BTRY B 529th FA BN
Lifer
- Nov 26, 2005
- 15,093
- 312
- 126
MrK6
Diamond Member
- Aug 9, 2004
- 4,458
- 4
- 81
I think this is a company size issue (correct me if I'm wrong), and I imagine DAMMIT's driver/software team is smaller than NVIDIA's, and therefore they must choose where to spend their time. In that sense, gaming performance sells graphics cards, not Linux distro's.
Regarding the benchmarks, that's some damn good information. I would hope the "4x0" part is a 470, although I don't know why they would hide it. Considering it's performance, you would think they would want to point out that it was the slower of the two cards coming out, but maybe it's an NDA thing. I also think it's interesting how close rumors were regarding the performance, i.e. I'm not surprised by any of the information at all. However, there's still too much info missing to make a definitive appraisal of the situation, but more info is always better
.Current state of the art for ATI on linux:
Open source drivers have decent support for golden oldies (R200, R300) and very good 2D performance on newer chips. Stability is excellent. Compiz and desktop effects function well on 4 series and older hardware. Rudimentary support for getting text and VGA with the 5 series. Gaming runs at about 1/20th windows performance in wine and 1/4th the speed of fglrx with native opengl games, never mind the speed of windows drivers. Power management is coming along on some GPUs, and some 4 series models no longer run at full blast 3d clocks and voltage 24x7. No video decode acceleration yet. Pot luck displayport and multihead support. No crossfire.
Closed source (fglrx) is a nightmarish mess. Power management functions, but often locks the entire machine coming out of standby or exiting X. Semi-functional video decode acceleration. Does not support latest (read: 8 month old released) X server or modern kernels. Extremely slow 2D. Wine performance is awful (though better than the open source driver), and oftentimes buggy. Composited desktops work with some monthly releases and break with others. Crossfire is mostly unsupported. No support for cards older than the 2xxx series or mobility versions. Hit and miss support for 5 series (some still show an "unsupported hardware" watermark in X).
I have no complaints about NV's closed driver support on Linux. It supports all the desktop usage functions I require, so I haven't had to investigate any issues. The general concensus is STAY THE HELL AWAY from ati hardware if you intend to run Linux with anything more advanced than a basic VGA framebuffer.
Nemesis 1
Lifer
- Dec 30, 2006
- 11,366
- 2
- 0
I bet ATI is going to have a refresh sooner than latter . Anyone but a fanboy would have already bought Ati 5000 seies/ As it is with NV 7 months late. TO buy NV after its released is idoit (fanboy only territory) Ati will be serving up Its next processor very soon . If the wait for fermi was OK its just as smart to wait on ATi refresh which only gives fermi 1 life of 6 months or less as a ATI rolls out refreshes in sept. To buy fermi 1 is plain retarded. With Fermi 2 and ATI releasing new processor in sept.(aTI) Fermi 2 who knows when it sees light of day.
You wouldn't buy a 470 for $259? How about $200? How about $150?
There is a price point at which NV's new offering is the clear choice above all others and not a "special" purchase at all. It's just that we all expect the initial pricepoint will be high enough to make the few boards available to a very narrow group of people with very specific needs. Be that fandom or the desire to run Linux.
There is a price point at which NV's new offering is the clear choice above all others and not a "special" purchase at all. It's just that we all expect the initial pricepoint will be high enough to make the few boards available to a very narrow group of people with very specific needs. Be that fandom or the desire to run Linux.
Well, I wouldn't go that far... if the performance numbers we have are true, as long as the price is right than Fermi might be a good buy. But, I'm willing to bet that the price will not make sense for the performance it offers.
Well the expectation is that ATI will keep their pricing competitive. We know that logistically NVidia can't win in a price war.
OTOH, NV has far more cash and is able to afford a "loss leader" strategy if it makes sense whereas ATI has far less leeway. But we're in complete agreement. Everyone expects Fermi I to be stupidly expensive (even more so than G200) for the product's mercifully short lifetime.
evolucion8
Platinum Member
- Jun 17, 2005
- 2,867
- 3
- 81
ATi's drivers leave a lot to be desired, specially the ones that come with Linux, they're pathettic, slow, resource consuming and problematic with video acceleration. The propietary drivers fares much better, it had been improved nicely since 2006, but they're still behind nVidia, specially in OpenGL.
