Hi,
JAG I think the customers server would have had RAID, which is slower than a single HDD as the drives have to kept in sync. Also the network would impact on performance.

You can test the differences by installing an IDE HDD in the server and do copy tests from the RAID to RAID, then RAID to IDE, the copy to the IDE will be much faster, but vastly less secure.

Also different RAID types have different performances, RAID5 is slow compared to RAID10 as RAID5 has to calculate the data offsets for all writes, and in a continuous test run it can add a considerable amount to the process.

>If money were no object (say $25,000 budget), what's the fastest server I can get to run native VFP (VFP 8) with local data? Need Raid for data for hd failure, and about 250GB storage.
>
>Have several tables approaching 2GB.
>
>Here's why I ask: We have a developer in the office who ran a 48 hour test on his machine over the weekend of our software. Don't know the exact specs, but a high-end P4 with 1GB RAM, 10,000 RPM HD. We ran the same test on a relatively new, maxed out 6 month old $10,000 customer machine with dual Xeon Processors, 2GB RAM, 15,000 RPM HDs. The test ran faster on our developer's machine! Does it have something to do with Xeon processors only being able to utilize 50% of a single processor for a single thread? I was shocked.
>
>Also, we have optimized the foreground/background memory on the server.
>
>In case you are wondering why dual processors with single-threaded VFP, it's because we occassionally run several instances of our app simultaneously. Most of the time we don't, but when we do, we need all the horses.
>
>Thanks