Test System Details

This is the ram in my laptop, it is the best DDR3 non XMP ram on the market

http://www.valueram.com/datasheets [...] 3K2_4G.pdf

Right off that spec sheet you see that the ram only consumes 1.8W maximum as is.....saving an extra watt is a joke (bear in mind under load my laptop is pulling 250W and it idles around 70W - I have an Alienware M17x)

Hah, I knew that this whole "green memory" thing was just an eco-fad. Memory sticks usually consume as much as generic case fans and usually lower. Why pay a premium for the same performance but a ~1 watt - 4 watt difference?

Low voltage memories like Kingston LoVo or G.Skill Eco doesn't necessarily mean reducing energy consumption. Low voltage can reduce heat produced, and also provide high oc potential under 1.65v. I got my G.Skill Eco 1.35v DDR3-1600 7-8-7 overclocked to 1.65v DDR3-2220 10-11-10.

^ That must've been the real use for those low voltage RAMs. A greater headroom for overclocking. But in stock, the only thing IMO that you could benefit from it was lower heat produced (which is negligible if you have a mem cooler and a good case airflow).

th

well u are right u can overclock it more at a lower voltage and u are kinda of wrong when u said it doesn't necessarily reduce energy, u should do some research of the loVos and compare them the regular ram and you will find that the ram do has lower energy consumption

Well, the increment in price isn't worth the saving if we compare it to increment in a power saving motherboard, or GPU. This should be something like coup de grace in designing the ultimate low power system, but definitely not the first step.

Hey, that 4W under load is 2W per module.
If you use a server with 18 ram slots filled, then that is 32W in all.
In the 32 ram slot Tyan s8812 it would mean 64W in power consumption.

Taking that these systems run like 0-24 7 days a week, it DOES count....

In this specific instance, then yes it does count. However, consider that most readers on this site are users and not server operators. Also consider the added cost of populating all those RAM slots with more expensive RAM. These modules cost ~$170 for 4GB of DDR3-1333. You can pick up some G.Skill off Newegg for $81. Multiply the difference by 8 to populate all slots in the Tyan s8812 and you're talking about an extra $712.

It never even crossed my mind to use low-voltage RAM as a mean to save power. To me the whole point of low voltage is either better overclocking headroom or lower heat output (and even that I'm not sure).

this doesnt make sense to me, considering a lot of the eco sticks are the exact same price as corresponding non-eco sticks, yet usually oc better:

http://www.newegg.com/Product/Prod [...] 6820231321

100 bucks for that, i think ill take it.

read some of the reviews on that post and to be honest...who the fuck buys eco friendly memory and expects it to be an overclocking champ and complaints about the heat sinks quality. seriously?

So for RAM it saves nothing

Personally, apart from fine-tuning my processor's CPU clock driver, the most power-sensitive steps I took were:
- using a large Heat sink with a slow fan
- using two 2.5" HDD in RAID 1 (the actual concern was on data safety, but power was a concern too)
- modifying my Radeon 4850's BIOS to use much, much lower voltage and clock speed when idle: 0.93 V and 160/500 MHz instead of 1.1V and 500/900 MHz. Probably cut the power requirements by 30W.

So there is a (small) point to LoVo ram instead of the silly Watt-usage comparisons done when SSDs are benched..

Test using a core i7 rig. Bit-tech.net did a similar comparison of these same modules 4 months ago. The reduction in voltage from "standard" 1.65v ddr3-1333 memory to these modules at 1.2v ddr3-1333 reduced system load wattage from 193w to 166.

THAT is a considerable gain, and I am disgusted that Toms chose to bench a single test configuration here - the IMC's in AMD and Intel chips are markedly different, and the only reason DDR3 is being made at this voltage is because Intel designed it's nehalem chips around the 1.5v DDR3 specifications standard, forcing manufacturers to start binning for voltage and not just speed as they had been doing previously.

For anyone interested, here's the feature article from bit-tech: http://www.bit-tech.net/hardware/m [...] ddr3-ram/1

any meaningful power saving can only come from the components that consume it in the first place : and that would be your graphics and cpu . from two sources : chip level power management (in tandem with the OS) and the fabrication (NM) of the chip itself .

i will hopefully replace my 5850 with something 28nm . someone (long back) on tom's said that graphics are going to be saturated , well , no . forget eyefinity , even with single monitors , the newer games will require a lot of juice , considering the holy grail - photo realism is still far .

I can imagine this is much more exaggerated with DDr2 2.1V running at 1.4V and 667mhz instead of stock 1066mhz .

The gigabyte mboard on demand tool helps a lot with power saving as well.

I think when the yearly bill gets cut down by $50.00 or more then people will start taking this serious or spend the money on burgers / beer instead.

Servers centre's with this low power ram or corporate company's 100 or more pc's will see a lower bill using low power parts and under clocking.

Honestly memory modules weren't the first thing that sprung to my mind when it comes to low-power components. But on the other hand, since it runs at frequencies with lower voltage than normal modules, I'm sure this will be a great potential for overclocking.

For consumers it seems to me this kind of low voltage memory should have been provided as SO-DIMMs for notebook use as the first market. Even a few watts could have a reasonable impact on battery life in such a platform.

For a desktop it seems pretty negligible though.