We have some efficient IDE RAID systems nowadays that - in my head - SHOULD
take that responsibility off the processor's shoulders.
If so, speed (mb/s) and response time should be similar to SCSI. I have a
MB with one of those RAID controllers, but am lacking the processor and
memory for it. But I've seen it work at a customer with VERY good results.
Yet, I still have to test it with HL/CS
At 04:03 PM 11/21/2002, you wrote:
But as HLDS is'nt written for SMP then it won't be that noticable
I believe it was on this list that we discussed the issue of IDE drives
causing random lag spikes because of their high priority request being
serviced by the CPU (as opposed to SCSI which are serviced by the
controller). I would assume these issues would go away completely with a
dual CPU box yielding much smoother play. Also, things like stats can be run
on the same box without impacting the HLDS process. There are some benefits.
You just don't get double performance of the HLDS process itself.
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
