Horrible CPU performance

[c33raz]

The Laptop was advertised ( at a retailer where i bought it), as being able to run when under loads at up to 3.4Ghz, however when running some games such as league and testing with some programs i saw that the cpu has an abyssmally bad performance, dropping down to 2.1Ghz. When under heavier loads such as multitasking it dropped under the specified 1.6Ghz frequency to 1.3Ghz.

My 5 year old laptop shows better performance in games and in benchmarking applications than this brand new one.

The laptop is less than a few weeks old and all the drivers and updates were installed. there are no issues when using diagnostic tool from hp and care center. The temp never rise to critical levels (max 80°C) and it has good ventilation.

Is it designed to have such poor performance, to the point of working below default frequency or is this a problem with the laptop 

[Huffer]

The picture is very hard to read. What is that application? "Trial Version"?? The built in task manager does a good job of showing CPU frequency history and CPU-Z works well, too. It will drop under the base frequency in power saving mode and you will very rearely see the max 3.4 ghz. Have you actually benchmarked the performance of the CPU in terms of its actual ability to crunch zeroes and ones? Clock speed is only a rough proxy for CPU performance. You need to run something like passmark cpu test:

This is my computer...app is a free download

You need to compare actual performance vs. others with an i5-8250U. If your 5 year old laptop is an i7 Quad Core then you may be right that is performs better than this new computer. Your i5-8250U would be expected to score about 7670 on the same test and I am going to make a small wager that if you run the same test on your 5 year old laptop the new laptop will outscore it. 8th gen Intel Core CPUs have a pretty good track record of performance and if yours is underscoring the norm then we know there is a problem with the CPU. 

Post back with any other questions. 

[c33raz]

I'll post test numbers of both laptops hopefully tomorrow or next monday, since I  sent the laptop to a service center.

The software in the background is aida64 free version, running stress tests. Monitoring software is HWmonitor i also used CpuZ to monitor, but HWmonitor shows more stats which I've found much more helpful.

I did run some other tests before such Cinebench and passmarks' software, however I don't have the results saved on my phone.

Kind Regards

[Huffer]

When you have numbers report back, please. 

[c33raz]

Allright it took me a bit longer than expected, I was busy the entire day yesterday so I couldn't pick the laptop up.

Here's the results, but first some notes:

-Asus was tested at home, HP was tested in a dorm

-Both Laptops were tested on a wooden desk with open air paths

-Results are in the form of screenshots of benchmark results

Here's the ASUS benchmarks:

CPU: Core i7-3630QM @ 2.4 Ghz

Cinebench score:

-579 CPU

-114 CPU (Single Core)

-5.09x MP ratio

CPUMark score (average over 10 runs):

-8502

Heavy Prime95 test:

-Package temp: 82 °C

-Clocks: @2.9 - 3.0 GHz

Aida64 stability test:

-Average clocks @ 3.0Ghz

HP benchmarks:

CPU: Core i5-8250U @ 1.60 Ghz

Cinebench score:

-465 CPU

-147 CPU (Single Core)

-3.16x MP ratio

CPUMark score (average over 10 runs):

-7553

Heavy Prime95 test:

-Package temp: 60 °C

-Clocks: @1.8 GHz

Aida64 stability test:

-Average clocks @ 1.3 Ghz

Example image of ASUS and HP run:ASUS run no. 1HP run no. 1

LINK:

Here's the link to the images and excel spreadsheets where i compiled the results, there is also some other images of other tests.

https://drive.google.com/drive/folders/1Ok4b0KACc9McsLYXL4FGzuTILaXIJkMH?usp=sharing

NOTES:

Asus runs were consistent throught the testing. The First CPUMark run was done WARM. meaning the Laptop already ran a trial test before. HP did not have that, the first test was done COLD, no other test were running before.

Both laptops ran the tests right after Restart, minimazing the impact other applications and miscelanious bloat would have on the tests.

Asus laptop's fans ran synchronized with the temperature (rising more or less linearly) , higher the temps. louder the fans. This was not the case with the HP laptop however where the fans spun up really fast for a few seconds and dropped down right after back to a middle level (exponential rise and drop).

CPU was not shown to be Throttling on the HP in the tests. However after starting up a test or loading the CPU the Frequency would stay at 3.4 Ghz, the power would max out at around 45W. This was for a few seconds, after that the power drops down to 15W and stays there. The CPU temps drop from 85 °C down to around 60  °C and stay there throught the period. The frequency also drops down to around around 2.2 Ghz.

