Do You Really Need More Than 6 GB Of RAM?

[Guest]

Microsoft should dumb the x86 already and start offering x64 to the mainstream. We see memory requirements growing all the time while software developers mainly concentrate to x86 architecture. And it will continue to be so until mainstream is forced to x64. Microsoft is the key component to make this happen.

 

[jkeelsnc]

Thank God for X64. Bye bye to only 8 General Purpose Registers which was too limiting. Having said that, X64 gave tremendous lease on life for the x86 architecture flawed as it was up until 32 bit levels.

 

[Guest]

I have been using 8GB on vista 64Bit for 2 years now. Recently, one of the sticks became unstable and I took it out to get it replaced, and I imediately noticed the HDD rattling away much more often.
Also, as a MMO player, I often run 3 x instances of the game at once, and that uses 5.6GB of ram, and with Excel running and a movie which I often watch while waiting in-game, it goes over 6GB.

As for XP, who cares? I could of told you 7 years ago its not worth putting in any extra ram in...

 

[filip007]

You need to know that 2GB is limit for 32-bit application that can be used as max and most games are 32-bit build so there is no point have more than 3GB (2x1GB+2x512MB) or (3x1GB) for max usability.

For 64-bit Photoshop profesional hardcore user that people need more RAM everyone else 3GB is just fine.

 

[awaken688]

I'd love to see some tests with Adobe Lightroom 2 being used. How about a wedding collection with 3000 images from a wedding with 24MP files from a new SLR. Especially when loaded in a catalog with over 100k other images. I know it would make a difference, but I'd love to see some benchmarks similar to the test above for importing and exporting RAW files.

 

[killmess]

[citation][nom]Anonymous[/nom]...Really Micro and very Soft long time suggested for people to upgrade to better windows ..... Windows X![/citation]

You mean "Xwindows" the graphical interface of linux.
So, a better windows, it's a non microsoft bu77sh1t. ;D

 

[warezme]

What a lazy write up. Yes its common knowledge that there is no difference in 3GB to 12GB if you are just opening up ie explorer windows and doing mundane pedestrian stuff. Turn around and test those same systems with a couple of 21MP RAW files in PS with layers or try to do some multiple vm's and watch those low memory systems start to crawl.

 

[Tindytim]

[citation][nom]warezme[/nom]What a lazy write up. Yes its common knowledge that there is no difference in 3GB to 12GB if you are just opening up ie explorer windows and doing mundane pedestrian stuff. Turn around and test those same systems with a couple of 21MP RAW files in PS with layers or try to do some multiple vm's and watch those low memory systems start to crawl.[/citation]
QF-F***ing-T

Try opening up some dense scenes in 3DsMax or Blender, then tell me 12GB doesn't benefit you. It seems like Tom's Hardware is really Tom's Gaming Hardware.

 

[robertbradbury]

It might be nice if you provided tests under Linux rather than just Windows. It is relatively easy to setup real "user" benchmarks under Linux though some of the more hardware related tests are a bit more difficult.

If you want some "real" memory using applications, consider Blast with some really large genome (.fasta protein sequence) files or molecular modelling/dynamics applications (gromacs, moldy, namd, etc.) with some very large molecules. I suspect these types of applications would tank if the programs were required to swap due to memory limitations. Running relational databases (Oracle, mysql, etc.) show significant performance variations dependent on cache size and table size.

 

[p05esto]

Meh, I'll keep my McMansion and Truck AND SUV thank you. The green weinies can save the planet for me....while I enjoy my life. I have 6GB and almost bought 12, next rig to have 12GB with Win7.

 

[guumbastic]

Sorry if this is covered by someone else, or even if my assumption seems stupid to some, but have you ever thought that after throwing so much hardware at it, perhaps the problem isn't in how much memory or CPU or video card muscle you throw at it, but the software itself just isn't capable of going any faster?

oh, and no surprise windows didn't run those games any faster. systems with 2GB and higher will see no real impact from gaming as long as you have a video card worth its weight in the system.

also, another benefit from running this much memory (I run 8GB DDR3-1600, an intel Q6600 @3.6GHz w/50% FSB OC on an ASUS Maximus Extreme myself), at least as far as windows is concerned, is that you can shut off your paging file and let your hardware do its job. i know my system runs a ton faster and i suffer no ill effects from lack of a page file.

just sonething to think about in the next benchmark, perhaps - shut off the windows page file and then let's see the numbers.

