Dear forum members,
We are delighted to inform you that our forums are back online. Every single topic and post are now on their places and everything works just as before, only better. Welcome back!
Discussion on Article:
2GB of RAM: Do We Really Need That Much?
Those who have excess cash can spring for 2 Gb. for gaming should they be inclined, though it's pretty tough to justify the expenditures fro Vid cards and DRAM.
Workstation folks always use large quantities of DRAM because it's useful for graphics intense applications, especially where time is money.
Most PC enthusiasts however really don't need 2 Gb.
1GB for standard PCs
2GB for high-end PCs
Run Quake4 @ Ultra settings & you'll see a 1GB system practically freeze
First of you, you did not define what you meant by "contemporary computer system". It seems all the tests revolve around standard computer operation and games, which does not seem very "comtemporary".
Secondly, RAM is storage, albeit temporary. Increasing the amount will never improve performance much unless you are using MORE RAM. What's next? An article proving a bigger hard drive doesn't increase performance?
I really wish you guys had looked over the data and wondered "Is this worth publishing?", because from the looks if it, there was no other reason to post this article besides to try to coax ignorant people to not buy more RAM.
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.
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.
because you can just buy a jump drive and a add on to the ram
U guys could add some tests with 2x512 modules off faster mem modules...
But this article it's very nice...
Go back and run the tests again on high and maybe you will see a large increase in FPS at 2gb ram. I certainly did...
The real question for enthusiasts should be whether running 2 gig of ram, whilst sacrificing timings is overall better than running 1 gig of ram at tighter timings.
Add your Comment
Enter your username and e-mail address. Password will be sent to you.