an 8750H.

Slowing the CPU down is what I would call a last resort. If your CPU runs hot, I would suggest blowing out the dust and replacing the thermal paste. For some games, a slower CPU will hardly be noticed so you might get away with that without the FPS dropping too much. I still prefer maximum performance.

 

[Valour549]

@unclewebb That's what I thought. I too prefer max performance. I did some research but it seems like everyone has different setup. Here's what I have so far. Please let me know if I need to change anything.

For Limit Reasons every time I start up it has EDP Other as well as PL1. But if I clear it and then game, it seems PL1 doesn't appear anymore. However, EDP Other is always there. I have an MSI GL73.

 

[unclewebb]

EDP OTHER under the RING column is not too important. You can try setting Power Limit 4 either sky high or setting it to 0 to see if this goes away. You can also try increasing the IccMax core and cache values to their maximum, 255.75.

-144 mV for the cache is on the edge of stability. If you see any BSODs or any errors in the TS Bench test, I would reduce this. Every CPU is unique. You might have a really good CPU that can run stable at that setting. In general, the majority of 8750H start showing errors when the cache is set beyond -125 mV.

No need to set the turbo time limit to more than 3 million seconds. I think the CPU is going to ignore that request. Setting this to the default 28 seconds should be OK. If your cooling can handle it, increase PL1 to 70W if you need to.

Can you run Cinebench R20 without any throttling? That is a good real world test.

MAXON Cinebench R20.0 Download

CINEBENCH is a real-world cross platform test suite that evaluates your computer's performance capabilities. CINEBENCH is based on MAXON's award-winn

www.techpowerup.com

The CPU core does not have to be set exactly double of the CPU cache offset. Some users think a difference of -100 mV is the maximum. A little too high is not really a problem. Any extra will automatically be ignored by the CPU. If you do not see an improvement in temperatures or performance when Cinebench testing when the core goes over -225 mV then there is no need to set the core offset to -289 mV.

Overall your settings look great. You have done your homework. Run a ThrottleStop log file if you want me to have a look at it.

 

[Valour549]

I haven't changed any settings (tried all combinations of Power Limit 4 and IccMax values as suggested, nothing removed EDP Other). No BSOD when running TS Bench or Cinebench 20. Just yesterday I completely cleaned both fans and did a repaste of both CPU/GPU with Coolmaster Mastergel Maker.

As usual I cleared the Limit Reasons PL1 on startup, then I ran the tests:
1) Ran TS Bench twice.
2) Then ran Cinebench 20 twice.
3) Let it cool to ambient temp.

No PL1 showing after running either test, but the CPU temp did hit max of 90 °C with max 5.1W as shown. Not sure why the temp still hits 90 even after my thorough cleaning and repaste yesterday. Log file of above tests attached.

Stress test with AIDA64 for twenty minutes causes temp to hit a max of 97 °C and thermal throttling of max 15%, though it does eventually stabilize so the throttling is only about 5% or less. I wonder if this is a definitive indicator that I screwed up the pasting somehow (line method). The temp in my room is about 27 °C as reference.

But that is stress test, I don't think games will do that right? Well I could always be checking Limits after a session to make sure it hasn't happened.

 

[unclewebb]

nothing removed EDP Other

EDP OTHER under the RING column is hyper sensitive on a lot of CPUs. The CORE column is the important one. I would not worry about only EDP OTHER lighting up yellow. I would worry about THERMAL lighting up yellow under the CORE column. That and PROCHOT 95°C turning red indicate thermal throttling.

wonder if this is a definitive indicator that I screwed up the pasting somehow

I hate to say it but you might be right. Some users need to replace the paste multiple times so they can learn what works best. Instead of the line method, maybe a thin even layer will work better.

temp did hit max of 90 °C with max 5.1W

That power consumption data is not correct. Intel CPUs use shared monitoring timers so if you were running ThrottleStop and AIDA64 at the same time, it is possible that they were interfering with each other. Try running only ThrottleStop for monitoring. Run a TS Bench test and see if power consumption is reported correctly. Let me know if this still does not look right. There might be a problem with the high performance timer that ThrottleStop uses.

Stressing the CPU and FPU in AIDA64 is kind of excessive for most laptops. Go play some games instead. Most users are happy if their games play smoothly without any throttling. If you need to crunch data and run both the CPU and GPU at full load for extended periods of time, getting a well cooled desktop computer would be a good idea.

 

[Valour549]

So I decided to repaste again, this time being careful to use less than the first time.

The first thing I noticed after starting TS is that now EDP OTHER is showing for all three columns: Core, GPU, Ring. Clearing it again does nothing. IDK why all three are showing this now, but if it doesn't really matter than I guess I need not worry.

However, after running the tests and doing another clear, EDP OTHER only shows for Ring now.

Running the tests I pretty much get identical max Throttling and Temp values (max Watt is now 13.9 though)

 

[unclewebb]

@Valour549 - I said that EDP OTHER showing yellow only under the RING column is not important. When it shows up across all 3 domains at the same time, that is important. It is not unusual for some of these boxes to light up while the CPU is booting up. I always suggest that you push the Clear button in Limit Reasons before you do any stress testing. If these 3 boxes light up red while stress testing, that shows that throttling is in progress. This can usually be fixed by adjusting the core and cache IccMax values higher as well as Power Limit 4.

Does other monitoring software report your CPU power consumption correctly? What laptop model do you have? Did you make any changes in the BIOS to IMON slope or IMON offset? Adjusting either of these can change the accuracy of the power consumption data. The manufacturer might have done this deliberately to avoid any power limit throttling issues. It is usually either that or maybe the high performance Windows timer (HPET) is not being properly calibrated when you boot up.