Thursday, February 12th 2009
Phenom II AM3 Plagued with DDR3-1333 Issue
Barely a week into the introduction of the DDR3-supportive AM3 socket CPUs, the processors seem to be having design flaws. This, as circulated by AMD in its revision guide document for the 10h family of processors (found here, page 80). The issue, as described by AMD, centers around the DDR3 memory sub-system. On machines with more than one PC3-10600 (1333 MHz) memory module populating a memory channel, the users may experience unreliable operation. The company does not get into the specifics of the symptoms. This issue however, does not affect systems with a module per channel (one or two modules installed in the motherboard), and only those with three to four modules installed.
The AMD K10 memory controllers on AM3-socket processor provide a 128-bit wide memory interface (with DRAM Ganged mode enabled), which amount to two 64-bit wide memory channels. On most motherboards, four DIMM slots with two slots sharing a memory channel are present. With this issue, one is not recommended to use more than one DDR3-1333 memory module per channel. AMD recommends a quick fix for the issue for systems using more than one DDR3-1333 module per memory channel: to manually specify the memory to run at 533 MHz (1066 MHz DDR), and accordingly set DRAM timings. As a little compensation, one can tighten DRAM timings with the drop in frequency. AMD will fix this issue in the next stepping (sub-version) of the CPUs. The "x-factor" with this erratum revolves around DRAM voltage, a significant factor. One might note AMD saying "the processor memory subsystem may exhibit unreliable operation over the allowable VDDIO voltage range", which leads us to think if there is a potential workaround with adjusting the DRAM voltage beyond the allowable range (read: over-volting the memory). We hope to hear more from AMD on this.
UPDATE (02/13): AMD replied to the report, saying that work is in progress toward fixing the issue. While not getting into the specifics, AMD indicated to us that apart from addressing the issue, the company is also working toward something "which will make you 'feel cozy' about the DDR3 support". The statement is ambiguous, and is best left uninterpreted at this point in time.
The AMD K10 memory controllers on AM3-socket processor provide a 128-bit wide memory interface (with DRAM Ganged mode enabled), which amount to two 64-bit wide memory channels. On most motherboards, four DIMM slots with two slots sharing a memory channel are present. With this issue, one is not recommended to use more than one DDR3-1333 memory module per channel. AMD recommends a quick fix for the issue for systems using more than one DDR3-1333 module per memory channel: to manually specify the memory to run at 533 MHz (1066 MHz DDR), and accordingly set DRAM timings. As a little compensation, one can tighten DRAM timings with the drop in frequency. AMD will fix this issue in the next stepping (sub-version) of the CPUs. The "x-factor" with this erratum revolves around DRAM voltage, a significant factor. One might note AMD saying "the processor memory subsystem may exhibit unreliable operation over the allowable VDDIO voltage range", which leads us to think if there is a potential workaround with adjusting the DRAM voltage beyond the allowable range (read: over-volting the memory). We hope to hear more from AMD on this.
UPDATE (02/13): AMD replied to the report, saying that work is in progress toward fixing the issue. While not getting into the specifics, AMD indicated to us that apart from addressing the issue, the company is also working toward something "which will make you 'feel cozy' about the DDR3 support". The statement is ambiguous, and is best left uninterpreted at this point in time.
105 Comments on Phenom II AM3 Plagued with DDR3-1333 Issue
Want to buy an intel but I might have to someday.
with no incentive to work for/towards & being in the middle of the credit crunch I suppose being part of AMDs workforce must be a pretty demoralising job not to mention how low the moral is in the work place. im guessing - to some workers they think:
# AMD has already had its day so theirs no longer a need to be work like a dog.
# AMD are probably gonna cut staff anyway - why work like a dog if you're not gonna be in the same seat next week??
# AMD are fighting a losing battle - why work like a dog if their efforts will never achieve anything??
#AMD care cutting staff salaries & bonus's - great! why come into work at all.
& so the curtain calls.....
But whenever a solution pops up I'll play again, I've been having fun with the 'new AMD' so far :toast:
Really if more than 2 people read the reviews you would have seen very clearly in the stats and mentioned by the reviewer that the motherboards in their current state will only officially support one dimm per channel at 1333Mhz. Anything other than that would be downclocked to 1066Mhz.
I'm ashamed that after all that talk of how great AMD was doing you "all" appear to have lost faith because they told you beforehand that something was not officially supported in the first place. :shadedshu:
Again I say, yes it sucks for a select few that you MAY have problems running 4 dimms at 1333Mhz or higher.....but if you could read something other than "bad" news you would've known beforehand.
Kei <--- f*ck the flame jacket :D
Running 4 DIMMs is always a bitch. My Gigabyte P35-DS3R mobo automatically cranks the latency behind the scenes and needs more MCH volts. This isn't some modern revelation either. You can wander back in time to the days of Pentium and K6s and see that loading up all banks on the mobo will result in the need to turn down the latencies. More DIMMs/SIMMs = more difficulties in signal integrity.
So far in my experience I haven't had too much trouble running with 4 dimms with my PII 920 AM2+ setup. The only 'problem' I have is that I can't get 1066Mhz stable using all four dimms using the 2.66 divider. I'm running two different types of ram right now which of course has something to do with it.
However both sets (4Gb total) run in excess of 1066Mhz by themselves. One set runs up to 1150Mhz and the other runs to 1120Mhz. No problems with them seperately, but if I use all four dimms then I can't get them stable using the 1066 (2.66) divider so I just use the HT Bus to clock them up using the 800 (2.0) divider instead to whatever speed I need. I usually just leave it at 1000Mhz which is close enough and I like whole numbers. :p
The 4 dimm thing isn't really new at all, and goes back a very long time. Heck, I'd be just fine with 8Gb 1066Mhz DDR3 haha. Once the new AM3 boards are a bit more available I"ll very likely give it a go. It'll probably be with 4Gb though in the beginning anyway....I think. ;)
Kei
(btw, I'm wondering when any other site is going to cover this "issue")
And oh, AMD responded to our email, we're expecting another short reply, after which I will update the news.
EP35C-DS3R x2
P35C-DS3R x2
P35-DS3R
EP35-DS3R
None of those needed any extra volts at MCH and they have mild OC's. I still have two in my house. Thats the most of the same series I have ever had. Funny thing you use that as an example being one of the best OC'ers of their time for the money. A few of those had 64bit Vista and 8GB of RAM like the one I gave to my GF. It had no issues that I have seen. :wtf:
I think your issue is the RAM you used. Its probably not on the Supported memory list. Common mistake really.