Question Is there any possible issues with my idea.

[ImmmCanadian]

Hey all!

I am currently working on an essay which will look at whether it is more important to increase CPU power, or if it saves more time to improve the efficiency of the algorithm itself. In testing this I will need to underclock my cpu and disable some cores in order to test how fast it runs at 1 core, ___ ghz. Is there anything that I should be wary of in terms of breaking hardware or something going wrong, thanks for any help.

 

[CountMike]

Hey all!

I am currently working on an essay which will look at whether it is more important to increase CPU power, or if it saves more time to improve the efficiency of the algorithm itself. In testing this I will need to underclock my cpu and disable some cores in order to test how fast it runs at 1 core, ___ ghz. Is there anything that I should be wary of in terms of breaking hardware or something going wrong, thanks for any help.

Which CPU are you talking about ? Not many of today's are able to drop to only one core.
You can't increase CPU "power" just it's voltage. Power used by CPU will be in accordance to voltage supplied and load.

 

[NightHawkRMX]

I have seen somone make a single core 9900k.

If you have a good bios it can be done.

 

[ImmmCanadian]

Which CPU are you talking about ? Not many of today's are able to drop to only one core.
You can't increase CPU "power" just it's voltage. Power used by CPU will be in accordance to voltage supplied and load.

Sorry for the late response. I have a Ryzen 5 1600. My bad if I misunderstood the process of overclocking and underclocking however it shouldn't affect my idea of measuring how fast the algorithm runs for different clock speeds correct?

Edit: Also the single core was an example,.I would like to test it using a single core however if it isn't possible I will simply just go to the smallest amount of cores I can use.

 

[ImmmCanadian]

I have seen somone make a single core 9900k.

If you have a good bios it can be done.

Do you have a link to the thread by any chance?

 

[CountMike]

Sorry for the late response. I have a Ryzen 5 1600. My bad if I misunderstood the process of overclocking and underclocking however it shouldn't affect my idea of measuring how fast the algorithm runs for different clock speeds correct?

Edit: Also the single core was an example,.I would like to test it using a single core however if it isn't possible I will simply just go to the smallest amount of cores I can use.

HWinfo https://www.hwinfo.com/download/ can give you all you need per core, Volts, Amps and Watts. Bios should be able to get 1600 down to 3cores/6threads.
One thing I can say that may be wrong with your idea is that no two cores are same even when they are in same package and unless you can check cores one by one you may get different results with each one.

 

[ImmmCanadian]

HWinfo https://www.hwinfo.com/download/ can give you all you need per core, Volts, Amps and Watts. Bios should be able to get 1600 down to 3cores/6threads.
One thing I can say that may be wrong with your idea is that no two cores are same even when they are in same package and unless you can check cores one by one you may get different results with each one.

Thanks for the quick response but I'm not sure how the cores can be different even after a quick google search, can you elaborate further. Do you mean that not each core is always going to be the same so this will cause an uncertainty in my data, but won't ruin my entire experiment . Also is using MSConfig in order lower the amount of processors the same as what I am trying to achieve? Thanks as always.

 

[CountMike]

Thanks for the quick response but I'm not sure how the cores can be different even after a quick google search, can you elaborate further. Do you mean that not each core is always going to be the same so this will cause an uncertainty in my data, but won't ruin my entire experiment . Also is using MSConfig in order lower the amount of processors the same as what I am trying to achieve? Thanks as always.

Ryzen Master for instance, can detect better cores which could be up to 10% better than lowest rated core.
In my CPU for instance, those are 02, Gold, 03, Silver and 07, Bronze rated. In theory, OS and programs would prefer those cores in that order.
Run a benchmark or a test and look at HWinfo, you will see that some cores would boost higher than others.

 

[ImmmCanadian]

Ryzen Master for instance, can detect better cores which could be up to 10% better than lowest rated core.
In my CPU for instance, those are 02, Gold, 03, Silver and 07, Bronze rated. In theory, OS and programs would prefer those cores in that order.
Run a benchmark or a test and look at HWinfo, you will see that some cores would boost higher than others.

Ok, thanks for bringing this up, I will look into this and include it in the uncertainty of my experiment.