>
>
>By adding three bits of colouring you're going to cut the collision of
>about 1/8.
>
No, Shuji is right:
You have just shifted the problem, without reducing collisions.
256 kB, 4 way cache with 32 byte linesize.
cacheline == bits 15..5
offset within cacheline: bits 4..0
The colouring must depend on more than just bits 13 to 15 - if these
bits are different, then the access goes into a different line even
without colouring, there won't be a collision.
Shuij, I don't understand why you need both a shift and a modulo:
address % odd_number should generate a random distribution (i.e. all
bits affect the result), even without the shift.
-- Manfred
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/