Universal GPU waterblocks.. hmm

[gbkinum1]

I recently added a MCW82 to my loop.. but I'm a little nervous about my VRMs and vRAM. I've never even seen my GPU hit 40C.. which is a huge improvement over the stock cooler where it was regularly at ~65C. The only issue I've found is that often times while I'm playing BFBC2, the game gets choppy for a quick second.. which could really be just about anything. I don't really remember that happening before I added my video card to my loop.. but honestly I didn't really pay attention to it either. I attached the vram sinks with the thermal paste that came with the kit (used only a small amount.. figured I'd use the same strategy as I do for applying thermal paste to a CPU), and the air temp around my ram sinks is usually in the 22-27C range.

It could really be a million things that could cause this (including just using an out dated video card.. 9800gtx+).. but I was curious if anyone else has ever had any trouble with the vram/vrms causing stability issues with a universal gpu block or if my time would be better spent looking into other possible causes. Thanks in advance,

gbkinum1

 

[guavasauce]

your sudden skipping sound more like texture fill or frame buffer issues. but really, there shouldnt be any. did you update your drivers during this addition? change any other settings?

i would like to think if the sinks werent working correctly then the skipping would last till most likely a bsod.

 

[rubix_1011]

I doubt the waterblock or the RAMsinks are the cause. I have run MCW60's and RAMsinks for years (including on a pair of 9800 GTX+ cards) and have never had issues with them. Stuttering like that in a game can be just about anything (like you mentioned) and might have more to do with internet latency than anything else.

 

[gbkinum1]

Makes sense to me, I can't really think of anything else I've changed.. but it could have always been there and I was just paying more attention for whatever reason. Or I could have had something running in the background that slowed it down for a sec.. dunno. I kinda doubt it was an internet problem.. usually lag issues have rubber-banding in game or people walking in place etc.

I actually just thought of something minor that has changed (I dunno if it would cause this..), I have EVGA precision running in the background to keep an eye on GPU load/temps etc.. perhaps there is a little more overhead associated with it than I expected. It's kinda difficult to test this because it was an intermittent problem (happened like 2-3 times in the couple hours I was playing).. although I guess that is often enough that I could try playing the game w/ and w/o evga precision to see if that fixes the problem.. it'd just take a while.

Thanks for the responses =]

 

[rubix_1011]

I've watercooled for 9+ years and haven't had any issues related to universal GPU blocks or the setups. If you had issues, you'd know it when you put load on your system...it would likely shut down, not just stutter intermittently. If your video card were having any problems, it's very noticeable and would continue getting worse and worse until you experienced crashing or stopped playing.

The only thing I can say would be some service/application running on your machine that causes the hitching (think virus scan kicking on, etc) , or you experience network lag/connection issues briefly either due to server you are playing on or issues local to your ISP and service area.

 

[wolfram23]

You have no idea how tempted I am to buy MCW82s for my 5850s ><

I saw these awesome full cover passive heat sinks for VRM/VRAM too,

http://www.crazypc.com/products/hd5850-hs-936605.html

$30 for heat sinks, $55 for block... $85 total. Or, $120 for full GPU block. Saves a fair bit with 2 cards, although I already have passive RAM sinks.

 

[rubix_1011]

Yeah, I do like Swiftech's approach on the single, large heatsink to accompany their universal blocks with a fairly cheap price tag. I only wish there were more options available for card compatibility.

 

[wolfram23]

Yeah, true! But that's always the case, even with full cover blocks. At least there's still the option of universal VRAM/VRM heat sinks. But of course having a full heat sink like this is a hundred times better.

 

[gbkinum1]

Yeah, true! But that's always the case, even with full cover blocks. At least there's still the option of universal VRAM/VRM heat sinks. But of course having a full heat sink like this is a hundred times better.

Why is it 100 times better? If ram sinks keep a card stable.. what other benefit would you get from a full heatsink?

 

[boiler1990]

It has its own clean looks like a full cover block. It's also much easier to apply to a reference design card since you don't have to struggle with attaching it.

It took me a good 20-30 mins to attach my RAMsinks to my 560Ti, and I knew what I was doing. A unisink like those probably takes half that.

 

[rubix_1011]

That Swiftech full cover heatsink probably takes like 3 minutes to install; that's only because it likely uses 12 of those weird, spring-loaded screws that nVidia likes to use on it's stock coolers.

 

[wolfram23]

It also has a massively larger surface area, so a lot more room to dissipate heat. The VRM and VRAM individual heat sinks aren't really that good, they are only the size of the components they are trying to cool. If I over volt to 1.2V or so, my VRM and VRAM temps get way too high during a burn test. At stock, it's all gravy, but watercooling to me is to enable higher overclocks and simultaneously lower noise levels.

Installation and aesthetics I consider a secondary, but convenient, bonus.

 

[rubix_1011]

I'd prefer a full-cover heatsink plate like this, but they don't make one for the GTX 260 Core 216's 55nm design (orig. was 65nm).

Bummer for me. It's about time to upgrade my GPUs, anyways.