Hardware BFG Tech GeForce GTX 285 H2O

http://www.bit-tech.net/hardware/graphics/2009/03/09/bfg-geforce-gtx-285-h2o/1

BFG has paired Nvidia's GeForce GTX 285 with a Danger Den waterblock and overclocked it. Will this make for a tempting purchase or just price it out of its depth? Probably the latter, but at least you can also read about Danger Den's new DD-GTXX285 block - it's like two reviews in one, you lucky blighters.

 

Hmm, good timing or bad? I've just spent the past hour pricing up a new watercooling system, and then this review crops up. The prospect of someone else fitting the block and a guarantee is very tempting, I've killed a gfx card before tring to fit a block to it, not something I want to repeat...

 

I think to really test the cooling efficiency of the block you should have used a 120.2. It's not very common at all to see anyone cool a GPU anymore with just a 120.1. Look at the builds being done and you'll even see that ppl are splitting the loops between the GPU and the CPU/NB so that the heat of each line is contained to itself. Couple that with a 120.2 and you have a better test.
Plus, what fan were you using? If no fans at all, I'm not surprised at all by the results. . . Be better have a high quality fan with that rad, especially when if you're spending the money on the card/block you're not gonna go cheap and get some low performing fan to cool your rads with.

 

Eh?

It was super cool at both idle and load using just a 120.1 radiator - not sure I see the merit is unnecessarily using a 120.2.

I was fitted with a high airflow 120mm fan to my knowledge - certainly not a cheap low flow model.

 

Yeah, like what he said. The fan was a Ruilian Science RDM1225S, which is a high-airflow model.

 

Please correct card name spelling in all the standings from H20 to H2O. Thank You!

 

Done!

 

The mem clock on this nice and steady card is 666MHz
But why are there no Clocked cards around?
If u gonna make a Water cooler card, make it have at least 702MHz :S

 

What about temps after the final overclock?

Oh, and it would've been nice for us lazy readers if you just included the 4870X2 results in the overclocked graphs.

 

Yeah, but then we'd have to overclock the X2 to be fair to it, and then all the other comparison cards to be fair to them. And by then there'd be no coffee left in the whole building...

 

But this is exactly what I pointed out. How common is it to have a 120.1 cooling your GPU? Especially a 285? not very common at all. Take a look at all the build logs here or at xtreme (for example), and you're not going to find ppl using a 120.1 for their GPU. I just don't feel that's an accurate real world test. Most builds being done today are multiple rads (120.3+120.2) in big cases - maybe in someones SFF or lanbox do you see a 120.1, but not desktops. . .

And try updating your fans a little, I googled that fan Claave listed, and uh. . nice model - google returned with 13 results and it was all forum rants from 2004 or earlier. Is that company in business still? If the fan is 4+ years old there's a big part of your problem as well. Try mounting that fan with a San Ace or Scythe S-Flex F and then report the differences. . .

 

666 sounds cooler. Plus it's guaranteed to work on every card; a bad card might be flakey at 702, so it's up to the user to push it and test the stability of the OC.

 

im starting to find the "comments" page filled rants or "tounge in cheek" comments a little tiresome.

(quite ironic)

I did like the review though

 

Point taken and agreed... but I'm lazy.

 

Should we test every possible variation of cooling loop? That's a ridiculous idea, as the loads on any loop, the actual flow rates, the airflow and a hundred other factors will completely change the temps you'll see at home. A 120.1 with a high-flow fan provides a benchmark to see what your temps might be like.

 

I'm sorry that everyone is quick to judge. All I suggested was changing the rad size and the fan to something that everyone uses today. I never mentioned the other hundred factors that you want to talk about, because I agree with you (surprise?), you want a benchmark test that you can use for all WB's regardless of if it's for CPU/GPU. Benchmark to me is a bigger rad+better fan. Sorry that I have an opinion.

Gr8 artikle d00d, you roxxxx00rs!!!!! kkthxbuhbye

Happy now?

 

I'm not criticising you for raising a good point, I just don't agree with you. It's great to see folks who think and criticise. That's what the forums are all about.

 

Enough brass to keep Fred Dibnah happy! I was wondering why the GTX 260 & 280 (Forceware 180.84 ) were tested with different drivers to the GTX 285 & 295 (Forceware 181.20). I'm currently debating whether to buy a GTX 260 "216" or a Powercolor 4870 1GB, and wanted to see how the GTX faired against the 4870 with the lastest drivers? P.S. both cards are around the same price.

 

I believe that all the cards have been tested with the same setup, and I do not think that it would be fair in the benchmarks and the people who do them to change the standard to a 120.2 or a 240. There is nothing wrong with that fan, even it is slightly old. It is a quality fan that has most likely been used for other benchmarks.
I get your point, but I think we should just stick to the standards.

 

A 3C rise over idle temp with a single 120mm radiator is pretty strong evidence that temps will be just fine with a larger radiator. It's not like performance is suddenly going to decrease when going to a better setup.