Does my Core Ultra 7 performance test mean something is wrong?

November 12, 2026 3 answers
Asked by 4J~Claskin

I upgraded to a Core Ultra 7 265KF for 4K video editing and gaming. PassMark gave me a CPU comparison that made me wonder if something is wrong, and I’m seeing 3.7 GHz rather than the advertised boost speed.

I was thinking about limiting it to 4 to 4.5 GHz for safety or adjusting the clocks. Is that a good idea, and what should I check first?

Answers

F2__Cromley

I wouldn’t cap the clock or overclock it yet. The advertised boost is a maximum under suitable conditions, not a speed every core should hold all the time. A reading of 3.7 GHz by itself doesn’t establish a problem.

Leave the CPU settings at stock while you investigate. Check temperatures, power draw and clock speeds during the test, then compare the actual PassMark score with results for the same 265KF. A screenshot or the numerical score would be more useful than the comparison label.

x2__Jorvex:

Also check whether that 3.7 GHz reading was taken at idle, during a single core test, or under a full CPU load. Those can look quite different.

Reply to this answer
Q5__Dornick

Core count alone is a poor way to judge that comparison. The 265KF has different types of cores, and a benchmark score depends on how the test uses them. The Ryzen model name in the comparison may also be worth checking, since the description could be mixing up model numbers.

Post the exact score and the name of the test. Without those, there isn’t enough to say whether the result is normal or unusually low.

Reply to this answer
v5__Jexlin

For a machine meant to render for years, I’d start with stock settings and make sure the cooling is doing its job under a sustained load. Fan curves help, but check the CPU temperature too. If it gets too hot and reduces its speed, look at cooler mounting and airflow before changing clocks.

If lower heat or power use is the goal, power or thermal limits are options. Undervolting can also reduce heat, but it needs stability testing, especially before trusting the system with long renders.

6T~Bexford:

Test any changes with the editing and rendering work you actually plan to do, not just a quick benchmark. A setting that completes a short test can still fail on a longer job.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions