fast_clear_page() uses mmx instructions for clearing a page, what about
using sse instructions?
sse instructions can store 128 bit in one instruction, mmx only 64 bit.

On a Pentium III it isn't faster, but it should be faster on a Pentium
4.

I've written a user space test app - could someone with a Pentium 4 run
it?

Just compile the attached program and run it for ~ 1 minute.
Then write down the lowest numbers for 'zerotest', 'zeropg_1',
"zeropg_sse_1' and 'zeropg_mmx_1' and post them to linux-kernel.

Thanks,
	Manfred
P.S: you need a recent binutils version to compile the app, I've
uploaded a precompiled binary to

http://colorfullife.com/~manfred/zero