Re: 4MB pages and framebuffer access, x11perf results, 2.1.125

ganesh@cse.iitb.ernet.in
Fri, 30 Oct 1998 23:38:08 +0530 (IST)

Messages sorted by: [ date ][ thread ][ subject ][ author ]
Next message: I Lee Hetherington: "Re: ifconfig weirdness on axp 2.1.126"
Previous message: Alan Cox: "Re: Bad TTY Magic Number -> Memory Initialization Problem?"
Maybe in reply to: MOLNAR Ingo: "4MB pages and framebuffer access, x11perf results, 2.1.125"

> Correct :( without the GLOBAL flag, the numbers are almost identical, only
> slight improvement for the 4M version, not worth to mention. Sorry, this
> really was unintentional ... :( It's amazing though how much difference
> this made.
>
> -- mingo

could someone clarify this please ? as far as I've understood,
performance
1. 4k pages 100
2. 4M pages 100+delta
3. 4M+global around 115

I thought the only difference between 2 and 3, with respect to 1, was that
case 2 would incur _one_ extra tlb miss for the 4M page whenever X was
switched in. how can this alone produce such a large difference ?

-- ganesh

-
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/

Next message: I Lee Hetherington: "Re: ifconfig weirdness on axp 2.1.126"
Previous message: Alan Cox: "Re: Bad TTY Magic Number -> Memory Initialization Problem?"
Maybe in reply to: MOLNAR Ingo: "4MB pages and framebuffer access, x11perf results, 2.1.125"