RAM Cooling : Overrated?

Message boards : Number crunching : RAM Cooling : Overrated?
Message board moderation

To post messages, you must log in.

AuthorMessage
Luke
Volunteer developer
Avatar

Send message
Joined: 31 Dec 06
Posts: 2546
Credit: 817,560
RAC: 0
New Zealand
Message 949055 - Posted: 22 Nov 2009, 7:59:58 UTC
Last modified: 22 Nov 2009, 8:01:04 UTC

Okay, some will know I have an upcoming system build scheduled for December.

And I do plan on OCing it to the max as far as the CPU is concerned.

And I have heard of many people out there using RAM heat spreaders/coolers to supposedly "cool down the memory". Do they actually make that much of a difference in regards to temperature? Can they give you a more stable OC? Or is it all just hot air? (pun not intended)

I'd think that standard memory modules (< 1600Mhz) by themselves wouldn't need it as they wouldn't get hot enough anyway and even if they did, the ambient case temperatures would keep it within safe limits. Perhaps modules with outrageous speeds (1800 Mhz, 2000Mhz) might need some form of cooling?

So I googled "Do you need RAM cooling?" (and other similar topics) and I skimmed through about 3 pages and so far I've come across quite conflicting opinions. Some trash the idea and others embrace it.
So SETI@home. Do you use passive/active RAM cooling installed? Do you notice a difference in the temps, if OC'd or running at stock? Or perhaps it's all for looks in side windows cases.

Because if I can get a clear consensus from all you folk, I can make a decision.

I would love to get all your views on this. Thanks all.

- Luke.
- Luke.
ID: 949055 · Report as offensive
Profile gizbar
Avatar

Send message
Joined: 7 Jan 01
Posts: 586
Credit: 21,087,774
RAC: 0
United Kingdom
Message 949063 - Posted: 22 Nov 2009, 10:19:17 UTC
Last modified: 22 Nov 2009, 10:30:59 UTC

Ok, my 2 penneth...

Most high quality memory come with heat spreaders already attached, it's only the value versions that don't anymore. I have found that if you have sufficient airflow in the case, then you don't normally need extra cooling for the ram.

In fact, I think (and I do stand to be corrected), that only one manufacturer made a ram cooler, and that was corsair for one of it's dominator series. It clipped over the top of the ram and forced air down onto it, via a couple of small fans. <edit> I believe that somebody on the boards here had one at one stage [msattler, was it you? ;-)] but again, I don't know how useful it was. Corsair obviously thought so, to have gone to the trouble to design it and supply it with some of their top-end ram. You could buy it separately as well, at least in the UK. </edit>

<edit2> You can still buy it in the UK in the form of a dual channel, and a triple channel version. You can also get the Antec Spotcool, which might help too, if needed. </edit2>

What I don't know is if it would fit anything alse apart from the corsair ram, but my guess is that it wouldn't.

regards, Gizbar.


A proud GPU User Server Donor!
ID: 949063 · Report as offensive
Profile Mamluk
Avatar

Send message
Joined: 10 Sep 09
Posts: 80
Credit: 2,448,048
RAC: 0
South Africa
Message 949075 - Posted: 22 Nov 2009, 14:52:57 UTC

My 2 pennies as well...

Your "need" to cool down your RAM will depend on what you are running and for how long. If you are running S@H apps, the memory usage is very limited, and it should not require cooling. On one of my (non-S@H) rigs, I run some OCZ Blades at 2000MHz, with prolonged and intensive usage. They have not needed any active cooling. Personally I think that RAM coolers are a waste of time for those who do not manufacture/market/sell them. Spend the money on better cooling for your OCed CPU...
ID: 949075 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 949078 - Posted: 22 Nov 2009, 15:02:50 UTC
Last modified: 22 Nov 2009, 15:07:35 UTC

On the i7 platform, RAM voltage is limited to about 1.65 volts.....

And at that voltage, RAM does not run that hot. I do have one of those Corsair coolers, and have it installed on the Frozen Nehi's bank of tri-channel.
But I think it is hardly needed.
My 920 rig has 2 channels of RAM installed, and the sticks are barely warm to the touch.

