Questions about upgrading RAM for my alienware aurora r4 system

Hello Corsair forum,

I recently got myself an Alienware aurora R4 desktop. I know the standard RAM supplied with it is crap so i opted for the minimum amount and want to upgrade.

I am by no means a RAM expert so I was hoping someone could help me answer a few questions and help me pick a corsair RAM kit.

Question 1: I've read a document posted here on the forum by corsair called AN806: Memory Upgrade Resource Guide. It stated that when using all 4 DIMM slots (4-up) on my motherboard, I would not be able to clock the cards as fast as when using 2 DIMM slots. I read somewhere else that RAM runs stock at half its advertized frequency, so when you buy a 1600Mhz stick, it will run stock at 800Mhz. Does this mean I can't run 4 1600Mhz sticks at the actual 1600? Or does this lower maximum frequency only apply to speeds higher than the advertised frequency of the sticks?

Question 2: I know that the lower the timings, the better, but can I select a RAM kit with timings as low as possible? Or is there some limitation I should be aware of?

Question 3: In the Memory Upgrade Resource Guide, it says the following about 4-up: "This increased load can result in instability if the BIOS is not tuned or tweaked to compensate. The user should make certain that their motherboard BIOS allows adjustments to the memory controller voltage." So does this mean I'm supposed to alter the voltage to my CPU? The CPU has the memory controller right? So how does this work and how do I know how much to increase/decrease the voltage?

Now to help me pick a RAM kit, I want to buy a 32GB kit for use with Maya, Photoshop, etc. Now I noticed there are several lines or RAM made by corsair that have similar tech specs, so is one line still better than the other? Is there some factor that dictates what line of ram sticks I should use? Or is it just a matter of aesthetics? I tried to use the corsair memory picker, but it listed the alienware aurora 4 times with no difference in naming (old models i think), so that didnt help me choose. Here are the similar types of RAM I'm talking about:

http://www.corsair.com/en/memory-by-product-family/dominator-platinum-ddr3-memory/dominator-platinum-with-corsair-link-connector-1-5v-32-gb-dual-quad-channel-ddr3-memory-kit-cmd32gx3m4a1600c9.html

http://www.corsair.com/en/memory-by-product-family/vengeance-pro-series-memory/vengeance-pro-series-32gb-4-x-8gb-ddr3-dram-1600mhz-c9-memory-kit-cmy32gx3m4a1600c9.html

http://www.corsair.com/us/memory-by-product-family/vengeance/vengeancer-32gb-quad-channel-ddr3-memory-kit-cmz32gx3m4a1600c9.html

All are 1600Mhz, 9-9-9-24 timings, 1.5v, quad channel kits.

Any answers / help would be greatly appreciated.

Question 1: I've read a document posted here on the forum by corsair called AN806: Memory Upgrade Resource Guide. It stated that when using all 4 DIMM slots (4-up) on my motherboard, I would not be able to clock the cards as fast as when using 2 DIMM slots. I read somewhere else that RAM runs stock at half its advertized frequency, so when you buy a 1600Mhz stick, it will run stock at 800Mhz. Does this mean I can't run 4 1600Mhz sticks at the actual 1600? Or does this lower maximum frequency only apply to speeds higher than the advertised frequency of the sticks?

DDR stands for Double Data Rate. So you take your actual frequency and double it for an effective frequency. In the case of 1600mhz memory in CPUz it wil read 800mhz. 800mhzx2=1600mhz

As far as running all four dimms is concerned, that is going to depend on a few things. How well your board overclocks, your CPU , and of course the memory you choose. If you buy a proper 4 stick matched kit, then you SHOULD be able to run all for DIMMs. However , that is given your CPU's memory controller can put up with the stress of running four sticks. This is why you can usually reach higher speeds with 2 sticks vs 4. It's less strain on the memory controller.

The other side of this is that your CPU supports 1600mhz memory natively, so there should be no issues with running four sticks at 1600mhz. It may require a slight bump in voltage to achieve stability,(this goes back to the strain on the memory controller again) but it shouldn't be a problem.

Question 2: I know that the lower the timings, the better, but can I select a RAM kit with timings as low as possible? Or is there some limitation I should be aware of?

Sure, the worst that could happen is that they would be too low for a stable machine and you would need to loosen them. However , memory made today is tailored towards the new CPU's and again shouldn't be an issue.

Question 3: In the Memory Upgrade Resource Guide, it says the following about 4-up: "This increased load can result in instability if the BIOS is not tuned or tweaked to compensate. The user should make certain that their motherboard BIOS allows adjustments to the memory controller voltage." So does this mean I'm supposed to alter the voltage to my CPU? The CPU has the memory controller right? So how does this work and how do I know how much to increase/decrease the voltage?

You only add what you need to make it stable. If your memory is 1.5v memory, you would start out at say 1.55v and then add .2v after that with a max voltage of 1.65v. That is the max amount you would want to go with your CPU.

Now to help me pick a RAM kit, I want to buy a 32GB kit for use with Maya, Photoshop, etc. Now I noticed there are several lines or RAM made by corsair that have similar tech specs, so is one line still better than the other?

Yes and no. A 32 gig kit at 1600mhz is going to perform just about dead even unless the timing as different and even then it would take a benchmark to see any difference.

For the system your working with Vengeance, Dominator Platinum, or Vengeance Pro would be the three top choices. They are made for the newer CPU's and support XMP v1.3.

All are 1600Mhz, 9-9-9-24 timings, 1.5v, quad channel kits.

