why are manufacturers pushing Z77 1.65v DDR3?

[crawlgsx]

Everything I read says not to use over 1.5v with IB (or even SB for that matter)

Yet so many memory manufacturers are already listing their ramas "ready for Z77" or speccced for it, yet they run at 1.65v

such as Gskill's new ram

http://www.newegg.com/Product/Product.aspx?Item=N82E16820231588

They have a bunch like that, and they are not the only ones.

Designed for 3rd Generation Intel Core Processors and Z77 platform Intel XMP 1.3 ready Removable top fin of the heat spreader XMP Ready for 3rd Generation Intel Core Processors

 

[drescherjm]

Because these are crazy high frequency ram that is overclocked to achieve this frequency. No RAM manufacturer makes DDR3 2400 chips. Some people will pay 3x the price for ram that will give them on average a 1 to 3% performance boost.

 

[Valshistixol]

You can run 1.65 on SB, just not over that.

 

[crawlgsx]

Because these are crazy high frequency ram that is overclocked to achieve this frequency. No RAM manufacturer makes DDR3 2400 chips. Some people will pay 3x the price for ram that will give them on average a 1 to 3% performance boost.

But that still doesn't explain the fact that they are advertising it for Z77 @ 1.65v when intel specifically states not to run over 1.5v

 

[crawlgsx]

You can run 1.65 on SB, just not over that.

I keep reading in places that it's somehow bad/not recommended, but I have been running @ 1.65v for just over a year with no issues.

I wonder if the Z77 is the same way (when paired with a SB cpu)?

 

[E4g1e]

But that still doesn't explain the fact that they are advertising it for Z77 @ 1.65v when intel specifically states not to run over 1.5v

Actually, Intel recommends against DIMM voltages above 1.575V for extended periods.

 

[firas]

+1 jpmeaney

With the 2nd generation Intel® Core™ processors only support DDR 3 at 1066/1333 and now on some 1600; that isnt what I am hearing as causing the issue in almost every case is that they are running the memory at 1.65v +. The 2nd generation Intel Core processor only support memory at 1.5v ±5% (1.425v to 1.575v) anything over this has a chance of damaging the processor memory controller.

Personally the performance increase of anything over 1600 doesnt seems to be worth the extra cost.

not sure who’s right, Intel guy always said 1.575v max but all DDR3 2400 models are running at 1.65V

but really, who cares? 0.5% real world benefit
http://www.anandtech.com/show/4503/sandy-bridge-memory-scaling-choosing-the-best-ddr3/6
http://www.xbitlabs.com/articles/memory/display/sandy-bridge-ddr3_7.html#sect0
http://www.bit-tech.net/hardware/memory/2011/01/11/the-best-memory-for-sandy-bridge/10

 

[jamsomito]

not sure who’s right, Intel guy always said 1.575v max but all DDR3 2400 models are running at 1.65V
*snip*

Intel guy also said SB only supports 1066/1333 which is a far cry from 2400.

 

[drescherjm]

Intel guy also said SB only supports 1066/1333 which is a far cry from 2400.

DDR3 1333 is the maximum official Intel supported rate of the memory controller on all SB processors although the chip does allow you to overclock the memory controller and raise the IMC voltage.

 

[LigTasm]

Basically, the problem is people keep forgetting that DDR3 at 2400mhz is a serious overclock, over 1000mhz over intel spec. It needs 1.65v to run at the advertised speed and timing, nothing says its going to be safe to do so though.

 

[jamsomito]

Basically, the problem is people keep forgetting that DDR3 at 2400mhz is a serious overclock, over 1000mhz over intel spec. It needs 1.65v to run at the advertised speed and timing, nothing says its going to be safe to do so though.

Right, but to come full circle with the thread, manufacturers are producing RAM with these voltages and clocks and saying they're "compatible" or even "designed for" the SB processors, when Intel clearly says they're not in the specs.

 

[Kardonxt]

I'm sure the 2400 defaults to 1.5v 1333 just like almost any ram over 1333. So the ram is 100% compatible at default settings. And is guaranteed to be able to overclock to speeds that aren't supported with voltages that aren't supported.

