Slashdot Mirror
AMD Announces Fiji-based Radeon R9 Fury X, 'Project Quantum', Radeon 300 Series
MojoKid writes: Today AMD announced new graphics solutions ranging from the bottom to the top ($99 on up to $649). First up is the new range of R7 300 Series cards that is aimed squarely at gamers AMD says are typically running at 1080p. For gamers that want a little bit more power, there's the new R9 300 Series (think of them as R9 280s with higher clocks and 8GB of memory). Finally, AMD unveiled its Fiji graphics cards that feature onboard High Bandwidth Memory (HBM), offering 3x the performance-per-watt of GDDR5. Fiji has 1.5x the performance-per-watt of the R9 290X, and was built with a focus on 4K gaming. The chip itself features 4096 stream processors and is comprised of 8.9 billion transistors. It has a graphics core clock of 1050MHz and is rated at 8.6 TFLOPs. AMD says there will also be plenty of overhead for overclocking. Finally, AMD also took the opportunity to showcase its "Project Quantum," which is a small form-factor PC that manages to cram two Fiji GPUs inside. The processor, GPUs, and all other hardware are incorporated into the bottom of the chassis, while the cooling solution is built into the top of the case.
But...but....but.... WHERE are the benchmarks AMD?!?!
There, I said it.
Interesting way to mount the graphics card in the tower case though.
These graphics cards are becoming so bulky, they're just about ready and willing to snap the PCI-express ports with their sheer bulk, assisted by gravity. Perhaps a better question to ask is whether case designers are willing to go along with AMD's proposed redesign?
READY.
PRINT ""+-0
I don't understand why the marketing people are so intent on telling us what we must do with their products. This whole 'gaming at 4K' seems like they are shooting themselves in the foot by excluding a huge segment of enthusiasts who are looking for any excuse to find a use for all that power. Why try to only sell your top of the line products to people with 4k monitors? I realize that consoles and just the overall cost of photorealistic graphics have somewhat reduced the need for high end cards, but jeez. At least try to sell high end products. Pathetic marketing strategy.
Quite an experience to live in fear, isn't it? That's what it is to be a slave.
The summary should say "(think of them as R9 290s with higher clocks and 8GB of memory)." Currently, it incorrectly says "R9 280s" instead of "R9 290s". That's a big performance difference between the 280 series and 290 series.
That Project Quantum looks very interesting @ 17 Teraflops! More info here: http://wccftech.com/amd-intros...
Cue all the "AMD drivers suck" trolls.
Who cares about a short video card unless it's also low-profile? That's what's needed to cram it into a tiny system.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
But is it powerful enough to run the Windows 10 Minesweeper game?! http://wscont1.apps.microsoft....
Seriously, no joke. The Win10 version of games are horribly resource hungry for fuck knows what reason. In the time it took to just load Minesweeper on the Win 10 tech preview, I loaded up a web browser, played an entire game of mines in it, closed the browser, came back, and it was STILL loading.
I originally played Minesweeper in Windows 3.1 on a 386sx 16MHz. I'm now on a 3GHz quad-core. On raw cycle processing power alone, that is literally 1,000 the speed (this is before accounting for enhancements to the architecture over the past 20 years). And yet the game struggles on modern hardware!? If this isn't the definition of bloat, I don't know what is!
Apparently the Leap models with Project Quantum have been having problem with users inadvertently causing time-space distortions including memory loss with at least one user vanishing without a trace.
And I say this as a lifelong nVidia card buyer (first card was a Canopus Spectra Riva TNT2, back in the day).
Si instead of outsourcing to India, AMD outsources to Indians living on a couple of islands in the middle of the Pacific.
So it's a slashvertisement for the same site as every other hardware-related article, but in this case not a single review site has benchmarks. AMD is being a giant cocktease, first they flash a picture, then they give us some speeches and marketing slides, with the promise that they're oh so worth waiting for. Meanwhile nVidia is putting it out there saying if you want it, come get it. I don't think this "hard to get" strategy is working out to AMDs advantage, I'm sure a lot of people are tired of getting stringed along.
In high end CPUs it's been years with no real acknowledgement from AMD that Bulldozer was dead, they're going to string you along as long as possible in the hope that maybe you haven't gone Intel/nVidia by the time they get the product out the door. But their credibility is wearing extremely thin, you know they have review samples ready if they're about to launch the product. That they don't want the facts out there speaks volumes that the facts are against them and the fanboy hype will turn into a backlash when they don't live up to the wishful thinking.
Live today, because you never know what tomorrow brings
MultiSound?
One day not too far off, graphics power will be so mainstream that separate companies doing only graphics will be ... doing Better Housekeeping type name selling.
Short sell.
I have to question AMD's future given these kinds of products. They tend to not address the masses but rather the niche markets. That maybe more profitable per unit sold. But how much demand is really there?
In February of 2012, I ordered a graphics card made on the TSMC 28nm process node, the Radeon HD7970. The cards hit the market on January 9, 2012. It has 4.3 billion transistors and a die size of 352 mm^2. It has 2048 GCN cores.
In June of 2015, the Radeon Fury X is a graphics card made on the TSMC 28nm process node, with 8.9 billion transistors, and a die size likely to be somewhere around 600 mm^2 based on a quadratic fit regression analysis using the existing GCN 1.1 parts as data points. It has 4096 GCN cores.
Aside from notable improvements to the memory bandwidth, are you really going to sit here and tell me that this card is much more than just a clever packing of two Tahiti (HD7970) chips onto a single die? They had to add a more effective cooling solution (liquid) to cope with the increase in heat generation in a small area that was caused by adding this many transistors in a very small area, which goes to show that they did fairly little in the way of power consumpton savings.
What is the likelihood that, in three years' time, they have made any significant innovations on the hardware front whatsoever, aside from stacking memory modules on top of one another?
To me this looks like an attempt to continue to milk yesterday's fabrication processes and throw in a few minor bones (like improved VCE, new API support) while not really improving in areas that count, like power efficiency, performance per compute core, cost per compute core, and overall performance per dollar.
When the HD7970's Tahiti cores are being sold as a re-branded R9 280X, and most games except Star Citizen don't seem to really demand more than one Tahiti worth of horsepower to run appreciably, there's very little motivation for me to "upgrade" to a chip that's basically two of what I already have packed onto one die with better cooling and faster memory. Especially when it's likely to come at a very steep price, which is much more expensive than simply buying another R9 280X and running them in CrossfireX.
As a gamer, I think I'm going to keep on waiting until TSMC and AMD/Nvidia stop dragging their heels. I've had enough of the 28nm node. That's three distinct families of GPU now that they've released on the 28nm node. It has gone on for too long. Time to move to a smaller process node. Until then, they won't be getting my money.