No, but there are a few dual channel kits that are marketed as both dual and 4 channel kits. It doesn't matter. A 4 stick dual channel kit is the exact same as a 4 stick quad channel kit. The main thing there is that all sticks were tested and packaged together.

Thanks peanutz, for this very clear and complete reply. So basically what you are saying is that my rig should be able to run a 4 stick matched set at 1600mhz and should require no extra voltage, but there is a chance that it will. How do I know when my system is fully stable? Do I use some sort of testing tool? Or will the rig just not boot?

Also you mean that memory with an advertised speed of 1600MHz is in fact running at 1600MHz because its actually 2x800MHz due to double data rate. So no need to overclock them to run at this speed right?

As for the difference in series, what your saying is that it does mostly come down to aesthetics for the 3 mentioned ram kits correct?

Is there a difference in things like cooling? The dominator platinum mentiones DHX cooling, how much added value is that with respect to the vengeance ram?

Thanks peanutz, for this very clear and complete reply. So basically what you are saying is that my rig should be able to run a 4 stick matched set at 1600mhz and should require no extra voltage, but there is a chance that it will.

Yuppers! Thats it in a nut shell!

How do I know when my system is fully stable? Do I use some sort of testing tool? Or will the rig just not boot?

Just test it with memtest86+. There is a link on the left if you need the download. But let it run for three passas. It will take a few hours to run, so best to set it to run before you go to bed and let it run overnight. But anyway. If by chance it fails, then take your voltage up to 1.55v and retest. Rinse and repeat until it does passs. Just don't go any higher than 1.65v with that CPU. If you have to go that high then there are other issues somewhere.

As for the difference in series, what your saying is that it does mostly come down to aesthetics for the 3 mentioned ram kits correct?

To a point , yes. A 1600mhz kit at say CL9 is of the Dominator Plats is going to run or perform the same as the same 1600mhz CL9 VengPro's. However those same two may provide a little higher headroom if you were planning on trying to overclock them past 1600mhz.

In real world performance such as applications and such, you'll see absolutely no difference. It would take a benchmark run to compare numbers to see anything tangible.

Is there a difference in things like cooling? The dominator platinum mentiones DHX cooling, how much added value is that with respect to the vengeance ram?

Reply With Quote

Again, we go back to that overclocking thing. If they are run at their rated speed of 1600mhz, then heat will be a moot point. They will be barely warm. Now if you were going to or planned on pushing them further then I would take the platinum's over the Vengeance seies my self. I have been a lifelong Dominator fan and can personally vouch for the cooling abilities of them. The Vengeance Pro's are supposed to handle heat equally as well, but i can't say from personal use. And the standard Vengeance series will also perform about the same when run at rated speeds, but you may find they might heat up a little faster than the other two.

If your looking for the best bang for your buck and for a good performance boost, then i would seriously consider getting either an 1866mhz kit or even 2133mhz. Both of those would run equally as well and are small easily achievable overclocks using the XMP profile. A couple of clicks and your off to the races. I can only assume your looking for 32gigs for a reason...maybe photo editing,videos or rendering? This would speed up all of those tasks. And if you are just wanting to because you can it will also boost gaming performance and frame rates.

I hope this helps answer some of your questions. If you have anymore, I check in about a half a dozen times or so each day.

Cheers!

Peanutz, you're awesome. Thanks for filling in the blanks.

I apologise for the late reply. I went ahead and ordered 32GB of 1600Mhz Dominator platinum. The reason I'm not going over 1600Mhz is that apparantly my motherboard has issues running that amount of ram at higher speeds. (Alienware explicitly warns about this fact, and I've seen other people around the web with issues caused by doing it anyway).

Computer booted fine upon installing the RAM, until I got a memory related blue screen.

I tried to run memtest 86+, but for some reason, I cant get my machine to boot from the CD. This can be two things: 1. I read somewhere that with windows 8, you boot in UEFI mode, while a boot disk like memtest requires EFI mode (legacy), if I set legacy in my BIOS, it still wont boot from CD and also does not recognise my hard drive(obviously for it is in UEFI mode). 2. I'm not burning the ISO in the right way. (tutorials say that the CD should be named BOOT after burning, but mine isn't. It says MT420, like the memtest86+ version i downloaded. Did I burn the disk wrong? Or is this a windows related issue?

Anyway, because I could not load up memtest86+, I looked for alternatives. Turns out there is a built-in memory test in Dell/Alienware computers called PSA (or ePSA in alienware). I ran this diagnostic, it performs the right tests like WMATS and stuff. It popped out memory errors on 1 of the 4 sticks.

So I RMA'd the RAM, and got a new set. This time, ALL 4 sticks showed errors in the PSA test almost immediately, I reseated the RAM and retried: no luck.

At this point I was beginning to think that PSA is possibly not reliable. So I looked for another alternative. Found out windows 8 ships with an integrated memory test called Windows memory diagnostic. I've been running it on extended tests for 4 passes per stick(takes about 3 hours per stick to complete those settings). So far I've tested two sticks and no errors show up.

If all 4 sticks pass that test, Is it safe to say the RAM is working correctly? Is it wise to run a stress test on all 4 in prime95? Or is memtest86+ really the only reliable memory test, and if so, could anyone help me to get it running? Usually I'm quite capable of getting things to work (especially with the infinite number of guides for everything on the web), but this has me puzzled.

Thanks in advance for any help/advice.

You really need to test with a strees test that does not work in a windows enviornment.

Memetest does just that. But your correct, I wouldn't trust the Alienware test or Win8

What kind of troubles are you having getting memtest to run?

And are you using a DVD or a USB stick?