Separate names with a comma.
Discussion in 'AMD Flavor' started by Kyle_Bennett, May 27, 2016.
Wait, so is Intel buying RTG? or just licensing some tech for their iGPU?
I guess good for AMD, if they made this deal right.
does this mean dogs will be sleeping with cats?
maybe we get the phantom console too...
Licensing. Edited to make clear.
Intel in no way wants this to be public.
Wow I bet they don't. Wasn't Kaby Lake suppose to have an impressive iGPU? So much for that fucking CPU then in the iGPU department.
I mean Intel won't talk about it sure, but did they not think AMD would? And how would this not get out into the open.
There is 1 bad thing about this. Nvidia. If you can get an Intel CPU with a badass amd iGPU (specially laptops)......That is not a good sign for Nvidia. Let alone the Adoption rate for AMD tech in the future.
WOW I am shocked Kyle...Never would of thought of this happening
licensing, makes sense. I guess AMD could use some extra revenue coming in from this. Intel did have some agreement with Nvidia in 2011 to pay 1.5 billion over 5 years in cross licensing agreement. Looks like that expired and that little extra cash will go to AMD. Good for them. Hope that in return will help their R&D a little lol. Now to zen.
Your secret is safe with me.
What about the guy running the team at Intel who's work is now going by the wayside. Wonder how he feels?
This needs to be on the Front page! Let Hardocp be the website to leak the info!!!
Maybe he can work at AMD? I mean hell he might be the working closely with AMD to implement this.....Man I am floored to hear this, This is going to change the landscape. PEople might bash me, but Intel iGPU's are #1 in video card market share. With AMD in those iGPU's wonder who would get the credit?
No, the GPU on Kaby Lake (and pretty much 90% of the reason for Kaby Lake) was to support 4k, the GPU while better still lags behind a Iris Pro for example (I'm guessing on performance). They needed to have integrated support to keep themselves in the AIO department for when 4k content eventually becomes bigger in the PC business.
So I'm guessing this is more to do with licensing GPU technologies and not the actual GPU itself. Hopefully this partner helps AMD out in making better GPU's as AMD is doing enough to hurt themselves. Regardless they have nothing to gain to keep it secret other than to prevent Nvidia from flipping their lids.
TBH I think Intel doesn't give a Rats Arse about what Nvidia thinks lol.
I'm sure any corporation has the toolset to make business decisions like this a pain in the ass.
For instance if Nvidia has the capability to take this license to court, that would cost Intel money, while keeping it secret won't make any noise.
Well Nvidia tried this with Samsung and lost. So I do not see them doing the same thing here. This is a licensing agreement. Intel did not buy AMD. This is like Intel using AMD's X64 tech, and AMD using Intel SSE instruction set. I am sure AMD will get something in return to letting Intel use its GPU tech.
This actually makes sense. win win for both sides. AMD tech in Intel CPU so even if
no they wont. Intel paid nvidia 1.5 billion over 5 years starting in 2011. Guess what its 2016 and its good riddance. Intel never really wanted that anyways I think it was more like a legal issue that made them both kiss each other. So now we are 5 years in and this is being done.
Oh and I am starting to think Kyle's source is Koduri. ROFL
There were already rumors of this early this year in march when a few sites reported it. So looks like its finally a done deal. Intel's agreement with nvidia expires march 2017. So probably will be officially announced closer to that if ever.
Wow. My head hurts thinking of Intel APU's. I wonder if NVidia will get a lump in their throat.
Whoa that's really happening? This could be a legit threat to nvidia -- I like it.
Guys, you are blowing the deal out of proportion. Especially those that think it will make Intel iGPUs any less rubbish.
So I'm a little behind, AMD and Intel have concluded a licensing agreement for AMD tech to be used on Intel igpus?
Now why would Nvidia be pissed about that? This 2011 agreement was along the same lines of nvidia licensing tech to intel?
Also I remember a rumor? That Intel was cutting down on its graphics department some time ago, that isn't true right?
The licensing agreement I think was more from legal shit they were going through. From what I read Intel never really used any of the nvidia tech. It was more like shut the fuck up nvidia price I think lol.
So I found this article on the matter.
Yep confirms what I was sayiing, thanks. I didn't research much this time. Looks like intel didn't really want nvidia making boards for their chips. Now I am not going to lie, I had nothing but fucking shit load of problems with nvidia chipsets. Good riddance!!
