Theory of DDR3 Voltage Limitations for Bloomfield Gains Ground

[btarunr]

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.

View at TechPowerUp Main Site

 

[Laurijan]

Damit that meens that the new Kingston HyperX 2000MHz rated @ 1.9V could fry Core i7 processors

 

[TRIPTEX_CAN]

Wow everyone who purchased DDR3 in the hopes of saving time and energy shopping for new RAM for their x58 builds just got screwed. Although, anyone who would switch to this new platform so quickly likely wouldn't care about the costs.

 

[Tau]

This is a downer all the current DDR3 ram is useless now, as most is 1.8V+

 

[Morgoth]

is it posible to instal the 1,8 volt ram and lower it in bios to 1.65?

 

[Tau]

is it posible to instal the 1,8 volt ram and lower it in bios to 1.65?

Sure, but it might not be stable.

 

[InnocentCriminal]

Why would it be useless? IIRC, you can lower the voltage of RAM as well as increase it. So what's stopping you lowering your RAM speed and voltage to the point where it works? I would have thought BIOS updates for the motherboard could add support for higher voltage RAM and it's most likely just a deterrent, it probably won't fuck anything up.

 

[Tau]

Why would it be useless? IIRC, you can lower the voltage of RAM as well as increase it. So what's stopping you lowering your RAM speed and voltage to the point where it works? I would have thought BIOS updates for the motherboard could add support for higher voltage RAM and it's most likely just a deterrent, it probably won't fuck anything up.

Just a deterrent? Since the memory controller is now on the chip i doubt that it is. Will have to see what the reviews say, and wait for someone to stuff on 2.2V+ through it and see what happens.

 

[npp]

It's nice to see that vendors will be pushed to comply with JEDEC standards at last. Most "high performance" DDR3 modules operate at voltages in excess of 1,8V - the maximum JEDEC allowed for DDR2 memory! I have rarely spotted DDR3 modules working at 1,5V, actually, and I certainly don't find this to be a good practice - if JEDEC state a maximum voltage for given memory technology, they surely have the reason to do so. But I guess it will be hard to justify water-cooled modlues working @1,5V, right .

 

[btarunr]

Just a deterrent? Since the memory controller is now on the chip i doubt that it is. Will have to see what the reviews say, and wait for someone to stuff on 2.2V+ through it and see what happens.

DIMM @ 2.2V + Bloomfield = Boomfield.

 

[Tau]

DIMM @ 2.2V + Bloomfield = Boomfield.

Yeah i know, i was trying to make a point

 

[EastCoasthandle]

It's nice to see that vendors will be pushed to comply with JEDEC standards at last. Most "high performance" DDR3 modules operate at voltages in excess of 1,8V - the maximum JEDEC allowed for DDR2 memory! I have rarely spotted DDR3 modules working at 1,5V, actually, and I certainly don't find this to be a good practice - if JEDEC state a maximum voltage for given memory technology, they surely have the reason to do so. But I guess it will be hard to justify water-cooled modlues working @1,5V, right .

Here, here! This is what we should be seeing out of DDR3 memory to begin with. If they want to up the speed then invest in doing so lower then 1.65V or don't introduce anything at all. DDR3 IMO has a been a complete mess with higher then normal voltage and higher then normal latency. Maybe with this requirement the memory market will finally get it together!

We should be seeing timing of 7-7-7 or lower and voltage of 1.5V DDR3 1600. Instead we are seeing 9-9-9 at 1.8V+ or 7-7-7 at 1.8V+. It shouldn't be that way if they want to convince people that DDR3 is next gen. That's why it's not doing well IMO.

 

[mdm-adph]

I find this odd -- how is it that AMD gets around having different voltages for their memory even though they're using the same kind of integrated memory controller? Is this something unique to DDR3 ram?

 

[Sasqui]

I find this odd -- how is it that AMD gets around having different voltages for their memory even though they're using the same kind of integrated memory controller? Is this something unique to DDR3 ram?

Same question here...

Wonder when we'll see DDR3 modules with 6-pin PCIE power connectors?

 

[EastCoasthandle]

The JEDEC of DDR2 is 1.8V. You can buy 4Gigs of ram at 2.0V which is only 0.2V difference. I can run my 4Gig kit just below 2.0V. These 4Gig kits usually come with PSC memory ICs.

Regardless of what voltage AMD can accept it shouldn't be that way. If PSC can bring 4gigs of DDR2 of memory close to JEDEC there is no reason why everyone else can't do the same with DDR3. This would bring the difference down from whopping (1.5V - 1.9V =) 0.4V to 0.1V using 4Gigs at 7-7-7!

It's funny how we went from 2Gigs of DDR2 at 2.2V (where 2.1V is more common) down to 4Gigs of ram below 2.0V! But can't see the same innovation with DDR3 which is suppose to be next gen!

 

[thebeephaha]

I see this as a major fail if this is actually true.

 

[InnocentCriminal]

I find this odd -- how is it that AMD gets around having different voltages for their memory even though they're using the same kind of integrated memory controller? Is this something unique to DDR3 ram?

Exactly what I'm thinking.

 

[wolf2009]

DIMM @ 2.2V + Bloomfield = Boomfield.

lol

 

[Fitseries3]

WTF is the point of this crap even comming out? wth... this is total garbage.

 

[btarunr]

WTF is the point of this crap even comming out? wth... this is total garbage.

What are you refering to?

 

[Fitseries3]

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.

 

[razaron]

the soon-to-be released a-data tri-channel ddr3 is at 1.65-1.75v. so upcoming Tri-Channel kits should have the right voltage.

 

[Poisonsnak]

I find this odd -- how is it that AMD gets around having different voltages for their memory even though they're using the same kind of integrated memory controller? Is this something unique to DDR3 ram?

Back when socket 939 was new and the DFI Lanparty nF4 Ultra-D was the best AMD motherboard around these same kinds of problems existed. The board had a jumper so you could pull RAM power from the +5V rail (max DIMM voltage 4V) as opposed to +3.3V rail (max DIMM voltage 3.1V). Guys who were using the +5V jumper were frying CPUs if you had too big of a difference between Vcore and Vdimm. Keep in mind these were the days of Winbond BH-5 DDR1 chips where they would take as much voltage as you could give them.

 

[Fitseries3]

the soon-to-be released a-data tri-channel ddr3 is at 1.65-1.75v. so upcoming Tri-Channel kits should have the right voltage.

yeah but that kit needs 1.65v to boot and nehalem is 1.65v MAX. that means that the ram will have to run as low as 1.0v

 

[MrMilli]

You guys shouldn't worry too much:
http://www.techpowerup.com/72319/Elpida_Buffalo_Release_Prototype_DDR3_2400_MHz_Memory.html

The module was tested stable at 2096 MHz, at amazingly low 1.5V, with much tighter timings of 9-10-9-24.