[PSR] PSR Cache memory benchmark

[klapajar at fel.cvut.cz]

Hello,
    let me highlight the 's' in 'arrays' (upload system responds that  
the arrays doesn't look random). In test 4, multiple arrays are  
generated using 'ranArray' to determine, whether the function isn't  
just a predefined shifting (e.g. staticly saved randomization).
    For a proper solution, it is more like a matter of probability,  
but the odds of getting a similar ordering are quite low.
    In your case, you are defining around 15 possible solutions (i.e.  
not so random, especially in contrast to n!), and, therefore, reducing  
the chance of passing through the test.
    I would recommend you to add some random variable, and probably  
use 'srand()' to mix with the arrays a 'little bit more'. (Last part  
can is relevant for everybody.)

Sincerely
Jaroslav Klapálek

Cituji kasltoma at fel.cvut.cz:

> Hello,
>
> thank you for the note of wrong size of the structure, I missed that.
>
> After setting it's size to correct value, it seems to pass without  
> any issues.
>
> However, now I have a similar problem with the second function,  
> 'ranArray', where Brute says the layout of the array is not random,  
> even though the access pattern quite corresponds to a random one and  
> it is different on each function call.
>
> How much more random can it be?
>
> Thank you again very much,
> have a nice day,
> Tomas Kasl - kasltoma
>
>
> _______________________________________________
> PSR mailing list
> PSR at rtime.felk.cvut.cz
> You can unsubscribe from the list at  
> https://rtime.felk.cvut.cz/mailman/listinfo/psr