iMac (Late 2013) 27" i7 Haswell constant crashing with 32GB RAM

Agreed. I can make any of the 3 brands I've tried, work. As long as I took out the last stick, and only used 24gigs worth of RAM. This is a 32gigs only discussion, at least for me.

Yes, this is absolutely a problem specific to 32GB in use on the Late 2013 iMacs. I also never had any problems with any other configuration.

So far so good with my replacement Crucial RAM, although it is obviously far too early to know for sure. One thing I noticed about the new modules is that they are the high density version, meaning they have 8 total memory chips on the module (4 to a side) as opposed to the low density sticks (8 chips each side, 16 total) that I returned. This can also be discerned from the full model numbers, in particular the last few digits on the labels on the DIMMs themselves. Thus, I know that Ginooken is also using the high density (8DED) even though he never said so. I have no idea if the density has any relevance to anything, but I am making note of all details in the hopes that we may be able to pinpoint some consistency in the problems.

I checked with Crucial and they confirmed that both low and high density versions are completely compatible with the Haswell iMacs. When you order it, they just ship whichever one they have in stock. The Crucial model/part number for both is the same. As mentioned before, the SPD specs for this set are identical to the Apple Elpida, unlike my last set. The Apple Elpida is low density.

I continue to follow this thread closely and am hoping for a resolution from all kernel panics for everyone here.

I'm going to add another data point to this discussion. I bought two late 2013 iMac 27" 10 days ago. One has 16GB and a 1TB serial ATA drive; the other 32GB and a 1TB SSD drive. All memory is from Apple. The 16GB machine has been running continuously for 10 days without a single crash. The 32GB machine crashes at least once a day (screen goes black and it reboots itself). Happens whether I'm using it or not. I've sent all the crash reports to Apple.

Specs of the crashing Mac:

3.5GHz i7, turbo boost up to 3.9GHz

GTX 780M 4GB GDDR5

32GB RAM

Mavericks 10.9.1; loaded 10.9.2 yesterday and have already had a crash since.

Has passed all memtest runs. Off to the Genius Bar Saturday. Not optimistic.

Colojo wrote:

I'm going to add another data point to this discussion. I bought two late 2013 iMac 27" 10 days ago. One has 16GB and a 1TB serial ATA drive; the other 32GB and a 1TB SSD drive. All memory is from Apple. The 16GB machine has been running continuously for 10 days without a single crash. The 32GB machine crashes at least once a day (screen goes black and it reboots itself). Happens whether I'm using it or not. I've sent all the crash reports to Apple.

That is strange! Please follow up with your results after visiting Apple. Have you tried updating it to 10.9.2? Are there kernel panics in the logs? Sometimes they can glean valuable information. With the last set of Crucial I returned I saw memory corruption noted and indeed they failed a Memtest.

That might not be RAM at all but something else bad under the hood like a faulty GPU or logic board.

Not to jinx myself here but all's still well with my Crucial replacement set. 🙂

Update on my new 32gigs of Crucial Ram(part#CT4951624 on website):

After several other types of RAM tried with the 32 gigs, I can say today with the new Crucial; 4 days only 1 crash-- which was when I left the machine running all night to test the robustness which we need in my line of work. All other days, flawless with no crashes. And very quick. It's been a tough 3 month slog, but,

Success! Thanks guys for sharing your success, as well as thanks to Crucial!

Hi, I would like to double confirm that my system is working for 68 days without any problem/reboot/crash. This is for the people suffering this problem, at least you can have some hope, my nightmare took me 55 days till I got a good RMA'd pack of memories.

Summary: Got the iMac 27” Late 2013 + 32GB from Crucial on Nov 1st 2013. Multiple crashes as everybody when maxed out (32GB). Best workaround 24GB (8 Apple + 16 Crucial). RMA’d the iMac once. RMA'd Crucial memories twice. With 2nd RMA received a new "version" which is working perfect. WORKING (2nd RMA): SKU CT2C8G3S160BMCEU.C16FER. 8GB DDR3L-1600 SODIMM MAC. SPD Data Part Number: CT102464BF160B.C16 (Reported by Techtool Pro 7)

Booted on Monday Dec 23th with the new ones and since them I've had no problems, so 68 days since I made my last RMA the system is working perfect without crashes.

Regards,

Luis

Thanks for the update, Luis. Actually, I'd like to especially thank you because your encouraging posts about the Crucial back when I was just a lurker in this thread were a part of my inspiration to order the Crucial myself from the US. I appreciated you posting your detailed part numbers and SPD screenshots from TechTool Pro as well.

Cheers.

Joining the parade...

I suspect this has nothing to do with RAM, although Apple Tech Support was clear with me that non-Apple peripherals are the usual cause of Kernel Panics. (hmmm...maybe yes, maybe no?)

I THINK - at least in my own case - it's Chrome that's causing the problem. So...far...after removing Chrome...kernel panics have stopped. This, of course is a real problem (at least until I retire) since I maintain three Gmail accounts and the switch accounts feature in Chrome is really handy.

I have the same setup as most of you - 27" iMac - 32GB RAM (all Apple RAM) - 1TB SSDD.

Like an IDIOT, I wasn't reading early crash reports - WHAT...am I now an unpaid beta tester for Apple, Google, Amazon, all the rest??? ANSWER: YES - WE ALL ARE BETA TESTERS!!

Anyhow...the last two crash reports SEEMED TO INDICATE (I'm not a native Sanskrit reader, so I can't be sure)...Chrome "helper" just before the crash. So...I suspect that's the problem.

I have absolutely NO ISSUES with Chrome on my work MacPro, but that's a late 2009.

