Full Cover blocks

[Eddy_EK]

Originally Posted by Bevel

As gpu's become more powerful they would really benefit from better cooling. I think this idea has merit.

Yes, the GPUs are wery much gaining on CPUs. I think that the engeneers that develop graphic cards will soon be forced to stream on lowering the power consumptioin.

[Bevel]

Interesting times ahead indeed.

[ProHandyman]

I agree that GPU temps and power consumption are hitting the same walls that CPUs have now, with less then adequate factory cooling.

I agree with some special machining, and possibly another pump, rad in the circuit... things would improve... but over-all cost will be the limiting factor.

I think water cooling has just about reached a pinnacle, that unfortunately may fall off rapidly. Advancements in chip technology may very well substantially lower the generated temps and power requirements in less then a decade.

I'm still going "wet" though!

[Lucanus]

them thar blocks sure do look mighty purrdy.

[Long Haired Git]

Originally Posted by Bevel

Now pro your thoughts please. I have had this idea for awhile now, you see the thing is if the Storm block works so well on a bare core why not produce a gpu block using the same principles ie jets? The concept is proven to cool better than pins or channels. Worth investigating?Hmmm

Storm design is too "tall" for the video card market. Need something flatter, low profile. Can go to G9 level in order to reduce stand off, but machine costs (aerospace level) put $$$ to far out.

Cathar did make a low profile prototype block: "Hydra" and this, I believe, was what was cooling his dual-loop peltier water chiller circuit.

GPUs have bigger contact patches than AMD XP CPUs. Storm design coped well with this density issue. Now we have IHS or large dies as per the GPUs, do we need Storm "density"? Some say Apogee and Storm perform the same on IHS type CPUs... MCW55 is Apogee for GPU?

Things ARE moving in this arena, but work commitments keep the progress slow.

MCP655 @ 5 with MCR220QP and Storm and 1.5m of tubing = 6 LPM
MCP655 @ 5 with MCR220QP and 2 Storms and 2.0m of tubing = 4.5 LPM
MCP655 @ 5 with MCR220QP and 3 Storms and 2.2m of tubing (SLI) = 3.8 LPM

Certainly think such flow rates are acceptable.

Assuming 110W CPU (over-volted, over-clocked) and dual 70W GPUs (ditto mods) then the impact of the two extra storms is 6 degrees C to the CPU. However, for that sort of heat load, I'd be going for dual MCR220's or a PA120.3.

[ProHandyman]

Thanks for the input, LHG! The effect on flow wasn't as severe as I thought, as your approximator shows. However, there is still the incresed heatload, as you have mentioned

Yes, I believe Stew had something along these lines... at least unconfirmed rumours had it. Unfortunately, I don't think anytime in the near future we will see anything come out of Little River. Cathar is all but MIA, with a rare post here and there about the web.

All is interesting... further developments will be interesting to watch. Time for that double cup o' joe!

[Bevel]

Originally Posted by ProHandyman

Unfortunately, I don't think anytime in the near future we will see anything come out of Little River. Cathar is all but MIA, with a rare post here and there about the web.

Well maybe if we ask nicely he may grace us with his presence and share his thoughts with us on the future of GPU cooling, I for one would be very interested to hear his views. So how about it Cathar?
Oh and hows the R1 going?

[unreal]

Where's the review for this?!

[Eddy_EK]

No reviews yet, but hoplefully will be soon.

[Eddy_EK]

My friend has done some testing and he got some good clocks. But had prolems with crossfire and got poor results, perhaps because of CPU. His post is here

[unreal]

Which temps are which in screenshot?

[Hansfragger]

Originally Posted by ProHandyman

Ed, being the one in your corner, I believe that getting to a fine machine finish that is just short of "mirror" is more then adequate! (I myself would strive for the micron surface ripple that a well machined racing cylinder head surface would have). This would have to be with limited to no obvious visual "fly cut" markings.

Full polished-mirror finish is an extreme operative and esoteric visual for reviewers and "those easily impressed by bling". If the occasional client wants the "ultimate finish", make it available at an added cost, due to extra "time and effort" used. I being an entrupeneur myself, totally understand the level of desired finish to what's fiscally reasonable. This level of manufacture could be offered as a limited or special edition, possibly offered with other "bling", such as your plating and LED lighting effects, or even a "signiture edition".

Offering higher finished models at an increased price sounds like a great idea. Careful - you might start a new trend.
Eddy - You do very nice work.

[ShoNuff]

Eddy you have a PM.

Very nice blocks BTW. I like the fact that I don't have to "mix metals" and have the option of cooling the vreg's without having to use an aluminum block.

I’m sold!