Looks like more of the same, then.
As far as I'm concerned, I'll stick with Nvidia until they make the drivers work properly. There's a whole lot of rather nifty 3D effects I've become accustomed to while running Ubuntu, and it would be a shame to lose them.
MrK6
Diamond Member
- Aug 9, 2004
- 4,458
- 4
- 81
As I said before, there really isn't enough information to qualify that unless you buy graphics cards only based on performance numbers. We still don't have any confirmed data on power consumption, noise (actual dB and quality), as well as feature support (how well does nfinity work?), etc.
I'd imagine the 5xxx series cards are much cheaper to make than the Fermi cards though. Add in the fact that Fermi is probably going to be available in extremely limited quantities for some time, I doubt you'll be seeing NVIDIA enlist in such a strategy. My guess is you'll see 4x0 parts go for a high, but competitive MSRP on launch (or whenever they're available), followed by a few weeks of retailer price gouging. Depending on how supply and demand meets up after that, you'll see prices adjust accordingly. I could be totally wrong here, but it's just a hypothesis based on the current rumors/info floating around.
Oh, if all you want is compiz cube you'll be fine with either open or closed source. Hell, that stuff works fine without 3d hardware with the latest Mesa releases and a zippy CPU.
It's only when you want tear-free video playback with correct colors, not boiling along at max clock rate & voltage, suspend/resume, TV out, switching VTs, wine, not having to hold your Xorg packages back or other more advanced desktop stuff where both open and closed source ATI drivers explode in different, creative ways depending on the revision. Catalyst 10.2 in particular seems like a trip to 2006 in the wayback machine.
This is not to say NV is prefect -- some people report problems with 6150 and other products. But since I also own an 8800GT it's been completely zero effort to have everything work. Especially in comparison to keeping my ATI powered machines running.
NV could price it for $2, and I wouldn't (really couldn't) buy it since all the cards will be bought out and resold for exorbitant prices on ebay. Simply put 8000 units isn't enough for the U.S. market, and I'm pretty sure out of all of those some will be defective and imagine the pain you'd have to go through trying to RMA a product you could barely buy in the first place.
Fact of the matter is almost everyone will have to wait for Fermi 2 anyways if they want a fermi card.
I think at this point availability is no longer a rumor since more reliable sources have gotten info after talking to manufacturers.
With so few Fermi (and Tesla?) cards I'm wondering which businesses will risk buying them.
You bring up an excellent point re: RMA and dealing with failures. If I knew a product I'm buying for business use is EOLed even before it is released I certainly would give it a pass no matter how good seems. If it fails the risk is either a lengthy downtime to find an exact replacement or a round of qualifying for a new product. Or both.
It simply doesn't make sense to do such a limited run of Fermi considering tooling costs. Something just doesn't add up.
You bring up an excellent point re: RMA and dealing with failures. If I knew a product I'm buying for business use is EOLed even before it is released I certainly would give it a pass no matter how good seems. If it fails the risk is either a lengthy downtime to find an exact replacement or a round of qualifying for a new product. Or both.
It simply doesn't make sense to do such a limited run of Fermi considering tooling costs. Something just doesn't add up.
It's the price of progress. NV probably thought TSMC could give them better yields, but what's done is done so they'll have to make due and cut their losses by selling what they've got.
I think a good example of what to expect would be ATI's 4770. It was supposed to be the first 40nm under $100, but manufacturing problems and shortages just drove the prices up so a 4850 ended up being cheaper. Now that was a small cheap chip so NV's problem is going to be worse since their die size is much bigger.
As for business I think they'd be OK considering how much they pay for professional cards NV will bend over backwards to provide them with support. I'm sure in the tesla line they will keep some spares around, but I don't expect the same in the consumer grade cards. It's one reason I never buy at the high end since I'm sure even RMAing a 5970 is difficult though not nearly as difficult as I imagine RMAing a 480 would be until supply issues are resolved.
So you know ATI's next gen product is 100% on track and will be released on time without any problems and will perform way better than current gen and will also be affordable?
Nemesis 1
Lifer
- Dec 30, 2006
- 11,366
- 2
- 0
You have ZERO rights to ask me that question. As I have read all your post. I don't need to know the facts as far as your concerned . You wouldn't know facts from fiction if it ran you over.
