By the way, why was Quake4 ONLY run @ High Quality?
Run Quake4 @ Ultra settings & you'll see a 1GB system practically freeze

Mallissin makes a very good point. More RAM will only noticably improve the speed of a computer if there previously was not enough RAM available and data that should've been stored in RAM was paging.

If you really wanted to explore performance increases, you should've been working with applications that require more than 1 gig of memory and show the performance increase.

You don't increase your RAM to speed up useable apps, you increase your RAM to use apps that NEEDED more RAM than you had.

I'm sorry to seem hostile, but the article is testing something out of context and without proper variations of use to prove the point.

Most games all run the same way. Content (textures,geometry files, etc.) is loaded to RAM so they can be passed to the video card quickly for rendering. This does NOT STRESS THE RAM and actually makes a point for adding MORE RAM, since more content is always preffered and having as much of it in quickly accessible storage a big plus. Most games hard code their engines with set static limits to the amount of RAM used, and this is not always variable. Once more game companies begin using multi-resolution compression on content, RAM will become extremely important and in large quantities. The multi-resolution will require more variable caching for the video card and data to be available for parsing by the processor.

Video editting applications are the complete opposite. There is too much content to be loaded to RAM, so while you're working most of the data is cached to hard drive, making the case for a fast hard drive array and not more RAM. More of the cache can be moved to RAM though letting the user move through the data faster if more is available. I'm not sure how Adobe Premire was tested, since I don't see a description, but I'm willing to bet the test was using a small set of data which is also not very helpful. The spot on Photoshop also annoyed me, since it *IS* typically professionals who use this program, and the reviewer clearly ignores this fact.

Compression benchmarks don't help at all either, since most use streams that depends more on how fast the processor can do it's job than anything else.

The only benchmark that I thought would show the benifits of more RAM was the 3-D content creation, using 3DS Max 7. Anyone who has used MAX knows RAM is extremely important, and that is because most CAD programs use vector data. If more of this vector data is thrown into RAM, screen refreshing is much smoother because the processor can run it's calculations without paging to disk. Microstation, AutoCAD, and nearly any other CAD program all run the exact same way. The fact this benchmark showed no gains, made me the most annoyed because I work with Architects and 3D artists, making my living from making sure they can work quickly.

I really believe the reviewer should have threated the RAM more like a hard drive. They should have thrown different sets of data at each program and showed the results. I believe from experience that the data produced from this sort of benchmarking would clearly make the case for MORE RAM.

First off if you want to play games get a $500 Xbox 360 and save a LOT of money. Second computers were not invented to play games so take games out of the test and 2 gigs of ram is a total waste. Any current day "cumputer" runs fine on 1 gig. I.M.H.O