Friday, October 3rd 2008

Theory of DDR3 Voltage Limitations for Bloomfield Gains Ground

Nehalem does promise to be a processor to look out for, it would be Intel's next installment, post the successful Core 2 series processors. This time however, Intel made a core modification with the way the system handles memory. The Bloomfield processors house a massive 192-bit wide memory controller for supporting tri-channel DDR3 memory. It however was found that the controller could bring in limitations to the DIMM voltages that the system could support.

The retail version of ASUS P6T Deluxe OC Palm Edition motherboard was unboxed by XFastest. Being the retail product, as usually, it comes with precautionary labels attached to parts of the motherboard. The one that covers the 6 DDR3 DIMM slots reads:
According to Intel CPU SPEC, DIMMs with voltage setting over 1.65V may damage the CPU permanently. We recommend you to install DIMMs with voltage setting below 1.65V.
It could have implications on the current DDR3 memory market as well as you, if you happen to have DDR3 modules, which you plan to retain for use in the future platform, that operate above the said voltage. It also means that in the near future, we could be seeing memory sticks that facilitate overclocking at much lower voltages. From a technology standpoint, companies such as Samsung, Elpida, Micron, etc., are working on releasing DRAM chips based on newer silicon fabrication technologies, that operate at lower voltages.
Add your own comment

89 Comments on Theory of DDR3 Voltage Limitations for Bloomfield Gains Ground

#1
MrMilli
fitseries3 said:
why release a motherboard or better yet an entire platform when nothing that exists today will work with it? sure... maybe a few sets of ram will work, but why even have ddr3 when its gonna be running ddr1 and ddr2 speeds?

im sure memory companies will make ram that complies but jeeez.... it would be better for intel to unlock this BS voltage link.
Dude only OC kits need higher than 1.5V.
Any standard PC1066 or PC1333 memory module needs 1.5V. (= 99% of the market)

Example:
http://www.ec.kingston.com/ecom/configurator_new/partsinfo.asp?root=us&LinkBack=&ktcpartno=KVR1333D3N9/2G

Why do people say stupid stuff? Nobody knows ...
Posted on Reply
#2
btarunr
Editor & Senior Moderator
zithe said:
It's not like they can prove you were using high voltage sticks, can they? Give it a shot and RMA it in hopes they'll take it. (That's if it's true.)
The only thing is, it's your processor, and not motherboard or RAM that gets bricked. 'Tis a painful task...to RMA a processor.
Posted on Reply
#3
Fitseries3
Eleet Hardware Junkie
you dont call ppl stupid on TPU and get away with it.

the fact is nehalem operates around 1.0v and ddr3 CANNOT OPERATE AT THAT VOLTAGE.
Posted on Reply
#4
s3rv3r
My OCZ DDR3 1600 PLATINUM 1.90V, I can put on garbage, because the Core i7 just work with memories 1.65V. If the people have DDR3, can forget. You need spend more money with DDR3 AGAIN
Posted on Reply
#5
btarunr
Editor & Senior Moderator
MrMilli said:
Dude only OC kits need higher than 1.5V.
Any standard PC1066 or PC1333 memory module needs 1.5V. (= 99% of the market)

Example:
http://www.ec.kingston.com/ecom/configurator_new/partsinfo.asp?root=us&LinkBack=&ktcpartno=KVR1333D3N9/2G

Why do people say stupid stuff? Nobody knows ...
At least going by the Newegg, a sizable number of DDR3-1333 (the standard Bloomfield IMC officially supports), are > 1.5V kits.



^ PC3-10600 kits http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=2010170147+1052129233+1052429248&Configurator=&Subcategory=147&description=&Ntk=&SpeTabStoreType=&Order=BESTMATCH&srchInDesc=



^The only PC3-10660 kit there



^ PC3-10666 kits http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=2010170147+1052129233+1052430325&Configurator=&Subcategory=147&description=&Ntk=&SpeTabStoreType=&Order=BESTMATCH&srchInDesc=

Click on the links > advanced search > click on the "Voltage" drop-down, to see for yourself.

And please, don't call people stupid, we are t3h ov3rcl0ckers. We are worried because our overclocking gets into a mess, at least with the memory modules you get today.
Posted on Reply
#6
Fitseries3
Eleet Hardware Junkie
no one is seeing my point at all.
Posted on Reply
#8
hat
Enthusiast
I find it hilarious that Nehalem only officially supports 1333.
Posted on Reply
#9
KBD
fitseries3 said:
no one is seeing my point at all.
I think what hes trying to say is that most DDR3-1066 and DDR3-1333 operate at approx 1.5 to 1.8 and if someone wants to overclock them they will have to raise the voltage and since they cant go over 1.65 this will limit the RAM overclock. Furthermore DDR3-1333 is inferior to current DDR2-1066, besides some DDR2 operates at 1200+ at good timings while DDR3 modules dont so there is no advantage to using these DDR3 modules.

@hat, i beleive its 1066, 1333 is overclocked not officially supported, correct?
Posted on Reply
#10
phanbuey
I wonder if this will affect all boards or just the X58 (and other boards from Intel Chipsets)... This would be a perfect time for nvidia to bring out a board with unlinked voltages.
Posted on Reply
#11
pjladyfox
fitseries3 said:
why release a motherboard or better yet an entire platform when nothing that exists today will work with it? sure... maybe a few sets of ram will work, but why even have ddr3 when its gonna be running ddr1 and ddr2 speeds?

im sure memory companies will make ram that complies but jeeez.... it would be better for intel to unlock this BS voltage link.
I'm thinking, and keep in mind this is just a guess, that Intel may be trying to force either individually or both JDEC and the RAM makers to come up with a ratified spec for memory speed and voltage. If memory serves JDEC was going to ratify at one time a DDR2/DDR3 2.0V spec but then memory makers just started cranking out faster and faster speeds while bypassing the verification process and they just stopped just short of making it final figuring they were going to do whatever they wanted JDEC be damned.