I've also noticed a lot of log entries concerning Adobe Creative Cloud using outdated drivers (or something foolishness like that).

But...all in all...I'm happy as a pig in slop with my new iMac. I'm also glad the problem is not my external WD HDD.

Just keep telling yourself...every day...in every way...I'm getting better and better! (Sometimes it helps)

John

Got my iMac back from the Genius Bar yesterday morning. Their tests showed no problems with the RAM, logic board, GPU or any other hardware, so they reinstalled Mavericks 10.9.2. Brought the computer home, fired it up and it's been running just fine ever since. This is the longest it has gone without crashing.

Keeping my fingers crossed... I'll post another update at the end of the week.

Hi John,

John.Orban.who.lives.in.USA wrote:

I suspect this has nothing to do with RAM, although Apple Tech Support was clear with me that non-Apple peripherals are the usual cause of Kernel Panics. (hmmm...maybe yes, maybe no?)

Err, sort of. Hardware issues (even Apple RAM can be or go bad!) are the usual cause of kernel panics. Did you speak to them over the phone, or have they had a look at your machine? Have you tried running the Apple Hardware Test (hold down D at boot)? I would really recommend trying to run a few passes with Memtest86 in multi-CPU parallel mode as well. I wrote about that earlier in this thread.

I THINK - at least in my own case - it's Chrome that's causing the problem. So...far...after removing Chrome...kernel panics have stopped. This, of course is a real problem (at least until I retire) since I maintain three Gmail accounts and the switch accounts feature in Chrome is really handy.

Actually, I think what it probably indicates is that Chrome was the active app when the system panic occurred. I've seen others with the RAM issues report Chrome in their panics.

That said, there are alternatives to the Chrome Switch Accounts feature in other browsers like CookieSwap for Firefox. 🙂

After running the tests above, if you still get panics and haven't taken the machine to Apple yet, I'd recommend doing so.

Dan

I've spoken with Tech Support and they were the ones who told me that hardware (peripherals usually) are the cause of kernel panics.

All of the hardware devices I had connected worked fine with my Late 2009 iMac, but that doesn't mean anything.

The only devices I had connected were an Apple DVD drive, WD Studio II External HDD, DELL monitor, CyberPower UPS (via USB cable). I had a kernel panic with the WD disconnected, so I know it's not that.

The Apple Tech said it could be the DVD drive. At this point the only thing I use it for is to rip my movie collection into iTunes which I've finished doing at this point.

I'd hate to think it's the DELL monitor, but even with the Late 2009 I had problems with the monitor going dark. I'd have to disconnect the adapter and reconnect and then it would start working again. I haven't had THAT problem since I got the new iMac, but maybe that's because it's causing the panic. I plan to get a second 27" monitor from Apple before I retire, so that should fix that problem if the monitor is the culprit.

I ran a hardware test this morning and it showed no problems, so I'm pretty sure it's a peripheral issue.

Makes sense about the active app. That's good to hear though. I'll check out CookieSwap. Always liked Firefox. I like the developer features in Chrome, but it appears Firefox has them as well (Firebug, etc.).

Thanks for the reply - it was helpful!

Well, back to the drawing board. My iMac crashed again this morning.

The only other new item in my setup is a Thunderbolt Drobo Mini. Drobo uses their own Thunderbolt driver, so I'm going to switch it over to USB and see if that helps.

Well, to clarify a little bit, what I found out in my long, terrible odyssey to getting 32GB working in this machine was not just that hardware issues are almost always the cause of kernel panics, but that a huge majoriry of those were bad RAM.

Do you have any hardware drivers loaded? They can also be problematic. Look in the kernel panic log for the line that reads

loaded kexts:

Anything non-Apple will be at the top of the list, and those are the only ones you need to worry about. Once you see Com.apple.xxxxx you're good. Those kext drivers from third party software have been known to be the source of kernel panics.

Other than that, no matter how you slice it kernel panics almost always point to bad RAM. I doubt it will be anything you're plugging into the machine. You're lucky since the RAM is 100% Apple. Do you have an Apple Store nearby? Take it in and let them swap it out for you. Believe me, it can pass an Apple Hardware Test and still be bad.

Have you tried making a bootable CD of Memtest86 from here?

Just burn the CD, pop it in your Apple DVD drive and then hold down the alt key when you hear the startup chime. Choose the EFI icon in the shape of a CD. Once it boots to Memtest86 hit Esc to get to the settings. You don't want to run the slow, default test. set it to run in Multi-CPU mode, Parallel. This way it will use all cores of the CPU at once when testing. A full pass should take about 30-40 minutes and if you see any errors in the tests at all you will have found your problem.

I recommend the same advice to Colojo.

Good luck!

Thanks, cdanr. I did run Memtest86 as you described and it reported no errors.

Here are the loaded kexts from this morning's crash... pretty damning for the Drobo. TrustedData is Drobo's old name:

loaded kexts:

com.TrustedData.driver.VendorSpecificType00

1.7.0

com.drobo.SCSI.ThunderBolt

1.1 [64171]

com.attotech.driver.ATTOiSCSI

3.4.1b1

<< com.apple stuff starts here >>

I'll check my previous crash logs as well.

Great news about the RAM, Colojo. There is an easy way to find out if the Drobo driver is behind the panics (and it may very well be). First, check with Drobo to see if there are any driver updates to the one you've got. I'd try upgrading first if there are any.

If that doesn't work or there aren't any updates, temporarily uninstall the driver using whatever uninstaller software/script Drobo provides and see if the panics subside.

I'm not sure what that attotech driver is but it may be part of the Drobo drivers.