From my own understanding it would seem that the CPU is Thermal throttling which is being controlled by HP software, either that or limited power delivery.

The HP AIDA64 stability test was extreme, only showing the most extreme of situation where the load is on the CPU, iGPU, GPU and CACHE. This result goes way beyond realistic load so it can be ignored.

TLDR;

All the nice graphs and images are under the Images and here:

https://drive.google.com/drive/folders/1Ok4b0KACc9McsLYXL4FGzuTILaXIJkMH?usp=sharing

For any questions about the tests, custom benchmarks or the laptops feel free to contact, I'll try to reply as soon as i can.

Kind Regards

Raz

[c33raz]

Just a quick followup

My friend has a similar laptop with the same CPU, his scores on CPUMark were about the same as my 5 year old laptop.

[Huffer]

Exactly as I predicted. Your old laptop is an i7 Quad Core and your new one is an i5, also a Quad Core but lower clock speed. The tests look to be about what would be expected. 

[c33raz]

Interesting. 

Although a bit displeasing, seeing my local vendor advertised this laptop as a gaming laptop. I expected more when they told me in the store that it easily outperform my old one.

It's also interesting that while the tests were running the CPU never throttled, would that be due to HP's custom bios settings and overheating protection?

It also seems that on this laptop the cooling solution has a hard time keeping up with TDP over 30W, since it ran at stable 2.1 Ghz at 60 °C at 15W TDP.

Makes me wonder how the design of this machine went. Putting in a 1050M a very capable GPU with this CPU and limiting it. Doesn't make much sense.

It's also interesting that when i searched for the results for this CPU there was a huge difference between scores in various benchmarks. I suppose that I should've paid more attention to that and lack of benchmarks for this model.

Any suggestions about possible fixes?

Maybe custom bios?

Kind Regards

Raz

[elus]

---

@Huffer wrote:

Exactly as I predicted. Your old laptop is an i7 Quad Core and your new one is an i5, also a Quad Core but lower clock speed. The tests look to be about what would be expected. 

---

Core i7-3630QM Turbo boost speed = 3.4Ghz. Core i5-8250u Turbo boost speed = 3.4Ghz.

3.4Ghz = 3.4Ghz.

But 5 year old ASUS with core i7 can hold high clock speed under load in games or stress tests, because this one have Power limit 1 (Long duration power limit) = 45w. HP Laptop have Power limit(PL) 1 = 15w.

P=f*C*V^2, where "P" is power, "C" is capacitance, "f" is frequency, and "V" is voltage.

So with PL1 = 15w CPU cant hold full speed specified by Intel. But Intel CPUs never limit self performance depending on power. Limitations based on other things.

I know several systems which control Power limit settings:

1.  OS. For example Windows 10. There is no problem to find some articles and turn it off.
2.  BIOS. I have AMI Bios and advanced setting are hidden by HP manifactures. But we can dump current bios image by special software and view advanced settings (CPU - Power Management Control). I used AMIBCP 5 and in my case there are no power limitations in the BIOS.
3.  Intel Management Engine. Its used to control a lot of basic functions of laptop such as sleep states, fan speed and others. It is based on the Intel Quark x86-based 32-bit coprocessor. And it is uncharted territory for me.

Ok, let's go back a bit. Power limit based on TDP(thermal design power). TDP is the maximum amount of heat generated by a computer chip or component that the cooling system in a computer is designed to dissipate under any workload. Actually, its recommended PL1.

Core i5-8250u specification :

• TDP : 15 W;
• Configurable TDP-up : 25 W;
• Configurable TDP-down : 10 W.

25w! Thats why PL1 must be customizable. Every single cooling system is differs from others (f/e i can buy cooling pad!). And advanced users must be able to change PL1 to find balance between performance and heat generation. Moreover, no one has canceled thermal throttling, which will not allow the cpu to overheat too much. Different manufacturers allow users to castomize PL1 in special programs (Intel XTU, ThrottleStop), but HP doesn't. Its so sadly. I bought expensive laptop with such performance as in twice chipper laptop. HP! Community needs help! Can you hear and add custom BIOS/IME firmware?

Vlad Kaminsky, Belarus.