SV: Efficient IPC mechanism on Linux
From: Lars Hammarstrand
Date: Wed Sep 10 2003 - 06:24:05 EST
>>>On Wed, Sep 10, 2003 at 12:54:32PM +0200, Luca Veraldi wrote:
>> > Memory is sort of starting to be like disk IO in this regard.
>>
>> Good. So the less you copy memory all around, the better you permorm.
>>
>Actually it's "the less discontiguos memory you touch the better you
perform". Just like a 128Kb read is about the same cost as a 4Kb disk
read, the same >>is starting to become true for memory copies. So in the
extreme the memory copy is one operation only.
Well yes, that might be true for a one user desktop system. But
shouldn't we meet the demands from the "data center" users as well? Now,
I do hereby claim that: 1000 process/threads doing 128Kb reads is
_absolute_ not the same as 1000 process/threads doing 4Kb disk reads in
terms of i/o, memory and bus bandwidth.
Anyone willing to bet on this one?
Bottom line - don't waste unnecessary bandwidth, others might need it.
--
Lars.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/