Tux stable with hyperthreading?

Hi!

I have been able to successfully run Tux on a high-traffic site with RedHat 8.0 and 2.4.18-kernel, dual AMD 1600+, Adaptec SCSI.
But the site has grown too big for that server so I changed to a dual Xeon 2.4Ghz w. hyperthreading (so the system shows four CPU:s, and start four tux-threads).

It ran nicely for about two days, and then the whole system crasched. In the middle of the night (4.10 am). The load was only 0.6 at that time. And the disc-activity was low, backup starts at 5 am. The system responded to nothing exept ping and the capslock-key on the keyboard still turned the capslock-led on and off. The screen was totally blank. I had a SSH connection open from another computer which was not disconnected, but neither responded to anything.

The configuration is exactly as on the old stable machine.. No logging is enabled... And the hardware is very much the same except for motherboard and processors.

One thing is quite strange.. The load on the four tux-workers is mostly like two of them is doing all the work, and the other two is mostly sleeping?

$ ps ax |grep TUX
21458 ?        S      0:01 [TUX date]
21459 ?        SW     0:00 [TUX logger]
21460 ?        S      0:00 [TUX manager]
21461 ?        S     15:48 [TUX worker 0]
21462 ?        S      2:43 [TUX worker 1]
21465 ?        S      5:12 [TUX worker 2]
21470 ?        S     45:39 [TUX worker 3]

Anyone got the same experiences?
Anyone know if I should run only two tux-workers?

Anyone know if I can gain stability or speed by applying any of these hyperthreading-patches avaiable on the net? I tried a little but was not able to find any patches for the 2.4.18 kernel (only 2.4.18.xxx, and where do I find sources for that strange one..?)...

Thank you 
//Andreas