SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : How high will Microsoft fly? -- Ignore unavailable to you. Want to Upgrade?


To: Rusty Johnson who wrote (47498)7/5/2000 5:00:52 PM
From: SunSpot  Read Replies (2) | Respond to of 74651
 
In the real world, most servers don't have 8GB RAM. The less RAM, the better Linux gets compared to Windows. A 32MB web-server can handle quite a load.



To: Rusty Johnson who wrote (47498)7/6/2000 12:17:02 AM
From: mozek  Respond to of 74651
 
It's interesting that the maximum cache set and cache file size on the Linux server was 100MB and the maximum IIS cache on the Win2K server was 2MB with a max file size just shy of the 1MB max file size used in the static test.

It's also interesting that the test required dynamic CGI get with forced createprocess/fork for serving an ad. In the real world, process pools or threads are used to make the performance much better. This is especially interesting since it is already widely known that Win2K has more createprocess overhead than *nix has for fork. In the real world of Web servers, this overhead is usually moot since either thread or process pools are used to serve requests.

It is impossible to tell from these results what would happen in a real world, process-pooled environment where Win2K had the same cache size as Linux. While someone could make an argument for the forced createprocess, it REALLY surprises me that such different cache sizes would be published as if they were reasonable benchmark settings. It's almost as if they weren't trying to fairly represent real world performance of the two systems.

Mike