Using the Fedora Core 4 SMP kernel the system time goes exactly twice as fast as normal all the time ( 2 seconds of system time = 1 second real time ). I'm using a Pentium 4 processor and SiS chipset. Is there a way to fix it ?


P.S. I know there are posts here similar to this one, but this is different from other posts I've seen cuz all the others were dealing with either ATI chipsets or AMD processors which I do not have. And the system time is like this all the time not just in bursts. I have yet to find a solution that works.

no your right it isnt ATI chips responsible it is a direct result of SMP

i still have no idea to fix it ,

mandriva appending
notsc
and
no_timer_check

to the bootloader fixed it

however it doesnt help i SUSE

running default Kernel ie Single processor however works fine, sort of defies the point of a dual processor system.

If you have P4 with HyperTheading (not a dual Core)
since the processor itself delegates the 2nd logical (non Existant CPU you dont need a SMP kernel to make use of it.

only use smp kernels if you have 2 or more processors, or dual core

I have two pentium 4 CPUs (in different systems) and I have never had that problem before. Though, I compiled my own kernel in Slackware.

Wouldn't it be better to have SMP enabled on a HT enabled CPU? I notcied after compiling the kernel the second time around (with SMP support) that using the -jx option gave me a ~40% decrease in compile times.
x being a number of simultanious threads to execute at once, I usually use 2 for single CPU, and 3+ for dual/HT, since even on a single CPU (without HT) compile times dropped by 20% to 30%. It increases in speed because there is time in between when it finishes compiling one source file, to starting to compile another.

no, it doesnt make a difference

Well ... Fedora Core 4 chose to install the SMP kernel, I didn't. But my Pentium 4 has Hyperthreading support, and it works faster with the SMP kernel so I'd rather keep it. I tested both SMP and non-SMP kernels and I'm pretty sure the SMP is faster.

I suggest reading this: http://www.linuxelectrons.com/articl...40226231747944

If you still don't beleive, give it a try yourself.

(and sorry for digging up an old thread. I just saw that article and figured I should get my point accross.)

Also, H_TeXMeX_H, did you fix the problem with the clock?