 

[Guest]

Whatever.

 

[dtla1730]

I think the only reason I wasn't able to get completely into this article is the fact that RAM is so inexpensive at the moment.

The question "How much RAM is necessary?" in the current marketplace seems less interesting than, "Is more RAM going to hurt anything?"

If there were negative side effects related to "excessive RAM installation" (aside from 'it's not all being used with 32 bit addressing') then I might be more interested.

 

[Tindytim]

I just bought 12GB of RAM myself.

And as many have stated, the right apps and settings aren't being used to take advantage of more RAM.

I'll be getting rid of my swap partition, and increase the overall quality of my 3D work (larger, more detailed scenes, more poly dense models, higher res textures, etc).

 

[Guest]

Seems this whole study was done at the wrong side of the playing field. FPS gaming? How about Photoshop loading times, file copy, multiple apps running, etc.?

 

[WilliamMGeorge]

On my main home system I run 6GB of memory (64-bit Win7) with 1GB assigned as a RAM drive for temp files and other usage. I also turn the swap file off, which seems to make a huge difference in performance to me - though I've not done any benchmarking to back that up, so it may just be in my head I still run into situations when multitasking, though, where Windows informs me that memory is running low; I've not yet had it crash out an application because of that, but it happens from time to time.

While this article was interesting, I would like to see simulations of more intense workloads to see how 6GB vs 12GB fare there - something more than the summary at the end of the article would be nice.

 

[Guest]

One thing these tests didn't look into was multiple users on the same computer, all logged in all the time.

 

[b3l13v3r]

I know this article is a bit old now but as an update, using Windows 7 makes this conversation obsolete and will prove the author of this article completely correct.

 

[Guest]

If you want to have multiple programs that are memory intensive at the same time... Like running multiple compression programs (lets say 7-Zips LZMA as it it memory intensive) I bet you would see a huge differance then.

Not to mention ramdrives. Though their use might be somewhat limited for normal users, there are some situations where using it as a temporary storage location would be way faster than a drive. (I mean mapping a section of ram and using it as a drive, not the other type of ramdrive that exists as a card in a pci slot.)

Though for sure, there's no game out there that can make use of 4 GB of ram... yet. (But I'm pretty darn sure that there will be eventually.)

 

[Narg]

Virtual desktops is a VERY good reason. I run 64-bit Vista, and use a lot of virtual XP desktops. With less than 4 Gig of memory, my system would become unresponsive once I was at my normal running configuration. I now run with 8 gig of ram and my system is very happy.

This was a poor test. One game at a time? Games are coded to be small and quick, so as to not be memory limited on lesser machines. Very poor choice for testing memory usage. Use 100 I.E. windows doesn't work either, as I.E. is also made to switch memory to lessen the load.

To get a REAL test, run 64-bit apps side by side and test their product with various memory sizes. THEN you'll get the real picture. Instead of this sophmorish attempt at testing.

 

[thomasxstewart]

If You Have vista or 7, 4 Gb of memory was 5.4 index score here, while 6 Gb of memory is 6.7 index score, so it must be better & that was with 32 bit. seems to recognize while thing.

SignedHYSICIAN THOMAS STEWART von DRASHEK M.D.

 

[deckeon]

After I got 8Gig on Vista, I got rid of the swap file. Yes it's faster. Maybe if you don't have a clue how to use the extra memory it doesn't matter.

This article is poor.

 

[vabeachboy0]

a good test would be supreme commander with 8 players at 1,000 units each

 

[EarlZ]

Wait, I must be missing something why "Ready Boost" was used to test the difference in launch time with higher memory capacities.

IIRC, ready boost uses external sources like flashdrives and such to cache files, so why is this feature used in testing larger memory capacities... Or do they mean Vista Superfetch instead?

 

[Crashman]

[citation][nom]earlz[/nom]Wait, I must be missing something why "Ready Boost" was used to test the difference in launch time with higher memory capacities.IIRC, ready boost uses external sources like flashdrives and such to cache files, so why is this feature used in testing larger memory capacities... Or do they mean Vista Superfetch instead?[/citation]

Great catch! Yes, things get misworded sometimes.