On Tuesday 31 August 2004 08:49 pm, Randall R Schulz wrote:
The one OS is running under the other OS. They don't compete any more than any other software competes for cycles.
Neither Linux or Windows is written with a VM environment in mind. They both assume they have complete control of the machine. Neither end up operating as responsively in a VMware configuration on any given physical hardware setup as they would without VMware and the other OS present.
You assume that the FAIR way to compare this is to compare a VM to the same OS running on the raw hardware. Thats hardly a reasonable comparison. Its not what VMware sets out to do. Unless you have infinite budget to run one Windows and one Linux computer.... Or several of each. But I don't necessarily believe its as bad as you say, because my co-worker and I have identical laptops (P4 1.6gighz 768meg ram). He does his compiles in WinXP. I do mine in WinXP running inside A VM on top of SuSE 8.2. When benched against the same long running compile job, mine finished within 3 seconds of his. Occasionally faster, occasionally slower. Apparently Linux caching and ReiserFS3 for my Virtual hard drive beats the living snot out of real NTFS. I have run as many as 3 VMs at the same time testing client server application software that we write. This is slightly slower than using two real machines, because the VM acting as the server side (not being displayed) has less priority and there is a slight pause (sub 10th second) while it answers requests from the Client side where I am working. I still think its an excellent way to run some particular application that just will not cooperate under COO or wine, for which there exists no Linux equivalent. AutoCad comes to mind. But hey, Randy send me that 2.4 and I'll tune it up for you... ;-) -- _____________________________________ John Andersen