Slashdot Mirror
Intel Reportedly Designing Arctic Sound Discrete GPU For Gaming, Pro Graphics (hothardware.com)
MojoKid shares a report from HotHardware: When AMD's former graphics boss Raja Koduri landed at Intel after taking a much-earned hiatus from the company, it was seen as a major coup for the Santa Clara chip outfit, one that seemed to signal that Intel might be targeting to compete in the discrete graphics card market. While nothing has been announced in that regard, some analysts are claiming that there will indeed be a gaming variant of Intel's upcoming discrete "Arctic Sound" GPU. According to reports, Intel originally planned to build Arctic Sound graphics chips mainly for video streaming chores and data center activities. However, claims are surfacing that the company has since decided to build out a gaming variant at the behest of Koduri, who wants to "enter the market with a bang." Certainly a gaming GPU that could compete with AMD and NVIDIA would accomplish that goal. Reportedly, Intel could pull together two different version of Arctic Sound. One would be an integrated chip package, like the Core i7-8809G (Kaby Lake-G) but with Intel's own discrete graphics, as well as a standalone chip that will end up in a traditional graphics cards. Likely both of those will have variants designed for gaming, just as AMD and NVIDIA build GPUs for professional use and gaming as well.
While they may be able to get away with it in the cpu market, unless they actually make these GPUs price competitive it won't matter much. And since these GPUs will definitely have signed firmware and DRM, it isn't like they will be a compelling alternative for the Open Source crowd.
Get back to me when they offer a version with unsigned firmware and methods for technically savvy end-users to prove their GPUs are secure and we might have something to talk about. Otherwise it will be another has-been like the i740, i752, i754, all integrated models since the i810, and Larrabee for those who got to play with the Prototypes.
You need to go sit down...You're drunk.
Is it the third or are there even more failed attempts?
Intel 740, Larrabee and now this. Even if they are successful and finally get a miracle where they produce hardware that is actually good enough. They won't beat nvidia or AMD unless they use their fab tech to build a much bigger, much more expensive to fab chip. But let's just say they pull the miracle rabbit out of the hat. Their drivers will still suck for games. To be able to get a foot in this market you will need several years/generations with competitive hardware so game engines are written with explicit thought for you, games do tests and fixes on your hardware, a driver team works with game makers for a long time,etc. I just don't see the Intel videocard driver team being capable this way.
The only chance Intel maybe has is to convince the console makers to use theirs instead of AMD for the next consoles. With enough money/rebates and the great Intel sales magic to OEMs this might even work. But for discrete PC gaming this is all DOA. I just don't see how Intel can make money on this, not with the rebates they would have to give the console makers to actually "succeed". This sounds like another Atom/mobile CPU/ARM competitor fiasco where they burn billions.
They named a Graphics Processing Unit Arctic Sound .
Dumbasses.
#DeleteFacebook
Is it a Sound Blaster 16 with an Arctic Cooler, disguised as a Video Card?
I doubt i will ever buy an Intel GPU, but i would love it if they could put some pressure on nvidia. That being said, it would take years of losses for the project to become a serious threat in the market hopefully Intel is willing to suck-up those losses.
Normally it would be hard to crack this market but five stars are aligning right now that are going to make this easy.
1. Crypto Currency has made NVIDIA scarce and expensive. While that probably at it's peak and will wane for bitcoin and now etherium, new emerging currencies are going to emerge for which GPUs will still matter.
2. There about to be a paradigm shift to real-time ray tracing. GPUs have just reached the critical level of performance while new standards, drivers and libraries to support them are emerging that will bring this into the next generation of games about to be written.
3. VR and augmented reality are not come and gone. Far from dead they are just resting like parrot. Well maybe not like a parrot. Vapor ware like Magic leap is about to become real ware but the problem has been insufficient performance for real time augmented graphics.
4. And I save the best market for last. Driverless cars and self flying drones depend on GPUs. that market isn't even commerical yet. time to leap.
5. But none of the above matters. Nvidia and AMD can and are expanding into all those niches and they have the market channels and cost scales to do it. Competition cant get started. Unless of course you happen to have infitely deep pockets, a known history of selling loss leaders to crush competitors, and and superior channel to mobo makers who just gulp down your chip sets already.
