Problem with 16GB on 2010 iMac 27" i5 & i7.

With the 14 GB set up you have unmatched RAM. The downside is that there is a theoretical speed penalty. Some believe that any performance hit due to mismatched RAM is compensated by the increase in performance of the extra RAM though.

Here is one source. They site 3% improvement with matched pairs.
http://guides.macrumors.com/MatchedRAM_on_IntelMacs

But I agree with you - this could be an unverified legend. You should check benchmarks with 12 GB, 14 GB and 16 GB. It would be most informative.

I'm experiencing the same annoying issue with an iMac Corei7 but with 1066MHz soDIMM modules.

The system was able to run with 12GB RAM (2x2GB and 2x4GB). I decided to update the memory to 16GB. And the system doesn't start. But what's strange is that the system now have problem with 12GB as well!

I tried different combos:

2x4GB hynix modules + 2x2GB hynix (doesn't work or start randomly, once the system run, no problems at all)
4x4GB hynix modules (doesn't work or start randomly, once the system run, no problems at all)
2x4GB Transcend CL7 OMD22 D9LGK (it works)
2x4GB Transcend CL7 OMD22 D9LGK + 2x2GB Samsung CL7 (doesn't work or start randomly, once the system run, no problems at all)

The first configuration, before the 4x4GB update was 2x4GB hynix modules + 2x2GB hynix

Here is an other article, although it concerns an other mac, the theory is the same.
http://eshop.macsales.com/Reviews/Framework.cfm?page=/Benchmarks/Core2RAMbenchorig.html

.
Funny that it is written by OWC.

peterwillem

G'day everyone,

I've had my iMac for about two weeks and got this problem as well, glad to find a solution, sad to see the problem in the first place ;-(

I worked as an apple tech a few years ago and saw a similar problem. It took me five weeks of testing to discover that the ram was faulty, the machine always failed while no one was watching, or didn't fail at all. One day I just happened to be watching the ram test when it shut down. Ran a PC test on the ram and confirmed the problem. So when I first saw this happen my immediate thoughts were ram.

I too have OWC ram, and have now changed my configuration to the suggestion posted here. It's a shame that I can't have my 16Gig of ram, but then I'm not sure if I would ever use that much anyway.

I hope Apple are working on a firmware update.

By the way I upgraded from a PowerMac G5 1.6Ghz with 2Gb of ram so I'm lovin' it 😉

Michael

What a thread! I just stumbled across this, and finding this thread would have saved me a considerable amount of time. I only found it by putting in my model number. I thought I'd be safe buying a couple months down the road, after I could see what the issues were.

Unlike your reports, I can consistently trigger a complete and sudden power off by opening a file in Photoshop.

I have iMac Model iMac11,3 Intel Core i7.

I purchased 16GB 1333 MHz DDR3 RAM from OWC with Micron chips. I used memtest and tested each individual module, by plugging them in one by one. I ran the test once, but each module checked out. I had to do this, because memtest will not lock memory above 4GB or something. I then ran Rember in OSx 64-bit mode to test the memory. No errors there, but you can only test what you are not using in the OS, applications, drivers, etc.

I can consistently reproduce the problem with 4x4GB modules. If I take one out, it works. If I replace it with my Apple 2GB, it works.

I just got back from the Apple Store. They were going to replace my machine on the spot, but they did not have one available. They did allow me to try it with Apple RAM. The machine did not crash with the Apple RAM, but this test appears to be inconclusive, as others do not exhibit problems right away with new RAM. I did not see what chipset their RAM was, as it had a big sticker across it.

I'm unwilling to spend $2500 + extra RAM on a machine and not be able to upgrade to 16GB. Or, use part of it for years and then spend for 16GB again. Sorry.

I've just been through this thread, as well as the other two or three referenced. Is this the consensus so far for those shopping for RAM?
4x4 = 16GB Apple RAM (Micron chipset, possibly others) works sometimes
4x4 = 16GB Transcend Hynix chipsets work
4x4 = 16GB Transcend Micron chipsets work
4x4 = 16GB Samsung Factory Original chipsets work

All others have failed numerous times, unless combined in ways that result in less than 16GB total.

FWIW I upgraded to 16gb on my new iMac from OWC as well..was thinking it was RAM but after watching the logs noticed something peculiar...switching from Safari to Chrome has cured my intermittent crashes.

Don't recommend the newest Safari.

This is an interesting thought. Keep us posted on the progress, if it really keeps on working.
If it does, I will give my 16 G OWC ram another chance while changing to another browser, even though I have no idea as to why.


peterwillem

If a machine is shutting down, producing odd screen results, or even hard freezing, then the issue is a hardware issue, not Safari. I know some get the beach ball, which could be software.

If you have a spare drive, you can install the OS there and test it out. I did, and my machine still powers off, on command.

It may not be RAM, but something else with the hardware. I wasn't convinced of this myself, but I've read too much in other forums that made me realize that much of what I read in this thread is a hardware issue. You change your RAM configuration and it works? It's a hardware issue. I took my machine in last night, and the store was more interested in replacing my system than I was. (Unfortunately, they did not carry my model.)

Here's what an Adobe developer said about my issue. He read this thread last night before he made this reply:

"It's possible that all of them have a timing problem with the motherboard, or a power draw that exceeds the motherboard design.

They could be on the edge of termination resistance and causing problems as well (all 4 are needed to push it over).

But my guess is a power draw issue -- and that could be a fault in the RAM, or a motherboard that isn't supplying enough power (even a single bad decoupling cap could do it)."

I agree with you on the hardware issue. The Safari-other browser solution was to simple to be true.
So if I read and understand your post correctly you and/or the Adobe guy are saying that:

Apple ram, Samsung factory and transcend ram are drawing less power from the mobo than all the other rams , who are causing the shutdowns.
I don't think it is likely that the various mobo's of the people ,whom are having these problems, are not supplying enough power. They are spread out world wide and have all the same type of iMacs. That would be to much of a coincidence.

peterwillem

Good question. I'm not certain of the answer. The Adobe developer seemed fairly knowledgable. As he said, I don't think we can really know unless you can measure it as it occurs. I worked for Novell years ago. Our developers did not usually have success in correcting issues if they could not reproduce and observe it. The same would apply to hardware.

Given this is an issue worldwide, then it's likely a system issue, and not one with power regulation.

Others speculated wildly in the Photoshop forum based on issues they had, but we know from our results that it's the 16GB combination that triggers it, as well as a few other odd combinations.

Any positive or negative feedback from those with Samsung, or Transcend (Hynix/Micron) RAM?