Topic: Do You Really Need More Than 6 GB Of RAM?

[Pugnate]

http://www.tomshardware.com/reviews/memory-module-upgrade,2264.html

Ahem, as I have been saying for oh so long, nooooooooo...

In fact, it looks like there is NO advantage of going for more than 3GB RAM, for most high end users... you actually end up losing slightly, because you are spending the extra watts.

Disappointed in the lack of performance differentiation after the program had been launched, we began various tests to determine if any “normal” user would see larger benefits during multitasking. The most common “background” applications we could think of were file downloads, video encoding, and virus scans, while the most demanding top applications are games. Because virus scans tie up the hard drive, we limited our background applications to video encoding (TMPGEnc with DivX) and file transfers (Windows Vista SP1 x64 standalone update at 726.5 MB) while retesting our Far Cry 2 benchmark. Simultaneous use of these tasks had no noticeable effect on game frame rates, and while DivX encoding times did slow down by around 60%, memory capacity did not affect those results. We even tried opening more than 100 Internet Explorer windows, but there was no noticeable impairment even with only 3 GB available.

So there.

[WindAndConfusion]

While I don't necessarily doubt his conclusion that most people don't need ridiculous amounts of RAM, it seems to me that Tom has a raging huge hard-on for worthless benchmarks.

0) He didn't bother to establish the validity of his test method by examining systems with a very small amount of RAM (like 512 MB) and showing that the outcome of his tests actually dependa on the amount of RAM available. The fact that he got exactly the same results for page after page leads me to believe that his benchmarks are just a useless wankfest.
1) Instead of doing the same worthless tests over and over again, he should have identified the limits for each system (eg, "a system with X megabytes of RAM was able to cope with workloads of up to size Y before suffering performance degradation at rate Z").
2) He couldn't bother to provide Dtrace dumps whatever the Windows equivalent of Dtrace is, which would have been eminently useful in interpreting his results.
3) He didn't bother to test any really heavy loads, like virtualization, IDEs/SDKs/debuggers, or JIT interpreters.

[Pugnate]

0) Yea but he isn't conducting a scientific study. He knows who is target audience is, and that they understand that at some level, increasing levels of RAM does improve basic performance. 512MB may have been a waste of time, but I would have been interested to see results starting from 2GB.

1) Again, target audience... but yea, it would have been nice to have that info. Would have taken a lot more work though.

2) ....dude it was a simple test. What sort of interpretation do you need?

3) Again... again... target audience. It was a quick and dirty test. I think you missed the following quote:

We can only recommend larger capacities of 8 GB to 12 GB for professional applications where its usefulness has already been documented and for servers. None of our tests required high-memory capacities and wasted RAM is a burden both financially and ecologically.

[Xessive]

I think if he ran 3DMark or PCMark with different RAM for each benchmark his readers would have been content. It doesn't need to be a complicated experiment. It's reasonable to think that more RAM = better but the question is by how much. As it turned out more RAM is not always better, it can be wasteful in certain circumstances.

[Pugnate]

I think for someone like you, who uses photoshop extensively, lots of RAM and a quadcore is a good thing.

[Xessive]

I think for someone like you, who uses photoshop extensively, lots of RAM and a quadcore is a good thing.

Yeah, I can put it to good use especially when I'm preparing file for print. Never thought I'd need to to make a 1200dpi photobased image for print, and it pushed my system to its RAM limits. I could put the extra RAM to good use but the average gamer doesn't need more than 3GB (or whatever the games' reqs are) and the average user (multimedia, etc.) doesn't need more than 2GB.

I have 2GB DDR2 @ 800MHz and as an enthusiastic  gamer I have no complaints. Not yet anyway.

[Pugnate]

You have 2GB on the comp you work from? Or is that your gaming comp?

[WindAndConfusion]

0) Yea but he isn't conducting a scientific study. He knows who is target audience is, and that they understand that at some level, increasing levels of RAM does improve basic performance. 512MB may have been a waste of time, but I would have been interested to see results starting from 2GB.

1) Again, target audience... but yea, it would have been nice to have that info. Would have taken a lot more work though.

It wouldn't have taken much more work than it took to perform this ludicrous waste of a test.

2) ....dude it was a simple test. What sort of interpretation do you need?

Since this was explicitly a test of the usefulness of RAM capacity, it would have been nice to see how much of his RAM was actually getting used. And for fuck's sake, every modern OS has this shit built-in. All he had to do was Ctrl-Alt-Del and bring up the resource monitor.

3) Again... again... target audience.

Virtualization, IDEs, and JIT interpreters are all REALLY common, ESPECIALLY among Tom's target audience. (For instance, JIT interpreters exist or are planned for all existing web browsers. Are you saying that Tom's "target audience" doesn't browse the web?)

[Pugnate]

It wouldn't have taken much more work than it took to perform this ludicrous waste of a test.

For fucks sake, I think you may have missed the point of the article (hint: it is in the title). And the sort of gamers interested in the results of the article aren't armed with anything less than 2GB on Vista. You are talking about Vista tests starting with 512MB, that wouldn't have been worth the painstaking effort.

