>Well, really now I just started implementing RB-trees in the page cache (I
>am too much courious to see how they will perform ;).
In the meantime only if you have some spare time, could you do a bench
over my last 2.2.5_arca8.bz2? It has major changes in the buffer code
(fixed flushtime and fixed the way buffers are flushed to disk by
bdflush). It works as never before here also under swapping (performances
visible with eyes ;).
in arca8 I restored the o+offset in the hash function, feel free to remove
it before running the bench if you think it will harm (the +offset is not
needed until you'll be very low on memory and it looks like to me that
your machine never gets low on memory ;).
arca8 has _not_ my only-one-page-query code that was unstable and I that I
am not going to fix because with rb trees I just noticed (while developing
the code), that I'll automagically catch without overhead if I am
inserting a just-existing-page so I will _never_ need the two query even
if I slept in GFP.
ftp://e-mind.com/pub/linux/arca-tree/2.2.5_arca8.bz2
Ah and also the first pagemap-cachealigned patch was wrong (blame me). The
second patch (after Eric pointed out to think about it twice) does the
right thing (so you may consider to run the bench on it again... excuse
me). So probably this is the reason that the pagemap-aligned+irq-aligned
bench was weird...
Andrea Arcangeli
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/