Now, on the other hand, some of my older rigs are running RAM at voltages upwards of 2.5v......and there you have some amount of heat being generated.

In fact, if I recall, the fan setup originally came with the RAM I was running on the Frozen Penny.........which was at the higher voltages.

So cooling may be warranted there.....although I have never used anything beyond the heat spreaders supplied with your better grades of RAM, excepting the Corsair setup.

And most of the rigs have a case fan blowing in the direction of the RAM to help dissipate what heat there is.

Soooooooo.......

I would skip it if you are building an i7. Dunno about those AMD folks......
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 949078 · Report as offensive
spitfire_mk_2
Avatar

Send message
Joined: 14 Apr 00
Posts: 563
Credit: 27,306,885
RAC: 0
United States
Message 949087 - Posted: 22 Nov 2009, 17:02:13 UTC
Last modified: 22 Nov 2009, 17:03:11 UTC

You overclock until the weakest component fails to boot. That weakest component might be RAM, might be network adapter, might be video card. The notion that you will overclock cpu to the max is erroneous one.
ID: 949087 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65750
Credit: 55,293,173
RAC: 49
United States
Message 949091 - Posted: 22 Nov 2009, 17:33:11 UTC - in response to Message 949078.  

On the i7 platform, RAM voltage is limited to about 1.65 volts.....

And at that voltage, RAM does not run that hot. I do have one of those Corsair coolers, and have it installed on the Frozen Nehi's bank of tri-channel.
But I think it is hardly needed.
My 920 rig has 2 channels of RAM installed, and the sticks are barely warm to the touch.

Now, on the other hand, some of my older rigs are running RAM at voltages upwards of 2.5v......and there you have some amount of heat being generated.

In fact, if I recall, the fan setup originally came with the RAM I was running on the Frozen Penny.........which was at the higher voltages.

So cooling may be warranted there.....although I have never used anything beyond the heat spreaders supplied with your better grades of RAM, excepting the Corsair setup.

And most of the rigs have a case fan blowing in the direction of the RAM to help dissipate what heat there is.

Soooooooo.......

I would skip it if you are building an i7. Dunno about those AMD folks......

I agree with what Mark has said here, From what I've read in reviews, i7(DDR3) ram @ 1.65v barely gets warm, Yet DDR2 like LGA775 uses can get heated, I use Patriot Viper ram(DDR2) PC2-8500 and It uses fins and the system works 24/7 without fail. But then that's cause and effect, which some in this world reject...
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 949091 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 6 Apr 07
Posts: 7105
Credit: 147,663,825
RAC: 5
Germany
Message 949093 - Posted: 22 Nov 2009, 17:36:08 UTC
Last modified: 22 Nov 2009, 17:54:25 UTC


I can talk only about DDR2 RAM.

I had contact to Corsair - I have their DDR2 800/5-5-5-18 @ 1.8 V / 4-4-4-12 @ 2.1 V.
They said, if running @ 2.1 V I need to cool with active RAM cooler.

Current I have OCZ 1066/5-5-5-15 @ 2.30 V in the Intel Core2 Extreme QX6700 mobo, which run at 944/5-5-5-15 @ 2.32 V.
And OCZ active RAM cooler (selfmade mod without original holder).

I have OCZ 1066/5-5-5-15 @ 1.8 V (low voltage RAM) in the AMD Phenom II X4 940 BE mobo.
And runnig this without active RAM cooling.


But, BTW.. because of OC a CPU.
I would OC only to the max. without changing the Vcore.
With this you would get a well performance/wattage.
If you increase the Vcore this ratio going bad.


For best performance/wattage, I would add a (or more) GTX2xx series GPU/s into the PC case.
If you have little time, I would wait to the new GTX3xx series GPUs. I guess - same wattage/double performance - than GTX2xx series GPUs.

ID: 949093 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 6 Apr 07
Posts: 7105
Credit: 147,663,825
RAC: 5
Germany
Message 949100 - Posted: 22 Nov 2009, 18:08:30 UTC
Last modified: 22 Nov 2009, 18:16:49 UTC


Maybe additional..

