Re: using doubles in the kernel

[Seth Arnold <sarnold@wirex.com>]

On Sat, Jul 06, 2002 at 01:35:52PM +0100, Lars Olsson wrote:
> I'm implementing a neural network (executing in kernel
> space because of performance reasons)

I've heard "performance reasons" meaning things should be implemented in
the kernel over and over again, and just once, I'd like to try to find
out _why_ implementing something in the kernel will improve your
application's performance.

You still need to yield the processor to the other $N processes running
on the system, you still need to answer interrupts.. The only way
implementing this in the kernel is if you can save on data copies across
protection boundaries. Since neural networks are inherently
compute-bound, I'd be surprised if you could see any performance
improvements by moving this into the kernel.

Why move NN into the kernel?

-- 
http://www.wirex.com/

PGP signature