Intel's timing is pretty good. demand rising, shifting requirements, actual need for improved performance, and deep pockets mean they can enter the market at the top teir of performance while competing on price without the worry of market share.
Some drink at the fountain of knowledge. Others just gargle.
/Oblg. Nvidia poking fun of Intel a few years back.
Maybe _this_ time will be different. We'll have to wait and see ...
Intel has been half-assing GPU capabilities for decades now. Why would anyone believe they'll do a good job on this chip? Are they giving the new guy everything he wants?
Intel has a history of programs that exist because "Intel needs to sell chips in [whatever market]". Not at all based on what customers want. They tried it for phones. Customers continued not to want Intel's offerings. Intel eventually gave up trying to push on that rope.
At first I wanted to make some joke about having the sustained fury of crabs scuttling about and reaching long claws fruitlessly in the dark of the ocean floor. But all that actually came to mind was the sustained fury of the towering boat tossing waves of an arctic storm and in fact how a GPU could render that power and scale.
Some drink at the fountain of knowledge. Others just gargle.
Intel Reportedly Designing Arctic Sound Discrete GPU For Crypto Mining
Comment removed based on user account deletion
Used to be every PC had to have a graphics card you can say that Intel failed the first couple of times but more PCs use integrated graphics than don't. I don't think they even want to build a graphics card just get the tech up so it meets or beats the low end from nvidia and AMD for their next gen CPU. Right now AMD is in the lead if I want a low end system I buy an APU cheaper than buying an Intel and video card system.
Their highest end card will be just on par with the lowest end card from 1-2 years ago from AMD/NVidia.
I'm sure it will have Windows 10 only driver support , and barely work with the fb driver on linux.
A board requiring 3 slots to accomodate the massive heatsink and power connectors galore.
And just to top it off, I'm sure it will have it's own Management firmware to spy on, err "help" users.
Seriously there isn't a place called Arctic Sound but here follows the first paragraph on sound as a geographic feature from the Wikipedia.
"In geography, a sound is a large sea or ocean inlet larger than a bay, deeper than a bight, and wider than a fjord; or a narrow sea or ocean channel between two bodies of land (see also strait)"
And you can get them in the Arctic so there!
Yeah, I'm sorry, but Intel has to put up or shut up.
They've been pushing graphics garbage out the proverbial front door for so long, they are working against their own history. And every time there's a new Intel graphics technology, they hype it before release and then it sucks. Raja Koduri or no Raja Koduri.
Yes this time might be different. It's not my intention or interest to buy into Intel graphics hype anymore. Put up or shut up, and don't hype before you've impressed us.
It is significant that Intel has to partner with AMD to release their new chip with Radeon graphics. Only through this partnership did Intel acquire enough IP and credibility, to bring a decent APU to the table. None of Intel's HD designs, nor their Iris Pro stuff, was remotely good enough.
https://www.extremetech.com/computing/261646-intel-launches-radeon-powered-cpus-hp-dell-announce-upcoming-systems
I get that the in-house Intel IP "works", in that it is meant as a low-end solution, it achieves that, the drivers are OK, etc. However they are far, far behind even AMD APU designs in terms of graphics performance. And they are pathetically far behind discrete GPU designs. Intel GPU designs are weak sauce and unimpressive, none of which stopped Intel from claiming (always pre-release), "just wait until you see our new chips!"
I'm not buying it anymore.
Intel had i5-5675C and i7-5775C a few years back that matched the APUs in performance, but thanks to an expensive 128MB L4 cache. This got used in the macbookpro too. But they basically give up on the idea though there are a few expensive Skylake + L4 eDRAM motherboards, much slower than the new desktop APUs.
Unlike the others, Intel has a lot of capacity to build their own chips - the others have to get their chips built for them. If Intel comes up with a tech that's especially suitable for a GPU, but not part of a standard fab, they can build it. No time wasted trying to get a third party to build a new line for them.
They might start with chips aimed specifically at massively parallel computation - probably more profitable to start there; think about building graphics cards later (if at all).
I think they can do it. I'll be curious to see what they come up with.