CPU impact calculations

[consithe1]

Hello!

Going through the documentation of BoaviztAPI, I was trying to reproduce the impact calculations for the CPU. However, it seems like I am stuck in different ways and I was wondering if you could help me with it please.

1. In the documentation, the die_size_per_core corresponds to the number of cores divided by the die size.
   However, in the code, it seems like it's the other way around: die_size divided by the number of cores which is logic when you think about the variable name.
   My problem is unfortunately the following one: doesn't matter which formula you take, the die_size_per_core value still never fits with the results you find
   For example, if we take a die_size of 694 mm², a number of cores of 28. By applying the formula, we get a die_size_per_core of around 25mm². Why does your data indicates a die_size_per_core of 0.248 in this case please?

Best regards

[da-ekchajzer]

Hello,

Thank you for opening this discussion.

In the documentation, the die_size_per_core corresponds to the number of cores divided by the die size.
However, in the code, it seems like it's the other way around: die_size divided by the number of cores which is logic when you think about the variable name.

There is a typo in the documentation that we just corrected. Thank you for your report.

Commit : e659191

My problem is unfortunately the following one: doesn't matter which formula you take, the die_size_per_core value still never fits with the results you find
For example, if we take a die_size of 694 mm², a number of cores of 28. By applying the formula, we get a die_size_per_core of around 25mm². Why does your data indicates a die_size_per_core of 0.248 in this case please?

I just found that the cpu internal data in the API is not using mm2, but cm2. This is why you have the issue.

In https://github.com/Boavizta/boaviztapi/blob/main/boaviztapi/data/components/cpu_manufacture.csv, die_size column is in mm2 and die_size_per_core is in cm2

In the case you described : 694/28 = 24.8 mm2 ≈ 0.248 cm2 .

Since few users provide directly the die_size but use the pre-recorded die_size (which are in cm2) we just found out. Thank you for raising the issue.

I just updated the documentation and unit in the API response. You should now provide your die_size in cm2 (/100) and not in mm2.

The next version of the API (v1 which is still in development) uses only mm2 so this error will no longer occur.

Commit : f1b8096

Could you test and tell us if using cm2 fixes your issue ?

You are also welcome to join our public chat to tell us about your use case : https://chat.boavizta.org/signup_user_complete/?id=97a1cpe35by49jdc66ej7ktrjc

[consithe1]

Thank you so much for your reply and answers! I feel a bit stupid to not have thought about the fact that it was in cm² and not in mm² indeed!

I will test as soon as possible to tell you if it fixes the issue.
I might have another question for the CPU energy consumption impact, shoud I create another thread or should I write it here as well?

Good luck with this project and best regards

[da-ekchajzer]

Don't feel stupid for a poor documentation on our side ;).

For your other question, you could :

• Open an issue : https://github.com/Boavizta/boaviztapi/issues
• Open a discussion
• Exchange with us in our public chat : https://chat.boavizta.org/signup_user_complete/?id=97a1cpe35by49jdc66ej7ktrjc