Memory: 16GB at 1333Mhz or 12GB at 1600Mhz

What is better having 4 4GB 1333Mhz DDR3 RAM or 3 4GB 1600Mhz DDR3 Ram?

Thanks

 

unless your big on benchmarking get 16gb. but alot of apps wont use either especially games, so know your software needs too

3X 4 also means no dual channel on the second pair dropping speed

 

4x4gb 1333

 

Like KCETech said, it depends. You cannot say that 4x4GB at 1333 is better without knowing OPs needs.

 

Duh, it's a no brainer. 4x4gb 1333mhz is both faster and bigger than 3x4gb 1600mhz.

 

How is 4x4GB 1333 faster than 3x4GB 1600?

But in any case you won't notice the difference, except in very memory speed sensitive apps or when using the IGP, but it's very minimal. My suggestion, go for the more memory.

 

Definitely go with more vs faster. As far as noticeable difference--one step is meaningless; two steps is marginal.

 

Wouldn't the 4x4gb be faster due to the dual channel RAM? The 2nd set in the 3x4gb wouldn't benefit. Honest question, I've never really been that sure on the whole dual channel RAM thing.

 

Ever heard of dual-channel memory? Of course it's gonna be faster.

 

Ever heard of asynchronous dual channel? It will not be measurably faster. In any case you're splitting hairs in performance for 99.999% of users and applications.

 

No it doesn't, and yes i can.

The 4x4GB dual-channel performance increase will completely nullify any performance gains by getting 1600MHZ instead (not to mention most laptops only support up to 1333MHz) and then there's that whole extra 4GB of RAM.

And you're not? Ironic considering you're arguing a noticeable difference between 1333MHz & 1600MHz.

Plus, asynchronous dual channel applies to different capacities, you still need the 4 sticks of RAM for it to work. 3x 1600MHz still isn't going to help you there.

Keep in mind the OP is specifically asking about 3x4GB dimms, not 2x4GB & 2x2GB.

 

In a new Ivy Bridge based notebook, I had the chance to try 2x4GB 1600MHz RAM vs. 2x8GB 1333MHz RAM and I noticed a huge and immediate performance drop when the (16GB) lower spec'd but higher capacity SoDimms were installed.

The whole system felt sluggish; IE would take a few seconds to display new webpages and basically the i7 3610QM system responded like something from 2007.

Putting the 1600MHz (8GB total) RAM back, the system felt once again like the 2012 high performance system it is (even with the ancient 5400RPM HDD the system shipped with).

My advice: keep saving until you can afford the proper rated RAM for your system (I'm assuming this will be going into an Ivy Bridge system, of course). If we're talking about a Sandy Bridge based system, then I'll agree with most here that the 1333MHz SoDimms (16GB total) will be the superior configuration - unless it's installed in a very highend i7 (XM) SB based system).

Good luck.

 

Why would you be assuming this is for an Ivy Bridge system? What has been said where which would give you that indication?

Also you're comparing 8GB of 1600MHz with 16GB of 1333MHz, it's documented that systems can run slugish if too much memory is installed but unused and I've observed this myself. Not to mention not every chipset or board supports 8GB dimms.

 

Where have you seen that 'too much memory makes systems run sluggish'? ROTHFLMAO

By stating my assumptions, I don't have to care what system is being considered: (especially since none are mentioned by the OP at all) - and my observations stand on their own.

 

No you don't. Asynchronous works with any number of sticks. Since there's only two memory controllers having 8GB (2x4GB) in one channel and 4GB (1x4GB) in the other will still be asynchronous. Asynchronous works with single + dual channel and optimizes for use with RAM amount that is dual channel. So in my example 8GB would be dual channel and the other 4GB single. In Intel systems it's called "Flex Memory". In any case it really doesn't matter much.

 

Some older controllers run slower when using 8GB dimms, most don't have native support for them. Also from memory the only 1st gen i-series chip which supports 8GB dimms is the i7. 8GB in dimms in general have some compatibility issues and can not simply be used in every system.

 

My vote for OP: get whatever is on sale. Seriously. Assuming that your laptop has a dGPU (in iGPU case, 1600MHz might be better).

Ignoring iGPU, you won't notice **any** difference between the two choices, other than a 4GB difference. Even then, you probably won't max out 12GB (I've tried everything other than some scientific computing and 4+ VMs, which will max that out). 1333MHz and 1600MHz won't matter to you unless you're benchmarking (basically: "I got bigger numbers than you!! So mine's better!!" in my book).

 

8GB is MORE than adequate unless you run lots of VM's or have specialized software that needs tons of RAM. 1333 vs 1600 MHz is minor difference.

 

I'll repeat again that on an Ivy Bridge based system 1333MHz SoDimms made a new system basically unusable.

You've been warned...

