Do memory modules really need heatsinks

Soldato
Joined
20 Nov 2005
Posts
4,735
Location
Redditch
As title really, just received my new corsair vengeance memory order from overclockers & looking at the modules I'm not sure they will fit behind my full height radiator.

At work at the moment so not tried them yet but just looking at them it may be tight

So, do memory modules really need the heatsinks?
 
Man of Honour
Joined
13 Oct 2006
Posts
91,164
It used to be only the most cutting edge memory had heatsinks (and in some cases actually needed) now it seems memory without is the exception rather than the rule.

I remember someone had the heatsinks from their OCZ Reaper X or something start slipping off and ran them fine without the heatsinks heh.
 
Soldato
Joined
4 Nov 2006
Posts
2,944
Location
London
No, but that diversified the market and allows them to sell all sorts of them with different colours. I just go for the lowest height memory to make sure it's compatible with future CPU coolers
 
Associate
Joined
11 Jan 2008
Posts
468
Location
Liberty City
My Samsung Green takes a beating with tight timings and overvolting for over 4 years. Still cool, still stable.

Yeah my Samsung at 1.35v is completely cool.

I remember my Geil DDR2 had orange coloured heatsinks and they did actually get pretty hot. Having said that I think DDR2 ran at 2.2volts at 800MHz.
 
Soldato
OP
Joined
20 Nov 2005
Posts
4,735
Location
Redditch
Yeah my Samsung at 1.35v is completely cool.

I remember my Geil DDR2 had orange coloured heatsinks and they did actually get pretty hot. Having said that I think DDR2 ran at 2.2volts at 800MHz.

Ah, the good old days of putting 2.2volts though the memory & 2v though the cpu to get 2.5ghz.:D
 
Man of Honour
Joined
13 Oct 2006
Posts
91,164
Yeah my Samsung at 1.35v is completely cool.

I remember my Geil DDR2 had orange coloured heatsinks and they did actually get pretty hot. Having said that I think DDR2 ran at 2.2volts at 800MHz.
Typically with RAM the modules are rated to considerably higher temperatures in normal operation than they would ever actually get to in desktop PC operation - heatsinks in rare cases made a difference in stable overclocks but generally don't help - what people consider toasty for RAM is actually around 20-25C below their max (long term stable) operating temperature.

The real problem is voltage (and temperatures have only a slight impact here) there is generally a point where the effects of electromigration speeds up considerably - if anyone remembers the D9GMH chips there was a voltage point (which slightly differed on each) where they would die within weeks at or above that level regardless of the cooling used. This is why SRAM is often the first product pushed through to test new semiconductor nodes as the density and complexity is good for exposing effects like this early on.
 
Man of Honour
Joined
13 Oct 2006
Posts
91,164
I used to have the OCZ XTC cooler like that along with the RAM and DDR booster - with that thing you actually needed cooling - think I still have the setup in a box somewhere.

I miss the days when manufacturers did totally nuts things like that.
 
Soldato
Joined
5 Sep 2011
Posts
12,816
Location
Surrey
It's not so much the memory these days, or the DRAM voltage, or effect this has on the IC lol - more so the means (especially with OC socket rails) the CPU need to get there. A lot of mystery shrouded there, but CLT and CLK can scale quite high when pushing frequency. Generally speaking DDR4 IC's are pretty tough, I've not heard of any real solid evidence of degradation or failures.
 
Back
Top Bottom