Every RAM have their own specs.

For example.. AFAIK.. DDR2 800 came with 1.8 V .
Then the manufacturers solt OC RAMs, DDR2 800/1066 with higher voltage (1.9 - 2.3 V).
If you bought them, you needed to cool down.

Then after two years or something, they solt DDR2 1066 with 1.8 V.
No active RAM cooling needed.

So everything over 1.8 V (@ DDR2) -> active RAM cooler.


I guess you can take this 'farmers regularly'* also for DDR3 RAM.

It's 1.5 V for DDR3?

[* "Bauernregel" in german]

ID: 949100 · Report as offensive
Dena Wiltsie
Volunteer tester

Send message
Joined: 19 Apr 01
Posts: 1628
Credit: 24,230,968
RAC: 26
United States
Message 949101 - Posted: 22 Nov 2009, 18:12:25 UTC

I am unable to tell you about RAM cooling because I have never had the problem, but I have a related story. We designed a new board with a processor we never worked with before. The engineer was warning us that we might be facing major melt down without the correct heat sink and special air flow.
After we had the board functional, I put my finger on the processor and it was just a little warm to the touch. The verdict was we didn't need to do anything special and we have never lost one of those processors. I would know because they are soldered in and the one time we wanted to remove one, we ended up scrapping the board because we couldn't remove it.
ID: 949101 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 949120 - Posted: 22 Nov 2009, 21:07:58 UTC

In my opinion, this can't be answered with a simple yes or no.

Increasing switching rates in the chip (higher frequency operation) means more heat. Lower voltages means less heat.

If there is enough airflow in the case, and the environment is cool, the chips may stay safely below the manufacturer's maximum temperature. If they're close to the CPU and the CPU fan blows hot air on them, maybe not.

If the manufacturer is conservative, then you can operate outside the recommendations and still be safe.

If not, you may need to cool the RAM.

Just too many factors for a simple yes or no.
ID: 949120 · Report as offensive
Profile hiamps
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 4292
Credit: 72,971,319
RAC: 0
United States
Message 949142 - Posted: 22 Nov 2009, 23:09:27 UTC

Reading other posts I checked it out and with a mild overclock on my I7 machine the memory isn't even warm. My watercooler blows that way but I was shocked and had a couple people feel them. To the touch they heel same temp as any metal in my case. They do look like they have sheilds of some sort.
Official Abuser of Boinc Buttons...
And no good credit hound!
ID: 949142 · Report as offensive
DJStarfox

Send message
Joined: 23 May 01
Posts: 1066
Credit: 1,226,053
RAC: 2
United States
Message 949264 - Posted: 23 Nov 2009, 16:05:47 UTC

I can always recommend passive cooling on all RAM chips. As far as active cooling? I cannot recommend it except in specific cases where airflow is poor and the OC/voltage is high.

In general, electronics don't like to change temperature, and the cooler the better in almost all cases. So, that's why I recommend some minor passive cooling for all RAM chips.
ID: 949264 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 6 Apr 07
Posts: 7105
Credit: 147,663,825
RAC: 5
Germany
Message 949286 - Posted: 23 Nov 2009, 17:53:58 UTC
Last modified: 23 Nov 2009, 18:03:02 UTC


AFAIK, only 'cheap' or old SD-RAM didn't had heatspreader.


passive:
heatspreader (with or without little holes) directly on the RAM chip.
This have nearly all RAMs.


active:
fan

This I meant with my post.

If I would go higher as 1.8 V @ DDR2 RAMs I would use an active RAM cooler.

