It was worth a try, but I think that this is still a very driver dependent thing. Too bad though as it had some potential, I'm guessing it's the PCI-E bottleneck though, since SLI/Crossfire uses those plugs.
It's an example of a product where engineers who've said 'we can do this' have shaped the direction, but it's a product nobody needs at that price even if it did work.
Originally Posted by M7ck You mention that no price has been announced but Scan has it listed as in stock at £268.55 Inc VAT
Thank you! Didn't see that and MSI didn't tell us they had been shipped to the UK yet. I've added an addendum to the article in the price section, but because we never change scores after publishing we won't add a value score. :)
the ASIC design concept is good, and very well made SoC. but the software just wasn't isn't there.
perhaps the software intercepting API calls may not be the best solution? im not sure, but it sure has a lot more potential than nVidia/ATI's software split frame. even if the benchmark doesn't so it yet.
I can only assume this is a result of either hand picked games specifically chosen (as they show a significant peformance increase) or something happening at a driver level (as noted in this review its very complicated and picky).
I am quite glad that Bit have done this review (and a damn good one at that) as it shows in real world conditions Hydra still has some way to go (being polite).
Its annoying as the idea and principle are excellent.
I think that if Lucid had managed to develop a purely hardware based version with no additional drivers required then both Lucid and MSI would have been on to a winner here.
MSI have always been keen to develop and work with new technologies and bring together ideas in to one neat solution, this is one that looked and sounded darn good on paper and is a shame it didn't really meet expectations, but then this seems to be a common event for MSI.
Well, the hardware is obviously solid, and the software is fairly bullet-proof if you didn't get any driver issues (if I'm fair, I was expecting Hydra to be a horror-story of BSODs and corruption and... well, you get the idea), Provided the venture capital funding doesn't die because of the bad reviews, Lucid could get something quite special out if they can get the improvements seen in 3D Mark to translate to everything else.
I won't be buying one (I wouldn't be buying one anyway) but it's something to keep an eye on. Regardless, something like this will always be more of a headache than a single card of any flavour.
Originally Posted by tad2008 I think that if Lucid had managed to develop a purely hardware based version with no additional drivers required then both Lucid and MSI would have been on to a winner here.
Absolutely! The hardware is solid, but it was never to be the case.
It seems to me like an obvious and forgone conclusion. Adding an extra chip into the process of piping graphics through your motherboard and into your monitor, having that doing complex things etc, it just makes common sense that there would be a performance hit right off the bat.
The fact is though that they did it. It works. It can be done. Now maybe they might get some interest from the bigger companies to help it use it's full potential or to revise it for the next board. It would be nice, but I doubt it will happen.
Actually if we look at it from the perspective of chip makers, this really hasn't been done in the past, so Digi has a point, this is a good step forward, although it isn't very advantegeous now, later on it will prove to be much better.
Kudos to MSI and Lucid for getting it to work at all, but really, is there that much demand for something like this? I was under the impression that multi-gpu systems are still fairly niche and out of those prepared to spend £500 on graphics cards how many have thought "gee, I wish I could run an ati card with my nvidia one in sli?". The whole point of spening so much money on a setup is to get the maximum performance, and, considering that there is usually only one obvious card on the market at any one given time, would they not just go for two of the same? The main situation I see this being useful is in some sort of QA machine for software developers (mainly games), but the need for separate drivers from Lucid/MSI precludes that.
personally, i'm amazed this thing shipped. not only that, i'm dumbfounded the people involved on either side still have jobs. this is a massive fail. who in their right mind would buy this thing? potential? who cares.
i should try this at work today. im going to go tell my boss that i can't take on anymore projects because i want to take existing projects and design a way to make them run slower. then, if i still have a job, i want to do just that, and then put those hydra-inspired projects into production. surely when the company vp's have to wait to get their reports they will say "this works great, lets see if we can sell this process to other companies....give trig a raise."....lol
"This means it had to not only create a multi-GPU hardware solution that bypassed the considerably strong driver ecosystems of both ATI and Nvidia graphics cards, but also develop drivers for every possible graphics hardware combination to work in every game, efficiently."
From what I understand I REALLY don't think that's what's going on here.
thats the way things are done in the IT industry everywhere isn't it?
Actualy I'm still quite interested in the card, aparently it overclocks well, it does have CF and if I find I can get some fun coding with the thing then I'm in (if the price falls and the onboard sound is up to scratch)
Ok, first off, I thought this was a great review as always, but I have a bone to pick with the opening remarks, "We've been hounding MSI for a while to get us their retail Big Bang-FUZION board. We've waited patiently as the..." <--- This statement contradicts itself. You can't say you were hounding someone for something and then turn around and IN THE NEXT SENTENCE say you were waiting pariently. That's all I'm going to gripe about, I swear.
Originally Posted by Emon Unless you meant supplemental drivers.
The ATI and Nvidia driver multi-GPU ecosystem works on the presumption that it will be on a certified hardware with the link cables attached. The Hydra has to make the game software running work in a multi-GPU mode (if applicable), while keep the graphics cards in single GPU modes and forcing them to only render a portion of the screen or frames told to render by the Hydra software. Not necessarily odd/even frames given the dynamic load balancing. The software and hardware controlling the GPUs has to be a lot more dynamic than what Nvidia and ATI currently use.
