[llvm-dev] RFC: System (cache, etc.) model for LLVM

[David Greene via llvm-dev]

"Finkel, Hal J." <hfinkel at anl.gov> writes:

>> Ok, I understand better now, thanks.  L1P really is a prefetch buffer
>> but there's software control to move it to faster cache if desired.
>> Should we model it as part of the prefetching API?
>
> At this point, I'd not base any API-structuring decisions on the BG/Q
> specifically. The generic feature that might be worth modeling is: Into
> what level of cache does automated prefetching take place? I know of
> several architectures that don't do automated prefetching into the L1,
> but only into the L2 (or similar).

Ok, that shouldn't be difficult to add.

                             -David