Heres a fact. Since june of 09 these forums have been filled with NV hype BS. Its time for an old ATI fanboy to start hyping northern lights beings how Fermi will be here in 30 days. Time to market for ATI . I do it out of respect for the company and my brother inlaw use to work for ATI now with Intel . So I am self appointed ATI AEG member. LOL . The differance is I do it for fun and sport. I don't get free parts for hyping and marketing.
Last edited:
MrK6
Diamond Member
- Aug 9, 2004
- 4,458
- 4
- 81
Given that Fermi is at most competitive with ATI's offerings, anyone who wanted or needed an upgrade has already gotten a 5xxx part, save for a few NVIDIA hardcores (see Hitler video). Since Fermi doesn't add anything new to the market, there's no reason to jump on it. Logically, it therefore would be wise to wait another six months for Northern Lights (this is, of course, assuming it isn't delayed) to see what it has to offer, considering it is supposed to be a new architecture.
No official reviews or benchmarks from unbiased, believable sources yet counters the first part of your sentence, and the recent steam hardware survey that shows out of 25 million users, only 3% have a 5700, 5800, or 5970 radeon card I beg to differ with the second part of your "anyone who wanted or needed an upgrade has gotten a 5xxx part".
Keysplayr
Elite Member
- Jan 16, 2003
- 21,209
- 50
- 91
Is that really what you're going to do Nemesis? Is it really that "fun" and "sporting" for you to constantly disrupt the already volitile forums?
MrK6
Diamond Member
- Aug 9, 2004
- 4,458
- 4
- 81
You have a lot to learn about the enthusiast sector, young padawan; we are but a speck on the horizon that is the video card market.
Hauk
Platinum Member
- Nov 22, 2001
- 2,808
- 0
- 0
Is it getting hot in here, or is it just me?
I know I'm not alone in those who've been waiting for a +1GB offering. 5870 CF was blazing fast and allowed me to max all details in every game tested. However, I felt I was hitting the ceiling with 4x AA or higher. The GPU's could render anything happening in the scene, but pan and move, bam, accessing and churning of the buffer. A heavily debated topic in itself, but damn it I've wanted a larger frame buffer just to see.
So yes, there's still a lot left to for ATI to offer. It was rumored a 2GB card would launch shortly after 5870's debut. It didn't happen. And now, it's delayed until Fermi's launch window. All eye candy, +4 AA, still a lot to offer..
I know I'm not alone in those who've been waiting for a +1GB offering. 5870 CF was blazing fast and allowed me to max all details in every game tested. However, I felt I was hitting the ceiling with 4x AA or higher. The GPU's could render anything happening in the scene, but pan and move, bam, accessing and churning of the buffer. A heavily debated topic in itself, but damn it I've wanted a larger frame buffer just to see.
So yes, there's still a lot left to for ATI to offer. It was rumored a 2GB card would launch shortly after 5870's debut. It didn't happen. And now, it's delayed until Fermi's launch window. All eye candy, +4 AA, still a lot to offer..
Please I've been building my own gaming PC's since the mid 90's before 3dfx was on the scene. But you did a great job of avoiding my factual answers to your insinuations.
Terrible logic.......
Not everyone is reading BSN every day and keeping up with launches.
When you are in the market for a card, you do research and pick the card that you think suits you best. Who cares when it launched compared to a competitor?
You're missing one important thing.
Fermi isn't a business mainstream product.
It's a tester. IMO no business is going to invest in HPC Fermi. It's just not a sensible business decision.
Fermi is marketed somewhat at HPC, yes, but IMO it was never meant to be a high volume HPC product. It's a "Yes, we can" product.
Yes we can... deliver impressive HPC performance
Yes we can... make sure we have support for important HPC features (e.g. ECC)
Yes we can... architect a product which fits your needs and improves on deficits
Now go and start planning your next cluster based on our technology, sort out your applications so they will run on our hardware, and next year, when you're prepared, buy our product, it'll be even better.
Businesses (AFAIK) don't go out and buy products on a whim. They plan these things, Fermi IMO is a product to let them plan. Release a v1 now, get businesses interested, and when you release v2, it'll be even better, and they will be queuing up around the block to buy them in massive numbers.
Fermi gets them talking. Fermi 2 gets them buying.
TRENDING THREADS
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 9K
-
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 7K
-
Facebook
Twitter