Look the point of the article was to see if modern every day gamers needed to spend the extra $200 to double their RAM from 3 to 6GB. These aren't the sort of people interested in an elaborate scientific study. They just care about benchmarks.

The fact is that all three test systems were the exact same, and ran the exact same benchmarking software. All that was changed on each occasion was the memory size, which is a perfect little study for some pimply little basement nerd wondering if his $200 will be better spent on a faster processor, a better video card, or more RAM.

If you think the test was a waste, then you are an idiot.

Virtualization, IDEs, and JIT interpreters are all REALLY common, ESPECIALLY among Tom's target audience. (For instance, JIT interpreters exist or are planned for all existing web browsers. Are you saying that Tom's "target audience" doesn't browse the web?)

That's it...you are full of shit.

Christ, explain to me how most every day gamers would be interested in any form of virtualization? My cousin runs an IT servers business in Asia, and works on his own data for his clients. I guess he might be interested virtuialization benchmarks... but really for that sort of thing, I can imagine 3GB RAM being useful as you are when it comes to piecing together what is appropriate for a given situation.

[WindAndConfusion]

Where on Earth are you getting the idea that the article was written exclusively for gamers? He ran fucking productivity tests on page 4.

Maybe you're just taking this too personally, but the fact of the matter is that I think Tom is a useless twat.

[Pugnate]

Where on Earth are you getting the idea that the article was written exclusively for gamers?

Well not just gamers, but regular users. And ummm.... from the fact that he included none of the indepth tests you pointed out?

He ran fucking productivity tests on page 4.

Because no one is expected to run moderate productivity programs right?

Maybe you're just taking this too personally

ahahaah

confused by random dickishly rude comments? I was just trying to be you. But I think I fell short.

[Xessive]

You have 2GB on the comp you work from? Or is that your gaming comp?

Both. The work one is a laptop and my gaming rig is a tower. Although I often do my work on my gaming one anyway, I like my big monitor more

"Stop screwin' around, you're screwin' around too much!" - Mr. Adler

[MysterD]

http://www.tomshardware.com/reviews/memory-module-upgrade,2264.html

Ahem, as I have been saying for oh so long, nooooooooo...

In fact, it looks like there is NO advantage of going for more than 3GB RAM, for most high end users... you actually end up losing slightly, because you are spending the extra watts.

So there.

Maybe Crysis 3 and Sup Com 3 would need that kind of RAM.
Oh, and DNF...

[Pugnate]

I have to clarify my earlier statement. You aren't spending that many extra watts or whatever.

Also, these days you can get 6GB of DDR3 for about $90... which is pretty sweet. And it has to be said that once 6GB gets more popular, more applications will start taking advantage. And surprisingly, most DDR3 packages are coming at either 3 or 6GB.

[Cools!]

For video, having more ram can make a huge difference in the performance of your system. In editing, you often have to work with multiple video streams, the more ram you have the more streams you'll be able to work with in real time (with added effects, etc.).

If you are doing motion graphics, compositing, etc. then your demands will be even higher. While you are working, your graphics card will do most of the heavy lifting, but as soon as you want to have a preview you need to render, which means that you have to load each finished frame into your ram.

For example, if you are working at standard definition and your frame size is 720x480 pixels at 24bit (RGB plus alpha), at 30 frames per second for 15 seconds that's already over 3GB of ram just for playback.

So more ram equals more options and faster workflow.

Even photo is becoming crazy demanding these days.

I shot Toronto's fashion week in March. 23 shows, over 10,000 photos. All in raw format. I use Apple's Aperture for photo management and it'll tax your system like crazy once you start browsing all the photos. Don't even want to talk about the new HDR craze.

[Cobra951]

I have no doubt that eventually memory over 4 GB total (including system-reserved addresses) will be quite beneficial.  The problem to overcome I guess could be likened to inertia.  This has been a hardline barrier for quite a while.  Everything needs to be designed or modified to surpass it, the hardware, the OS, and the applications.  Any noncompliant link in the chain will nullify the potential advantages of a lot more RAM. 

[Pugnate]

Cools, absolutely, if you are working in the video/graphics field, then a lot of RAM is definitely an asset.

[Cools!]

Cobra brought up a good point. How do we know if getting more ram is not simply a way to patch up a badly designed system? A system that is not well optimized?

It seems that the more resources you have to offer, the more of it will be trashed. Applications get more and more complex and there's only a limited amount of development time. One of the biggest problems I see these has to do with parallel/concurrent programming for multiple cores, etc.. Parallel computing isn't easy, both in terms of setting a system up and programming for it. It's much harder to find bugs since they tend to happen at edge cases that are hard to test for. And finally, it can be just difficult to wrap your head around the whole thing.

[scottws]

Yeah when I took Java programming, I had a hard time with creating multiple threads.  I saw the benefits easily, but coding for it was very difficult compared to all the other work I'd done before, and even after that.