Holy crap...That's the death nail.
Intel already license Nvidia IP. To add AMD IP is trivial and gives the team more things to work with.
I'm not sure I get this. iGPU is horrible no matter how you cut it. AMD's solutions were only mildly faster than Intels. Limited resources, limited power, limited memory bandwidth. How exactly is this to be beneficial to each?
And while APU's weren't a big seller they were a niche market where AMD had a slight lead. You would think they would need every sale they could get! Those sales will go to Intel now.
Easy. HBM......If Intel is able to make an APU like Vega with HBM it would be 1 monster gaming chip. TBH AMD usually fails in doing something well. Now with Intel having the licenses, lets see what a real Tech company can do.
This is good news for the future.
The real question is, what did AMD get in return? Money? Intel cross licensing?
HBM isn't the issue. Think IP. And Intel is already far ahead than AMD on this area. See Xeon Phi.
You are very limited in what you can do graphics wise without the IP from Nvidia and AMD. Specially since Imagination Technologies went downhill.
AMD will get a little bit of cash from it I expect, just as Nvidia. Nothing big but easy steady money. Perhaps 50M$ per quarter.
Trivial isn't it they are going to require AMD drivers and someone will catch it sooner rather then later especially if it is benchmarked and suprisingly similar to an APU.
Depends a little on how much time and money Intel put into it.
It expires in march 2017. But its obviously up for renewal. Any IP changes or such can take years to implement. However they will simply license from both from the looks of it.
HBM is not the answer.
And CPU with an IGP is an economic trade-off.
HBM is expensive.
So you will not see normal desktop CPU's with HBM anytime soon.
And with HBM...you limit the amount of RAM...you cannot expand the amount, you are stuck with what ever the SKU was born with.
I am so tied of hearing "HBM" being touted as the solution to anything...by people ignoring those facts.
Yep. Even AMD said so and why its restricted to the top.
Its quite easy to see products that gets this tech.
Fiji/Vega type cards. HBM
Various FPGAs. HBM/HMC
Common denominator=very expensive.
Xeon Phi is shit for a GPU. Perf/watt and die size efficiency is terrible for that work load. Xeon Phi is HPC number crunching not graphics.
It was in regard to memory. But even then, the most feature complete GPU so far sits in Skylake.
Short term this is correct. Long term (ie. 1-3yr) is probably a different story. If they don't or can't use HBM to fix the iGPU bandwidth issue they'll have to do on package DRAM of some sort or quad channel DDR4. Either of those options are also expensive but unlike HBM they don't scale to well. HBM is the most likely thing that'll get used eventually when you think about it.
Food for thought: AMD Fury's are currently going for $259 on sale right now. $359 is more typical. Probably no or little profit at that lower price but still. HBM is probably closer to affordable mass manufacture that you might believe.
In terms of your Fury cards, I bet you those are sold with a direct loss. And even then they essentially dont sell. Fiji have been nothing but a loss for AMD.
Now, what is the supposed HBM/HMC CPU going to cost? It serves very little to no benefit on the CPU side. So it needs to offer something EDRAM for example doesn't, and then add enough value to pay for HBM/HMC. So what does it offer that's so great its worth 100$ or more and possible limit you to 8GB?
The problem with IGP graphics is there is always a better dGPU option in the perf/$ area. And that the IGP cant handle the TDP issues with the faster memory. Look at AMDs APUs, its throttle monsters beyond hell and you want to increase the GPU load?
HBM/HMC also uses a lot more power than DRAM in terms of density. So you are directly increasing the TDP as well.
None of that matters to the previous point you were trying to make though. A bunch of simple x86 cores aren't going to be able to compete with a properly designed GPU. The fixed function hardware that is used alone pretty much guarantees it. Sure you can slap all kinds of fixed function hardware on die with a bunch of simple x86 cores too but the die size + power efficiency problem doesn't go away. Even with Intel's processes and fabs it won't work.
Stop being obtuse and read "long term". Stating facts for todays cost say ABSOLUTELY nothing of tomorrow. HBM1 was likely somewhat expensive, going forward those cost do diminish.
Yet Xeon Phi sells like hotcakes. Maybe you should look on what HPC loads there are and when what is used. The GPU isn't some kind of universal fix.