hallo, i have the problem here why my my main board detect RAM 6400 not 8500 ok firstly im using the hyper x 1066,on cpu z also 6400 not 8500. i use e45 main board gigabyte. ill try to set the letecty like factory kinstone give but the number on the cpu z also the same.
DDR2-1066 is not DDR2-800 that runs at 1066 by overclocking.
DDR2-1066 does indeed have a SPD byte which will identify it as such when the proper conditions are met - that means it's not overclocked DDR2-800. It really is 1066 RAM - it may have been made by the same process as a company's 800 RAM, but it doesn't leave the factory without being changed from the 800 by having the SPD set.
And overclocking, traditionally, refers to modifying related BIOS settings to create something that doesn't otherwise exist in the BIOS. That's not what's done with 1066 RAM unless the board specs say "1066 OC". I have a setting that is proper in my BIOS for 1066 RAM, just like I have one for 800 and 667 RAM. I'm not overclocking at all by putting in 1066 RAM and setting the motherboard to the same.
If you have a motherboard that supports 1066 RAM, then first set the DDR voltage to match the specs of the RAM you purchase, then it may come up as 1066 automatically. If not, then set the speed as 1066 in the BIOS manually.
Once the POST sees the RAM as 1066, it means it's found the spd for the 1066 mode and CPUz should show it as both a 1066 model number and the spd for it. If the POST - when the system boots and shows the RAM speed - shows 800MHz RAM, then CPUz will only show the spd up to the 800 mode.
What does JEDEC standard say for ddr2 1066, I believe its 1.8v nominal 1.9v max, CL5. I think unless those requirements are met by the memory chips, the RAM module cannot have SPD set for 1066, CL5 @ lets say 2.1v, as that would be outside the JEDEC standard, right? Those settings can be stored as an EPP profile in SPD, but have to be manually loaded, right? So I guess were talking about the "real" 1066 that can run at 1.8v, and "factory overclocked" 800 that is sold as 1066, because it can achieve that frequency but only with more voltage pumped into it.
And in my book going beyond standard voltage to achieve higher frequencies is overclocking.
I agree that going over recommended voltage is overclocking. But what is the standard? Are we talking about the industry standard? IEEE? JEDEC? Look at all the DDR3's that require more than 1.65V - they break Intel's standard. Clearly, the current industry standard for 1066 modules is voltage over 1.9V - not a good thing in my opinion.
A quick check of DDR2 800 at NewEgg shows 170 modules. Only 63 are 1.8V - the "standard" for DDR2 800. So are all the others overclocked 667s? No, they're 800s that want more voltage. When you move to 1066 you find 54 modules - only 6 of them at 1.8V. Are these 6 really 800 modules - no, they're better designed modules that deliver 1066 at lower voltage. Some of the others, especially those wanting over 2.1V are no doubt 800s that are being pushed to 1066 by more voltage - but they still have 1066 SPDs. Not just a new stuck-on label.
I've used 3 different 1066 modules in my system so far. The first wanted 2.2V to run in 1066 mode - I couldn't do that. These were lousy RAM that needed high voltage to achieve 1066 mode - likely 800s being pushed too hard. But they still are 1066 modules because I've seen the SPD for them and they list that mode. (OCZ Platinum)
The 2nd only use 1.8V and come up automatically in POST as 1066, with no changes to timing, freq or voltage. They are without a doubt, REAL 1066 modules. Unfortunately, they had lousy timings and I returned them. (Crucial/Micron)
The 3rd come up as 800s until I set the freq to 1066, then come up as those. Again, real 1066 modules. (G.Skill Pi)
The 1st modules were made by OCZ. OCZ now makes a 1.8V version of their 1066 modules to avoid these problems. Again, real 1066 modules. (Also OCZ Plat but not widely available)
Let's agree that some modules sold as 1066 are "factory-altered" 800s - using higher voltage to perform at 1066s. But to say that 1066 modules really don't exist is false - they do and it's not just a marketing scam. Companies are putting out real 1066 RAM and I think we'll see much more of it in the near future as they get tired of people returning their "fake" 2.2V modules.
Im not up for a discussion here really, so in short: JEDEC sets the standards for memory, not ram manufacturers, not intel, not ieee. If i can remember correctly: voltages required to operate:
ddr2 - 1.8v nominal, 1.9v max(but must withstand 2.3v before permanent damage is made),
ddr3 - 1.5v nominal, 1.6v max(must withstand 1.9v before permanent damage is made)
Nobody said real ddr2 1066 doesnt exist, but majority on the market are those "fake" ones, and i dont believe in them having SPDs set for JEDEC 1066 as that would contradict the standard. They might have EPP profile for that though. And its the same with ddr3 that requires over 1.6v
Whatever is outside max voltage specified by JEDEC is not real quality product but lesser quality product forced to work(with more voltage) as if it was the quality one. And if someone bought one of those products, he pretty much has himself to blame. Thats the world we live in, unfortunately.
I'm hoping that as more and more people return their high-voltage RAM the manufacturers will get the message. But, as you said, too many will buy the RAM, put it in, and never know the difference.
Again, I started out as one of those folks, buying for price without seeing the reasons for the nice price. I got lucky since my motherboard wouldn't even allow 2.2V so my memory never even pretended to work at 1066.
Yeah, IMO that practice should not be allowed, is should be more like it is with cpus - we know it may run way faster than spec(additional voltage might be required) - its up to end user if he wants to run it faster. Or some sort of big red label "requires more than standard voltage - check compatibility" should be clearly visible. But its not, so what can you do, nothing really apart from not buying those products.