Quote:
Originally Posted by K3TT I thought the motherboard is meant to be really good, read the preview on 3 different sites.
It was a controlled preview run by the Hydra team. It did NOT use the MSI board but a Hydra engineering sample motherboard. Given I know what happens at this type of event, I can only allude that they might likely pick games, games areas and AA/AF modes that particularly showed a performance advantage. 3DMark scales particularly well in Hydra for example.
MSI itself denounced the previews as not representative of the final product and its Fuzion board.
We've tried to pick games from the support list you lot actually play. Popular titles that we know work and provide reproducible testing results.
EDIT: It could just be the 32-bit version works and the 64-bit does not? We don't test 32-bit any more here but the previews were done on 32-bit and MSI internally test with 32-bit.
Quote:
Originally Posted by Scootiep Ok, first off, I thought this was a great review as always, but I have a bone to pick with the opening remarks, "We've been hounding MSI for a while to get us their retail Big Bang-FUZION board. We've waited patiently as the..." <--- This statement contradicts itself. You can't say you were hounding someone for something and then turn around and IN THE NEXT SENTENCE say you were waiting pariently. That's all I'm going to gripe about, I swear.
Oh FFS. :p I'll remove the patiently then. Hounding was meant to be sarcastic.
Comments 1 to 25 of 47
Replynow i know why nvidia wasn't concerned about hydra
Or maybe it's just that there's not enough bandwidth.
Intel X58 and P55 chipset's have finally given the gaming market what they want, the ability too switch platforms with out changing the motherboard.
Now AMD need to do something similar, allowing the competition to run SLi on there chipsets.
Thank you! Didn't see that and MSI didn't tell us they had been shipped to the UK yet. I've added an addendum to the article in the price section, but because we never change scores after publishing we won't add a value score. :)
perhaps the software intercepting API calls may not be the best solution? im not sure, but it sure has a lot more potential than nVidia/ATI's software split frame. even if the benchmark doesn't so it yet.
http://www.techreport.com/articles.x/17934/3
I can only assume this is a result of either hand picked games specifically chosen (as they show a significant peformance increase) or something happening at a driver level (as noted in this review its very complicated and picky).
I am quite glad that Bit have done this review (and a damn good one at that) as it shows in real world conditions Hydra still has some way to go (being polite).
Its annoying as the idea and principle are excellent.
With a good amount of time it may be the alternative that we've been dreaming of.
MSI have always been keen to develop and work with new technologies and bring together ideas in to one neat solution, this is one that looked and sounded darn good on paper and is a shame it didn't really meet expectations, but then this seems to be a common event for MSI.
I won't be buying one (I wouldn't be buying one anyway) but it's something to keep an eye on. Regardless, something like this will always be more of a headache than a single card of any flavour.
Absolutely! The hardware is solid, but it was never to be the case.
The fact is though that they did it. It works. It can be done. Now maybe they might get some interest from the bigger companies to help it use it's full potential or to revise it for the next board. It would be nice, but I doubt it will happen.
i should try this at work today. im going to go tell my boss that i can't take on anymore projects because i want to take existing projects and design a way to make them run slower. then, if i still have a job, i want to do just that, and then put those hydra-inspired projects into production. surely when the company vp's have to wait to get their reports they will say "this works great, lets see if we can sell this process to other companies....give trig a raise."....lol
From what I understand I REALLY don't think that's what's going on here.
thats the way things are done in the IT industry everywhere isn't it?
Actualy I'm still quite interested in the card, aparently it overclocks well, it does have CF and if I find I can get some fun coding with the thing then I'm in (if the price falls and the onboard sound is up to scratch)
http://www.pcper.com/article.php?aid=815&type=expert
http://hothardware.com/Articles/Lucid-Hydra-200-MultiGPU-Performance-Revealed/?page=3
http://techreport.com/articles.x/17934/3
The ATI and Nvidia driver multi-GPU ecosystem works on the presumption that it will be on a certified hardware with the link cables attached. The Hydra has to make the game software running work in a multi-GPU mode (if applicable), while keep the graphics cards in single GPU modes and forcing them to only render a portion of the screen or frames told to render by the Hydra software. Not necessarily odd/even frames given the dynamic load balancing. The software and hardware controlling the GPUs has to be a lot more dynamic than what Nvidia and ATI currently use.
It was a controlled preview run by the Hydra team. It did NOT use the MSI board but a Hydra engineering sample motherboard. Given I know what happens at this type of event, I can only allude that they might likely pick games, games areas and AA/AF modes that particularly showed a performance advantage. 3DMark scales particularly well in Hydra for example.
MSI itself denounced the previews as not representative of the final product and its Fuzion board.
We've tried to pick games from the support list you lot actually play. Popular titles that we know work and provide reproducible testing results.
Read: http://www.bit-tech.net/blog/2009/11/16/what-s-really-happening-with-lucid-hydra/
EDIT: It could just be the 32-bit version works and the 64-bit does not? We don't test 32-bit any more here but the previews were done on 32-bit and MSI internally test with 32-bit.
Oh FFS. :p I'll remove the patiently then. Hounding was meant to be sarcastic.
-
« Previous
-
1
-
2
-
Next »
Discuss in the forums