Possible major BIOS bugs in HP Laptop

[NineD]

I have discovered a possible BIOS bug or other firmware issue that causes the HP Laptop 15-ef0023dx to lose performance after running under heavy load until the system is restarted. This system is 2 days old and is using the latest BIOS version F.17.

Software used for testing: HWiNFO64 v6.26, Geeks3D FurMark 1.21.1.0

Testing steps:

Test #1: Run FurMark CPU Burner (simulate heavy CPU workload without use of 3D GPU graphics)

Result: CPU cores 0-3 boost to 3.2 GHz and gradually reduce in speed as CPU heats up. Stop test, wait 3 minutes to allow CPU to cool. Run test again. CPU cores 0-3 again boost to 3.2 GHz and gradually reduce in speed as CPU heats up.

Conclusion: Normal behavior

Test #2: Run FurMark GPU Stress Test followed by FurMark CPU Burner (simulate use of 3D GPU graphics and heavy CPU workload separately)

Result: GPU runs as expected during GPU Stress Test, CPU cores 0-3 remain at about 1.8 GHz during GPU test. After stopping GPU test, run CPU test. CPU cores 0-3 boost to 3+ GHz, same as in test #1.

Conclusion: Normal behavior

Test #3: Run FurMark GPU Stress Test and FurMark CPU Burner at the same time (simulate heavy system load such as gaming)

Result: GPU test runs slower due to high CPU load. CPU cores 0-3 remain at about 1.8 GHz due to APU package TDP limit.

Conclusion: Normal behavior

Test #4: Run FurMark CPU burner after completing Test #3 and allowing 10 minutes for APU to cool

Result: CPU cores 0-1 remain at 1.8 GHz, CPU cores 2-3 remain at 2.7 GHz. CPU CAN NO LONGER OPERATE AT FULL SPEED.

Conclusion: Improper system operation possibly due to BIOS or firmware bug - the system should return to the behavior shown in Test #1 after the APU has cooled down.

After completing Test #3, the CPU cores will not run at normal speed until the system has been restarted. After restarting system, normal behavior is restored. The use of FurMark for testing was for convenience only. This behavior has been observed using other programs that produce high CPU load including real-world applications.

Secondary item: After running test #3, when the system goes to sleep many minutes after completion of the test and after the APU has cooled completely, the system will sometimes power off instead of going to sleep. When powering on, error 90D (thermal shutdown) is reported. It has been proven that there are no thermal problems with the laptop because all temperatures are within normal limits even when running test #3. This is also possibly caused by a BIOS or firmware bug.

As you can see from the details provided, this is not the type of issue that would be cleared by installing Windows updates, performing "hard reset", or other simplistic troubleshooting steps. Please provide any insight regarding this issue as soon as possible. If this is a known issue, please confirm a resolution plan. If I cannot obtain a resolution plan, I will be forced to return the laptop to the place of purchase within the next few days.

[Lil_Boy_Blue]

Hi

How do you know it is not a feature of the CPU architecture?

AMD Ryzen™ 5 3500U with Radeon™ Vega 8 Graphics

(2.1 GHz base clock, up to 3.7 GHz max boost clock, 6 MB cache, 4 cores)

[NineD]

I don't think this is a "feature" for the following reasons:

• No other mobile CPU/APU from AMD or Intel has a similar feature.
• This feature (2 cores crippled at minimum speed and 2 cores crippled at reduced speed until system is restarted) is not documented in any information about the CPU architecture.
• This would be an absolutely terrible feature and it's preposterous to think that any CPU designer would intentionally implement this behavior.

Maybe it's a bug in the CPU architecture, in which case the CPU needs to be recalled, but it's definitely not a feature.