Read more.Video published by Intel Graphics on Twitter appeals to engineers looking for a challenge.
Read more.Video published by Intel Graphics on Twitter appeals to engineers looking for a challenge.
If they do it... and how I pray they do... the graphics market can be a proper fight... at last!
Originally Posted by Advice Trinity by Knoxville
Hope so, is there a roadmap of development or still too early?
That's a lot of big talk. But pushing a graphics architecture in such a short window I hope they can pull it off how they say they are else they'll be a laughing stock.
It still goes that you can't just throw money, people and resources and it'll magically produce a winning product. But then again, Intel is so big and has so much money, they could somewhat design and build a product at a loss to compete with Nvidia in the HALO space and soak up losses for a while and build mindshare.
I wonder if, because of this, Nvidia may move into the CPU space (properly) in the next decade. Because it was original Intel vs AMD and ATI vs Nvidia. Now it is AMD vs Intel and AMD vs Nvidia. Soon it is going to be Intel vs AMD vs Nvidia and AMD vs Intel. But Intel and AMD will both have an advantage over Nvidia because they can build their hardware to work more intrinsically together?
And I hope Raja just doesn't spaff out an AMD architecture, they slap a new badge on it and go "hey everyone, look what we sto...cop...created!"
yawn, isnt this the same dude who used to keep talking every year when he was at AMD and at the end, a mediocre product release, so yah right.
I think they'd need to license some of Itanium instruction sets IIRC?
There is a cross licensing agreement between AMD and Intel on the x86 to share the improvements from both sides. That's why when microsoft was looking at doing emulation/bare metal on ARM, Intel went in heavy saying they'll sue if their proprietary instruction sets are infringed.
They would have to license x86 from AMD and the other improvements from Intel.
To support the full x86 instruction set you would have to license:
x86 from Intel
x86-64 from AMD as this only covers the 64 bit extensions to the original x86 architecture.
Intel also had their own x86-64 extensions so you might need some of those (non-Itanium ones).
IIRC the only other company that has licenses to build hardware for this is possible VIA with their EPIA chips like the Nano. They haven't really produced anything new for a good number of years and seem to be doing some stuff with ARM.
I can assure you it doesn'tVideo published by Intel Graphics on Twitter appeals to engineers looking for a challenge.
So many red flags there. I should be the target audience for that video, but it just tells me there isn't a bargepole long enough.
Last edited by DanceswithUnix; 31-01-2019 at 04:09 PM.
Pleiades (01-02-2019)
I think they burnt a lot of bridges with Tegra and now a lot of companies won't use them. It's a shame, my Nvidia tablet is one of the best I have used. I hope that the Switch being a success gives them a push and they get Denver back in the limelight.
I don't see them ever having an Intel license though. That was made quite clear in their last agreement after litigation with Intel that they weren't allowed to so much as emulate an x86 in software, and I don't see the relationship between the two companies improving.
Last I heard they were creating a RISC-V core to embed in their graphics cards. If that is decent performance they could be at the start of the wave of uptake for RISC-V and maybe do some interesting hardware based around that.
Ooh, a RISC-V core in their GPUs could be really neat! I would really like RISC-V to have an uptick, a massive community influenced general compute system that's open source to all, me likey.
Nvidia are pretty good at burning bridges, just like with Apple etc.
I do kinda want an Nvidia Shield for my TV but I just can't bring myself to give them dosh right now.
Pleiades (01-02-2019)
We've failed every time before, and this time it will be different with failures from everywhere else (raja etc). Is anyone on this team (gpu team I mean) from a project that was a great success? I see no reason this time will be different without better people who have not failed before at doing this stuff. I'm pretty sure Raja was behind most of Radeon vii which is IMHO also a failure. No VSR/DLSS/RT and same price with memory that makes the card unable to make a NET INCOME most likely & that is based on previous designs with HBM, which always require game bundles to make a dime (at least in consumer space). HBM should be PRO/workstation/server only until prices drop massively. If you can't make money, what is the point in the product release? I'm not surprised by the limited release despite AMD claiming otherwise. IE show sales if you're claiming that please, otherwise, I'll take the rumor. When people claimed that crap about the first Titans NV said they sold the first run (100k) in hours IIRC, and the same people (for the most part) buy them yearly as most of them make MONEY from them (they are not gamers by and large, they are PROS avoiding $5000-6000 quadro's etc).
No I don't think this time will be different, and the graveyard of defunct gpu companies keeps growing. Just google list of defunct gpu companies and go to wikipedia page Good luck Intel, hope you don't bring your old gpu chops to the table, as they suck and have no business competing to take down AMD/NV, both of which make pretty great products, whether you like either or not as a company. You get a great deal with AMD or the best product (at a premium, duh) from NV. You get the same story on cpu side, great deal AMD (but moving to best product soon, hope they CHARGE accordingly), or the best product again at a premium, from Intel.
That said, again, I think AMD wipes the floor with Intel for the next 18 months or more, the only question is if management is smart enough to choose INCOME over market share stupid pricing (as they've done repeatedly in the past, always going cheap/market craptastic pricing killing INCOME from great products). You can't win a price war with giants. You will go bankrupt LONG before you win anything. It is best to get INCOME, and grow from there with yet more R&D. You gain nothing gaining share if you can't afford the next round of R&D. This isn't rocket science, it's basic math. Yet another disappointing quarterly report of ~38mil for AMD. ROFL. Intel/NV just collectively LOL, while I cried yet again with AMD shares (LOL, ok, I know they'll make me money as 64 core servers hit, but still!! Why waste great R&D on cheapo pricing?). I really hope they charge what they are worth at 7nm, or they just deserve to go bankrupt for being so dumb (of course I'll sell stock long before then...LOL), or hopefully bought by SMARTER people that believe INCOME is better than share when broke. Even dumb management can't screw up 64 core pricing that badly to yet again blow net income for the next year. Once they hit, I think income goes upwards greatly and sends the shares higher (but run on hype, benchmarks etc, it won't last long if past management decisions are an indicator...LOL). ~337mil this year, probably more than a billion next year finally, but that needs to be a BILLION a quarter, not a year to catch NV etc. I don't think Intel will affect NV/AMD bottom lines in gpus at all, and I wouldn't want to be an Intel shareholder next year. IF AMD starts another dumb price war, Intel server income could be hit very hard. IF they price right, Intel won't be hurt much but AMD will gain massive INCOME (comparatively speaking vs now). It's comic Intel won't be able to knock off AMD even with paltry 38mil quarters (Not the first time they've tried as noted).
It's always fun reading a verbal diarrhoea from nobodyspecial. He has so many convoluted points and anti points in one post...
chj (01-02-2019)
There are currently 1 users browsing this thread. (0 members and 1 guests)