Multi CPU's Not the same revision

I am not running it overclocked I was just testing the capabilities of the system. The CPU is a "Black Edition" with unlocked multiplier. The subject problem was happening BEFORE any overclock. I am fully aware of the risks involved with any modifications that alter factory specs. but thank you for the reminder.

 

From what I understand, ACC automatically overclocks the system. You don't have to do anything - you just have to leave it On in BIOS.

 

My understanding was that ACC helped the system remain stable during overclock AND underclocking the system weather it is by manually changing the settings or by software changes such as Cool & Quiet, which still works with or without ACC. But by your explanation it got me thinking of the consequences to the end user of such a feature. Please correct me if I am mistaken here. I will use C&Q as an example. C&Q changes the settings to achieve an under or over clock situation based on user demand (among other things like cooling etc.) Hardware manufacturers all state that modifying factory settings voids the warranty. Does this mean that C&Q will potentially void the warranty of the CPU? If so then a large percent of users out there, myself included, are voiding their warranties by using the features that the manufacturers recommend you use. Or am I just blowing smoke up in the air?

 

C&Q is for keeping the CPU temp. & power consumption down. It does NOT overclock the CPU at any time. It only reduces the clock speed.

ACC on the other hand is for overclocking and is a feature supported by AMD. As such I don't think you are voiding your warranty.

 

Hockalugie,

I had never heard of Advanced Clock Calibrations but there are several things that fit into Rockster's "unheard of" category so I felt a speculative comment might not be too far out of line. I am not aware any feature that automatically overclocks any processor and would appreciate RSInfo's clarification on the statement made in his next to the last post.

My speculative thinking is that a slight boost to your core voltage may correct this problem. Not knowing what you are running for a PSU and thus not knowing what the +/- voltage tolerance specs are, it is possible that you are slighly underpowered. Check not only the voltage settings but also the current outputs as itemized in your BIOS. You can leave all power settings alone except core voltage which can be adjusted manually. Try a very small incremental change and report back with any results. If as stated, ACC helps create a more stable overclock, then I would suspect it has some influence on core voltage and I would further suspect that it "automatically" adjusts core voltage when either the bus speed (technically a misnomer for AMD CPU's but is more understandable theoretically) or the multiplier is increased.

Again - this is speculative but from my perspective it makes some sense and may correct your problem.

 

Rockster2U,

Till I read this post even I had forgotten about this AMD feature. It has been discussed last year in Anandtech http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3360&p=1.

Regarding Auto overclocking, ACC has an Auto setting too so if you leave it at that it tries to overclock on its own. Some motherboards may have implemented it differently http://www.xtremesystems.org/Forums/showthread.php?t=203890

 

Thank you for your suggestion to increase the CPU voltage. I have tried that. The M3A79-T Deluxe mobo allows voltage changes in .0125v intervals. 1.312v is default for the 9850 BE CPU. I did not change the voltage setting to get the overclock I mentioned above so I was reluctant to increase it to try to solve this "problem ". But, the published limits are 1.55v so, I bumped it up 0.0125v at a time, to 1.374, to no avail. I still think it's the mobo. This seems to be a rare situation which leads me to believe it's a hardware defect.

 

rsinfo - thanks for those links. Very interesting reading. I think there is a little confusion in your understanding of the auto setting re: "ACC automatically overclocks the system" but that doesn't diminish the value of the information you provided. Again, thanks.

Hockalugie - it certainly sounds like your investigation has been pretty thorough and its difficult to argue against the possibilty that you have a defective board. Before you RMA it, you may want to try reseating the CPU and you would be wise to check out the CPU pins and the socket itself under significant magnification. This behavior is not indicative of any kind of major issue but could be a result of some kind of very minor contamination affecting conductivity.

Good Luck.

 

Good idea! I,m so busy with work though it will be a week before I get to it but I will give it a go. Thanks