Personally I think this entire increasing voltage/speed willy-nilly is not really doing us any favors. I mean why the heck should you have to go in and manually tinker with your voltage and speed settings just to get something to run as advertised? Do not get me wrong I like tinkering as much as the next girl but this has gotten to insane amounts of silly in a rather short period of time.

I guess we'll see one way or the other as time goes by and word gets out about this.
Posted on Reply
#12
Fitseries3
Eleet Hardware Junkie
what im saying is this...

nehalem operates at ~1.0v STOCK CLOCKS

IF the ram voltage is the same 1.0v THE BOARD/RAM WONT EVEN POST

to get the ram to boot you'd have the chip pretty close to its max voltage.

to solve this memory manufactures would have to develop ddr3 that can run around 1.0-1.6v depending on clocks. that is a tall order if you ask me.
Posted on Reply
#13
robal
One thing I don't understand is why Athlon could support all DDR voltages from it's early days.

Does anyone know what's technically different in built in memory conroller in K8 / K10 ?
I mean... voltage-wise..
Posted on Reply
#14
Fitseries3
Eleet Hardware Junkie
idk..... i dont understand why the voltage is linked to the cpu at all.
Posted on Reply
#15
robal
Well, since the memory controller is integrated into the core, you have to think about CPU voltage.

There are obviously some drivers that seperate CPU logic from package pins, so there is possibility to 'translate' voltage.

I don't understand why it's so limited.

Athlons can drive 2.2V DDR2 while core voltage is at 1.1V for example.
Posted on Reply
#16
tkpenalty
fitseries3 said:
what im saying is this...

nehalem operates at ~1.0v STOCK CLOCKS

IF the ram voltage is the same 1.0v THE BOARD/RAM WONT EVEN POST

to get the ram to boot you'd have the chip pretty close to its max voltage.

to solve this memory manufactures would have to develop ddr3 that can run around 1.0-1.6v depending on clocks. that is a tall order if you ask me.
Chill out mate. The memory controller and the RAM are synced only.

The memory controller syncing is purely only a monopolistic deterrent which Intel is using since AMD's offerings wont provide any dividends to users; i.e. AMD's proccessors are having a hard time competing. Thus Intel CAN "screw the customer over" by preventing the users from overclocking through this so-called design flaw.

Sure it might turn people away, but then again when you look back to AMD's offerings you'd probably look at the other Core 2 offerings instead.

Basically in short intel doesnt want you having as much value for money as possible; by running low end core i7 as fast as the top end one. Theres NO competition in the CPU market atm and this is one of the repercussions to the consumer.
Posted on Reply
#17
$ReaPeR$
maybe intel is a little behind because this is its first try with integrated mem contr or they do not want the i7 to be oc able because none would by their top series cpu at around 2000$ maybe....
Posted on Reply
#18
Poisonsnak
tkpenalty said:
Chill out mate. The memory controller and the RAM are synced only.

The memory controller syncing is purely only a monopolistic deterrent which Intel is using since AMD's offerings wont provide any dividends to users; i.e. AMD's proccessors are having a hard time competing. Thus Intel CAN "screw the customer over" by preventing the users from overclocking through this so-called design flaw.

Sure it might turn people away, but then again when you look back to AMD's offerings you'd probably look at the other Core 2 offerings instead.

Basically in short intel doesnt want you having as much value for money as possible; by running low end core i7 as fast as the top end one. Theres NO competition in the CPU market atm and this is one of the repercussions to the consumer.
Agreed. Plus think about it, if the bloomfield boards seriously had to run DDR3 at 1.0V to post, and as a result no boards would post, do you honestly think intel would sell any?
Posted on Reply
#19
phanbuey
tkpenalty said:

Basically in short intel doesnt want you having as much value for money as possible; by running low end core i7 as fast as the top end one. Theres NO competition in the CPU market atm and this is one of the repercussions to the consumer.
I agree. :toast:. Nvidia chipsets FTW!
Posted on Reply
#20
Scrizz
phanbuey said:
I agree. :toast:. Nvidia chipsets FTW!
NVidia can't help you at all :slap:
the mem controller is in the CPU :laugh:
Posted on Reply
#21
CDdude55
Crazy 4 TPU!!!
That site that unboxes it looks like it has lots of viruses.
Posted on Reply
#22
CDdude55
Crazy 4 TPU!!!
Scrizz said:
NVidia can't help you at all :slap:
the mem controller is in the CPU :laugh:
lol

But i do believe, Nvidia chipset FTW.
Posted on Reply
#23
btarunr
Editor & Senior Moderator
CDdude55 said:
That site that unboxes it looks like it has lots of viruses.
What part looks like a virus? Is your AV program showing up something? If so, I'll rehost the images for you. If you're refering to the site text looking garbled, you need the Chinese font installed to display as well....Chinese.
Posted on Reply
#24
Frogger
$ReaPeR$ said:
maybe intel is a little behind because this is its first try with integrated mem contr or they do not want the i7 to be oc able because none would by their top series cpu at around 2000$ maybe....
:toast: Very close to the TRUTH ;);)
Posted on Reply
#25
Mussels
Moderprator
This is a good way to stop OCing based on FSB. It will make OC'ers need to use multipliers on extreme/unlocked CPU's.

(not that i LIKE this idea)
Posted on Reply
Add your own comment