FEATURED Scared of running memory over 1.5v on Ivy Bridge?

[moocow]

Since so many people have been scared of running memory over 1.5v on ivy Bridge I thought I would grab a set of memory to show that things won't explode when running higher voltage older memory.

When I found this set of very early DDR3 memory I had to have them.
Rated at 9-9-9@ 2000mhz with 2.0v they would be a perfect test to see how Ivy bridge does with older memory.




The IC is samsung HCF0 and is from sometime in 2008. Needless to say, they are pretty old.

Trying 1.65v these sticks were able to do the rated speeds. Not bad at all for a .35v undervolt!




Next up I pushed the voltage to 1.8v. I was able to do 2133mhz superpi32m stable. They would validate at 2200mhz+ with the same voltage but superpi would crash after a few minutes. Not bad at all for still being undervolted by .2v



Last up I bumped the voltage to 1.9v to see what they would do and they were able to hit 8-8-8@2000mhz. Speeds like this are usually found on sticks with Hyper IC




Unfortunately I was unable to go farther due to a limit with the Maximus V Gene only giving up to 1.92v
Not bad for a set of ram from 2008

 

[keny]

Nice write up cow I have a set of projectx that are rated 7-7-7@1800 but I havnt even tried them yet as I found out after buying them that they were 2.0v sticks and that part at the back of my small brain was twitching at the thought of 2.0v ddr3 on a modern board, but I think I will try them out now ...
Thanks

 

[moocow]

They should be fine.

I've got a set of team Xtreem 8-8-8@1800 1.9v sticks that will get benched, as well as a set of 9-9-9@1800 1.9v ballistix. If I have some time tomorrow I'll slap them on the bench to see what they can do.

 

[Woomack]

Try with DRAM fixed frequency up to 500 and then you can make up to 1.96V on M5G.
2133 is quite nice for older Samsungs and ~1.8V.
I have 3x1GB kit of OCZ Platinum 1600 7-7-7-20 1.65V that were making 1866 7-6-5-18 2.0V but I have never checked them on Ivy Bridge.

 

[hokiealumnus]

Looks like a good ol' kit kow, nice write-up. I run 1.65V 24/7 on Ivy.

Slight disclaimer, if you please - Lest someone come upon the thread and think 1.8V+ is ok 24/7, I wouldn't recommend voltages like these for 24/7 use.

Heh, if you feel like running that experiment, you're more than welcome. I bet the IMC degrades when used long term at 1.92V though.

 

[wagex]

nah them radioactive heatsinks make all the voltage he wants fo free so no stressed out IMC ;D

 

[Woomack]

If cpu socket has some issues ( slightly bent pins etc ) then while running on too high voltage you can burn pins or cpu pads.
I had burned one socket pin but all was working fine after I cleaned it with some alcohol ( I'm not sure if it was too high memory issue or something else ).
I'm not recommending so high voltage for 24/7 but for tests, why not
I was running 1.92V on probably all my sticks and so far both memory and IMC are fine.

 

[wagex]

i think i run my ram at 1.65v because i think that is the lowest option on my board for some weird reason i was never able to go lower... i dont remember because i dont mess with that stuff lol. perhaps that was my 1156 board that was like that.

 

[hokiealumnus]

Oh absolutely, for testing/benchmarking go for it, the sky's the limit. I'm just saying don't run it all day every day.

 

[wagex]

Oh absolutely, for testing/benchmarking go for it, the sky's the limit. I'm just saying don't run it all day every day.

i dont have a choice lol

 

[hokiealumnus]

1.65V is fine, that's what I run mine at 24/7 and have yet to run into a problem. I was saying don't run in the 1.8-1.9V range 24/7.

 

[keny]

I wonder the IMC on the FX CPUs is like for voltage, I know the board goes into red at only 1.7v which I think is a bit low but I ramp up the voltage to 1.85 on almost all my DDR3

 

[moocow]

How about some D9GTR results?

These sticks are rated at 8-8-8@1600mhz 1.8v stock on XMP profile #1, the second profile is 7-7-7@1600mhz 1.9v. Lets see how they do on a modern platform.



Starting out I decided to go straight ahead and set the memory to 1.85v for all of the testing. Using more didn't help at all since the maximus V gene memory limit is 1.92v+- and using less didn't exactly help either.

Lets start out with 8-8-8@1866mhz 1.85v



How about 9-8-8-@2000mhz 1.85v?


Not too bad at all.

Lets see if they can hit the memory speeds of more modern memory.

9-9-9@2133mhz 1.85v



Lets see if the timings can tighten up a bit.

9-9-8@2133mhz 1.85v




And 9-8-8@2133mhz 1.85v



Decent speeds!
These sticks were unable to do 2200mhz or higher on ambient temperatures. With extreme cooling and more volts I'm sure that they should easily be able to do 2200mhz to 2400mhz with very tight timings.

 

[EarthDog]

Excellent write up there Cow!

Like Hokie alluded to, I think the concern was with the 24/7 operation with voltages over 1.65... Crank that bad boy to 1.9 and let her rip for a couple months and if its still kicking, BUNKED!

 

[moocow]

someone want to cough up a cheap sb or ib chip? i've got a spare mvg i could do that on

 

[rogbur22]

What I think is funny is how your not supposed to use anything over 1.5v memory wise on these 2nd and 3rd gen chips yet the motherboards list memory over 1.5 on there QVL's.

I have an i5-3570 and this memory which is 1.65v

http://www.newegg.com/Product/Product.aspx?Item=N82E16820104173

I have ans ASrock Z77 Extreme 4 as many people here do also and guess what memory is on the QVL ? Thats why I picked it initially. I just looked and it's still on there along with others over 1.5. Kinda screwy no ?

http://www.asrock.com/mb/overview.us.asp?cat=Memory&Model=Z77 Extreme4

I e-mailed ASrock after I realized what was going on and they wrote me back a 1 line reply that says "Roger, use 1.5V memory" that was it.

I didn't listen

 

[moocow]

Intel says alot of things.

Technically you're not even supposed to overclock or overvolt the chip and we all know how accurate that is.