ID: 949286 · Report as offensive
_heinz
Volunteer tester

Send message
Joined: 25 Feb 05
Posts: 744
Credit: 5,539,270
RAC: 0
France
Message 949346 - Posted: 23 Nov 2009, 23:43:51 UTC

A sample with RAM-cooling is my V8-Xeon
it is a must for the quad-channel FB-DIMM's
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Informationsliste Wert
Sensor Eigenschaften
Sensortyp Dual ADT7490 (SMBus 2Ch, 2Eh)
GPU Sensortyp Diode (ATI-Diode)
Motherboard Name Intel D5400XS

Temperaturen
CPU1 58 °C (136 °F)
CPU2 63 °C (145 °F)
1. CPU / 1. Kern 55 °C (131 °F)
1. CPU / 2. Kern 44 °C (111 °F)
1. CPU / 3. Kern 51 °C (124 °F)
1. CPU / 4. Kern 50 °C (122 °F)
2. CPU / 1. Kern 44 °C (111 °F)
2. CPU / 2. Kern 41 °C (106 °F)
2. CPU / 3. Kern 43 °C (109 °F)
2. CPU / 4. Kern 43 °C (109 °F)
DIMM 78 °C (172 °F)
GPU Diode 67 °C (153 °F)
Temperatur 1 51 °C (124 °F)
Temperatur 2 51 °C (124 °F)
Temperatur 3 54 °C (129 °F)
FB-DIMM1 84 °C (183 °F)
FB-DIMM2 87 °C (189 °F)
FB-DIMM3 84 °C (183 °F)
FB-DIMM4 78 °C (172 °F)

Seagate ST31000340NS 39 °C (102 °F)
Seagate ST31000340NS 28 °C (82 °F)
Seagate ST31000340NS 29 °C (84 °F)

Kühllüfter
CPU1 654 RPM
CPU2 649 RPM
North Bridge 4623 RPM
South Bridge 4182 RPM
DIMM 3020 RPM
Aux 837 RPM
Grafikprozessor (GPU) 90%

Spannungswerte
CPU1 Kern 1.10 V
CPU2 Kern 1.11 V
+1.5 V 1.54 V
+3.3 V 3.35 V
+5 V 5.13 V
+12 V 12.31 V
FSB VTT 1.21 V
North Bridge Kern 1.25 V
DIMM 1.82 V
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
I had have already temps of 100 grd celsius on the RAM's
depending on the tasks are running.
now are 11 tasks running under BOINC
8 x akv8
2 x astropulse 5.06 Hybrid
1 x collatz 2.06
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
as you can see cpu's are cool fan run still 654 rpm
but ram temp increase, with passive cooling they will be overheated in very short time
D5400XS V8-Xeon
ID: 949346 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 949440 - Posted: 24 Nov 2009, 8:37:45 UTC - in response to Message 949346.  

A sample with RAM-cooling is my V8-Xeon
it is a must for the quad-channel FB-DIMM's

Yes, I remember it being a concern when the FB-DIMMs first came out. My 8-core (Dell Precision 490) likewise has a special fan just for the memory.
ID: 949440 · Report as offensive
Profile bloodrain
Volunteer tester
Avatar

Send message
Joined: 8 Dec 08
Posts: 231
Credit: 28,112,547
RAC: 1
Antarctica
Message 949490 - Posted: 24 Nov 2009, 15:47:51 UTC - in response to Message 949440.  

overall no its not needed if you have good airflow in your case. now if you dont or have a lan box type one. then it might be needed.
ID: 949490 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65750
Credit: 55,293,173
RAC: 49
United States
Message 949513 - Posted: 24 Nov 2009, 17:28:18 UTC - in response to Message 949440.  

A sample with RAM-cooling is my V8-Xeon
it is a must for the quad-channel FB-DIMM's

Yes, I remember it being a concern when the FB-DIMMs first came out. My 8-core (Dell Precision 490) likewise has a special fan just for the memory.

Oh yeah, FB-Dimms make DDR2 look cool by comparison I think. :D
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 949513 · Report as offensive
Luke
Volunteer developer
Avatar

Send message
Joined: 31 Dec 06
Posts: 2546
Credit: 817,560
RAC: 0
New Zealand
Message 949591 - Posted: 25 Nov 2009, 3:26:13 UTC

Thank you all for your answers. It looks like the general consensus is that passive cooling is fine but active cooling is unneeded. Thanks S@H!

Now locking this thread...

- Luke.
- Luke.
ID: 949591 · Report as offensive

Message boards : Number crunching : RAM Cooling : Overrated?


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.