 

What?

1333MHz vs 1600MHz for Ivy bridge? - AnandTech Forums

 

I'm not talking about desktops as that link seems to indicate...

I'm talking about my direct experience with a new HP DV7 notebook.

That's 'what'.

 

That's absolutely ridiculous. There is NO way that slugishness was a result of the speed difference between 1600mhz RAM and 1333 mhz RAM, because the performance difference is less than 5%. You most likely had some incompatibility issues with the DIMMs.

 

Yea and with my experience as a tech i've dealt with people not having their system running & having it run badly because they had incompatibility with 8GB dimms.

Why are you arguing over semantics? You're saying the 8Gb dimms performed badly, we're saying that's because they aren't universally compatible. I can almost guarantee you that the performance difference was NOT due to a frequency difference of 266MHz, it was most likely a compatibility issue.

 

Guys, I'm not arguing at all.

Simply what I observed on the latest platform with the latest available 1333 MHz 8GB SoDimms.

Even if this is a compatibility issue - it is a big enough concern (given the responsiveness/performance hit the system exhibited) that people should be made aware of it.

As this was a show stopper for me (I could easily use 32GB RAM on my systems...) I returned the system - but it is something to think about either way.

(How many people have posted on how many threads (including me) that 1333MHz vs. 1600MHz makes no realworld difference - maybe with IB, it finally does....).

 

If that is really what happened and not in your imagination, this has nothing to do with 1333Mhz vs 1600Mhz. NOTHING.

 

I've noticed some slight improvements in system responsiveness going from 1600 to 1866MHz, but nothing that I would consider 1333MHz as slow when I used it just for comparison sake

 

It's called the placebo effect difference is just too small to be humanly noticeable. Only benchmarks can show it.

 

I thought that too, but it's not. I was a huge skeptic until I was finally able to bump my RAM speed up to 2133 and was like But it really doesn't matter for daily tasks. It's one of those things though like with an SSD, you run with it for a while, then go back to HDD and are like, wow this is slow. Not quite that magnitude but same idea.

 

If that were true, it would show in benchmarks, wouldn't it?

 

What benchmarks do you want run to show improvement in system responsiveness? There are none. I'm not saying it's like going from 5400RPM HDD to a Samsung 830 SSD. But there is a noticeable improvement. Go to JamesD's thread on RAM and review my negative or skeptical comments until I actually experienced it first hand. I called shenanigans from the beginning too. But in the end, it doesn't matter much because there's little improvement in processing tasks.

 

Better responsiveness is a result of improved processing speed. Which is measurable in PCmark and SuperPI.

Back in the day when I was adjusting timings on my DDR2 desktop rig, it definitely showed up in SuperPI.

 

System responsiveness isn't really a result of faster processing speed. Putting an SSD in your PC vastly increases responsiveness but it's not processing any faster. The I/O is much faster. That's about the only measurable difference which would be/could be done with memory benchmarking which Aida64 shows considerable increases. Same can be said for SSD. You can't really measure the system responsiveness except measuring the SSD performance itself; the seek times, random and sustained read/writes.

 

Should I get 8GB DDR3 or stay at 6GB DDR3? I don't game as much as before by the way.

 

You will not see any benefit from more RAM with gaming especially 6GB to 8GB unless you are a REALLY HEAVY MULTITASKER and have 4-5GB of stuff already running on your system at any given time. Single or Dual channel RAM makes zero difference when gaming with the dedicated GPU too.

 

Cool man. Btw HT when will you buy sager 9150 with 680?

 

I'm probably going to go with 7970m although the driver issues with Enduro are a bit disconcerting. Hard to justify $300 though for a marginal performance increase.

 

Really? Didn't know that. How come?

 

i would go with the 1600 if you have a highend CPU and GPU but otherwise it really doesn't matter.

 

The dedicated GPU has it's own memory, separate from the system memory or RAM (GPU RAM is called VRAM). For example, my Quadro 2000M has 2GB of VRAM to use, as opposed to the 12GB 1333MHz system RAM. If I was using the Intel HD 3000, it would share said 12GB RAM since integrated GPUs do not come with their own VRAM.

 

This is how:

http://forum.notebookreview.com/gam...amd-radeon-hd-6750m-benchmarking-results.html

Tested with 1066MHz RAM, 1333MHz, 1600MHz, 1333 single channel, CAS 7, CAS 9, CAS 11. No difference in performance for the most part. Ignore the CrossfireX and overclock figures.

 

RAM is so cheap there is no reason not to have 8GB.

 

I do have to agree with this unless you are stuck with only one slot...hard to justify 40-50 bucks for 4GB difference :/

That is how I feel about my series 3 netbook

 

If you already have 6GB and have no real reason to buy another 4GB stick, you are probably fine not upgrading.