 

[Acer_Sheep]

Guys, it's on intel's site you can use 1.65V memory, I don't even know from where that 1.5V shit came from, I personally run mine RAMs at 1.65 without any issues and probably the thousands of other users as well.

I think the IntelEnthusiast is not well informed about technical specification about CPUs they are making.

They wouldn't list them on their site as supported if they would not really be.
http://www.intel.com/content/www/us...asheet.html?wapkw=intel+core+i5+system+memory

 

[LigTasm]

Right, but to come full circle with the thread, manufacturers are producing RAM with these voltages and clocks and saying they're "compatible" or even "designed for" the SB processors, when Intel clearly says they're not in the specs.

I'm sure the 2400 defaults to 1.5v 1333 just like almost any ram over 1333. So the ram is 100% compatible at default settings. And is guaranteed to be able to overclock to speeds that aren't supported with voltages that aren't supported.

Exactly, the RAM will run fine at intel spec, it will ALSO run fine at an XMP profile of 2400mhz. All they are saying is that it has the potential to do so, nothing about twisting your arm to run it that way.

 

[jamsomito]

This thread is a good read.

It points to this official spec sheet from Intel. I'm no electrical/computer engineer so I can't determine if the conclusions from the aforementioned thread are valid, but Table 7.1 (p. 53) may yield some useful info. Otherwise, I'm sure it's in there elsewhere but I don't have time to go through it right now.

I understand that anything greater than 1333 memory is going to be overclocking the memory controller on your CPU. I also understand that faster RAM will default to 1333, but be guaranteed to work at the rated speed / voltage. However, advertising as the faster ram and saying that the default works with your CPU seems a bit misleading, don't you think? I'd be willing to bet there's no documentation with your shiny new RAM that says it's rated faster than is approved by your CPU manufacturer, it's just claiming to be compatible. It seems like the effects of running at higher voltages are more or less unknown right now (some people say there's degradation effects over time). Think about if they were worse though. If the effects of running at the RAM's full capability (speed/voltage) actually were detrimental to your CPU, any numbskull like me would buy the RAM, herp derp it into it's OC'd setting and proceed to blissfully fry their CPU.

 

[thesecond]

Right, but to come full circle with the thread, manufacturers are producing RAM with these voltages and clocks and saying they're "compatible" or even "designed for" the SB processors, when Intel clearly says they're not in the specs.

they are compatible at stock clocks.

Overclocking is mildly compatible and definitely not in the specs. The user takes it out of spec by doing anything to the timings, clockspeeds or voltages.

This thread is a good read.

It points to this official spec sheet from Intel. I'm no electrical/computer engineer so I can't determine if the conclusions from the aforementioned thread are valid, but Table 7.1 (p. 53) may yield some useful info. Otherwise, I'm sure it's in there elsewhere but I don't have time to go through it right now.

I understand that anything greater than 1333 memory is going to be overclocking the memory controller on your CPU. I also understand that faster RAM will default to 1333, but be guaranteed to work at the rated speed / voltage. However, advertising as the faster ram and saying that the default works with your CPU seems a bit misleading, don't you think? I'd be willing to bet there's no documentation with your shiny new RAM that says it's rated faster than is approved by your CPU manufacturer, even if it's claiming to be compatible.

No. It is only misleading to somebody who misunderstands the labels on enthusiast ram. All the label tells you is the clockspeed that the manufacturer binned the dimms at - its sort of like a guaranteed overclock label. It does not tell you default values, all default values are mandated by JDEC standards.

It is compatible in its stock form. As i said before, the user takes it outside of intel specification by overclocking it, and the label on your ram tells you how far you can expect to OC the ram.

 

[jamsomito]

No. It is only misleading to somebody who misunderstands the labels on enthusiast ram. All the label tells you is the clockspeed that the manufacturer binned the dimms at - its sort of like a guaranteed overclock label. It does not tell you default values, all default values are mandated by JDEC standards.

It is compatible in its stock form. As i said before, the user takes it outside of intel specification by overclocking it, and the label on your ram tells you how far you can expect to OC the ram.

Guys, it's on intel's site you can use 1.65V memory, I don't even know from where that 1.5V shit came from, I personally run mine RAMs at 1.65 without any issues and probably the thousands of other users as well.
...
They wouldn't list them on their site as supported if they would not really be.
http://www.intel.com/content/www/us...asheet.html?wapkw=intel+core+i5+system+memory

So if I'm getting this right... any RAM on intel's list is guaranteed to work IF it's run at the RAM's default settings (which matches Intel specs): [email protected]. Acer_Sheep is my case in point. He's taking the RAM's rated (binned) specs of [email protected] and assuming that because Intel approved the default settings, which isn't clearly indicated, that he's fine. In actuality, he could be subject to the degradation issues or whatever else in the long run.

That is what most people would assume (not all - most of us [H]ere like doing our research). That is the issue I'm having.

 

[Acer_Sheep]

So if I'm getting this right... any RAM on intel's list is guaranteed to work IF it's run at the RAM's default settings (which matches Intel specs): [email protected]. Acer_Sheep is my case in point. He's taking the RAM's rated (binned) specs of [email protected] and assuming that because Intel approved the default settings, which isn't clearly indicated, that he's fine. In actuality, he could be subject to the degradation issues or whatever else in the long run.

That is what most people would assume (not all - most of us [H]ere like doing our research). That is the issue I'm having.

Well ok, so be it, but why than they are on the site, and why dozens of manufacturers are making 1.65V memory? I don't get it basically, why is this happening. Doing a fraud is only thing that comes to my mind.

 

[drescherjm]

Well ok, so be it, but why than they are on the site, and why dozens of manufacturers are making 1.65V memory? I don't get it basically, why is this happening. Doing a fraud is only thing that comes to my mind.

Because extreme overclockers do not care if their processor could / will die 3 years (they certainly will not keep it that long) or that they paid 3 times the price for the binned ram chips they just want the highest possible frequency. There is a market for this and the dimm manufacturers are selling to this market.

 

[jamsomito]

Well ok, so be it, but why than they are on the site, and why dozens of manufacturers are making 1.65V memory? I don't get it basically, why is this happening. Doing a fraud is only thing that comes to my mind.

That is exactly the OP's question.

Because extreme overclockers do not care if their processor could / will die 3 years (they certainly will not keep it that long) or that they paid 3 times the price for the binned ram chips they just want the highest possible frequency. There is a market for this and the dimm manufacturers are selling to this market.

True, but the enthusiast / extreme overclockers are not the ones who would be having this problem though. It's your everyday Joe Schmo who wants the bigger, better, faster, or best bang for the buck. The advertising, in my mind, is misleading.

Guys, thesecond really hit the nail on the head. I didn't mean to hijack this thread, but it's really interesting to think about. Thanks for allowing me to share my thoughts - handing it back over now.

 

[Acer_Sheep]

Because extreme overclockers do not care if their processor could / will die 3 years (they certainly will not keep it that long) or that they paid 3 times the price for the binned ram chips they just want the highest possible frequency. There is a market for this and the dimm manufacturers are selling to this market.

This is not very correct answer, because extreme overclockers use extreme hardware for that, my hyperX blu 1600 is not for extreme overclocking things at all, although it runs at 1.65

 

[extide]

Guys, thesecond really hit the nail on the head. I didn't mean to hijack this thread, but it's really interesting to think about. Thanks for allowing me to share my thoughts - handing it back over now.

BS, there is a 1.65V DDR3 1333 kit on that sheet. 1333 IS within Intel spec AND would be the default SPD programmed valued.

There are plenty of other sheets like that for 2011 cpu's, etc also.

 

[thesecond]

So if I'm getting this right... any RAM on intel's list is guaranteed to work IF it's run at the RAM's default settings (which matches Intel specs): [email protected]. Acer_Sheep is my case in point. He's taking the RAM's rated (binned) specs of [email protected] and assuming that because Intel approved the default settings, which isn't clearly indicated, that he's fine. In actuality, he could be subject to the degradation issues or whatever else in the long run.

That is what most people would assume (not all - most of us [H]ere like doing our research). That is the issue I'm having.

Precisely!

Basically, the way to figure out if its "safe" or not is by the XMP profile. XMP profiles are apparently all Intel approved, so they should be keeping you in safe ranges.

However, this requires the user to figure out if their ram is meant for lynnfield/bloomfield or if its meant for sandy//sandy-e/ivy, because that difference will determine the limitation on the XMP profiles.