CPU issues

[Adrian_ae]

Recently I've noticed alot of stutters during heavy to light gaming alot during youtube media playbacks usually during transitions between videos and at the beginning, my settings were optimal adjustments as follows; All core ratio of 47 and a negative offset of -0120 and xmp enabled everything left to default, I ran with these settings for years no issues which puts me into the situation where I did a bit of research and came a term which aligns with my symptoms feel free to clarify I'm not an expert but i do have a bit of advanced knowledge and experience; clock stretching. My core clocks reported via HWinfo are 4700MHz but when I look at my effective clocks they are around 100 - 800mhz there's a significant gap this is observed during light to medium gaming or general workloads the effective clock are very low and inconsistent. The only time its at level with the core ratio is only during Cinebench rendering test it runs fine. I've tried defaulting bios settings, updating the latest bios firmware, chipset, reseating the cpu, connectors, ram, fresh windows install, I just can't wrap my head around it any help would greatly be appreciated. i will provide screen shot of the behavior with both cinebench test and in windows navigation. Any additional information required feel free to reach out thank you.

 

I forgot to mention my cpu cooler which is the NH-D15 in a Corsair Air 540 i hope this can be useful.

[leclod]

How are your cpu/gpu temps when stutter arise ?

 

 

[Adrian_ae]

1 hour ago, leclod said:

How are your cpu/gpu temps when stutter arise ?

 

 

They're low the cpu max temp during a stress test is around mid 70s across each core, I've ruled out cooling problems as the temperatures dont cross 80 degrees celsius

[unclewebb]

@Adrian_ae

The Effective Clock speed is only meaningful when a core is fully loaded 100%. At that point, the Effective Clock speed should be equal to the Core Clock speed. Other than that, I would not use the Effective Clock data for any other purpose. 

 

Cinebench should load all the cores. The Effective Clock and Core Clock should be the same during this test. 

 

4 hours ago, Adrian_ae said:

clock stretching

Most clock stretching issues were caused by Windows Defender using the system timers. This interfered with HWiNFO correctly calculating the Effective Clock speed. If you have this problem during Cinebench, run ThrottleStop 9.6 and the clock stretching problem should go away. 

 

4 hours ago, Adrian_ae said:

All core ratio of 47

Both CPU-Z and HWiNFO report a maximum multiplier of 44. It looks like the BIOS is not setup correctly or is not setting your turbo ratios correctly. Try running ThrottleStop and post a screenshot of the FIVR window. Also post a TPL screenshot so I can see how your turbo power limits are set. 

[Adrian_ae]

23 hours ago, unclewebb said:

@Adrian_ae

The Effective Clock speed is only meaningful when a core is fully loaded 100%. At that point, the Effective Clock speed should be equal to the Core Clock speed. Other than that, I would not use the Effective Clock data for any other purpose. 

 

Cinebench should load all the cores. The Effective Clock and Core Clock should be the same during this test. 

 

Most clock stretching issues were caused by Windows Defender using the system timers. This interfered with HWiNFO correctly calculating the Effective Clock speed. If you have this problem during Cinebench, run ThrottleStop 9.6 and the clock stretching problem should go away. 

 

Both CPU-Z and HWiNFO report a maximum multiplier of 44. It looks like the BIOS is not setup correctly or is not setting your turbo ratios correctly. Try running ThrottleStop and post a screenshot of the FIVR window. Also post a TPL screenshot so I can see how your turbo power limits are set. 

Understood so i shouldnt expect the effective clocks during gaming sessions, neither light work loads.

I defaulted the settings as well as enabled intel default settings thats why the multiplier showed x44, i appreciate your feedback I'll download throttlestop and post the relevant info

[unclewebb]

1 hour ago, Adrian_ae said:

i shouldnt expect the effective clocks during gaming sessions

The Effective Clock is mostly a measure of how loaded a CPU core is. It would have made sense for HWiNFO to report this as a percentage instead of a MHz value.

 

When lightly loaded, it will read less than the maximum clock speed. When a core is fully loaded, it should show the same value as what the Core Clock value shows. It is mostly a meaningless number except when all cores of a CPU are 100% fully loaded.