>99% of the dedicated GPU market is owned by Nvidia (66%) and AMD (33%)
>Intel is going to release dedicated GPU units in 2020 off setting the market
Anybody else hyped that there will finally be competition for this dual-monopoly? Prices will be driven down, technology will vastly improve because of competition and gaymers will rejoice.
Intel should produce dual purpose CPU/GPU so gamers can actually build a 'console killer'
They tried that once and they cancelled it before they put it on shelfs.
*shelves
Don't get your hopes up.
Larabe is back in the game.
>raja goes to intel
>all of the poo memes stop
>man behind Evergreen and GCN
>only got handicapped because Lisa forced him to focus on Navi which could be a big hit and was entrusted with next-gen
AMD is fucked because he actually has money and buttbuddies at Intel
YOU MUST POO ON IT
>implying intel wants to compete in the first place
they'll go for the workspace meme and hope their name is enough for them to be adopteed in the professional environment, no gaymen cards
Holy shit
Intel CPUs are botnet housefires now and AMD CPUs are pretty good. Intel graphics are mediocre while AMD's recent integrated graphics are good enough to replace dGPUs for most tasks. Why would I care about Intel dGPUs? Why would I care about noVideo shit that won't work with Linux? Seems like AMD is the only sensible choice as of late. I'm going to be putting an AMD card in my Power9 workstation.
That makes no sense. Intel has historically had more poos
>Nvidia fans currently tend to stick with intel
>Intel releases GPUs
>Now Nvidia and Intel are in a GPU fanboy war
>Nvidia fanboys start to hate Intel
>Buy AMD CPUs
16th dimensional shuffleboard
>dual monopoly
>Nvidia fans currently tend to stick with intel
No they don't.
>dual-monopoly
Nigger, it's called a duopoly.
They want to get into ai hardware sector. And sell us their scraps.
I want an Apple GPU
They most certainly do
This.
Ironic, isn't it?
I don't mind their hypocricy though.
Momma Su is Queen.
milk > poo
check out this shill lmao
better pay twice the price for an AMD card aand half the perf top lel
>all of the poo memes stop
Only if Intel use their ME backdoor to stop everyone from posting them.
reddit post
Do you live under a rock?
wow, like what they tried to do in the early 2010s but never released it?
plz try pajeet
I've heard something about PowerVR wanting to re-launch in the desktop graphics space, mainly because they are afraid Apple are going to drop them for an internally designed GPU.
OH wait, i'm a slowpoke, Apple dropping powerVR caused their price to crash so hard a chinese company bought them out.
raja will inevitably poo in it so i'm not holding my breath (but am holding my nose)
raja wasnt behind gcn user
its not a secret that sony was throwing money at amd and designs at them for it...
Now we need Nvidia to start making x86 chips
intel and samsung. although i like amd cards for their open sourceness intel actually has better drivers hands down
I don't think they will try that again. :^)
>people who buy high-performance components tend to pair them with other high-performance components
of course they wouldn't buy amd dogshit
>Intel
>Prices will be driven down
hearty kek.
You're on rumors from 2 years ago user.
Apple already dropped Imagination like a rock and their stock price fell like a rock too.
They're fucked and have no money to do anything.
/thread
intel barely can release "new" CPUs,
now they are going to break the duopoly.
>Anybody else hyped
You now remember the days of nForce chipsets and when the go-to combo was nVidia / AMD 64
Based, I'm gonna build a dual system with Intel and NVIDIA only. Better keep out the poo in the loo (AMD)
It’s weird how the shills show up at a regular time every day.
t. PooMD shill
WE CRUSHED THE COMPETITION
This is wrong
Sony has nothing to do with GCN, they only affected Navi.
That makes nvidia's drivers the worst on the market. Wow.
>milk > poo
why do I laugh at this shit
Although AMD is paying us well, if we aren't up shilling by 8AM (whatever our local time zone happens to be) they send the storm troopers to break down our doors and raep us with stun batons. It also sucks having to wear tracking collars and head cameras, but making money in the shill game isn't as easy as it used to be.
> >man behind Evergreen and GCN
> >only got handicapped because Lisa forced him to focus on Navi
That doesn't make any sense.
> some man did good stuff
> told to make some more good stuff
> ???
> waah they handicapped him
Also he made some big steaming pile of shit known as Vega (that card without primitive shaders). He didn't live up to expectations about him, maybe he could do it at Intel.
nforce 4 sli master race! Socket 939 goodness. San Diego core Athlon 64 3700+ overclocked to 2.6GHz, dual 7900 GT's in SLi, 4GB Kingston Hyper-X. *sipps* those were the days.
>dual-monopoly
It's called a duopoly or an oligopoly.
And yes, fuck the duopoly. But, they are ALL American, which explains because they aren't hit by antitrust laws. No American gives a shit that the market is abysmal because all 3 responsible are American.
Wait... So the Intel shills were real?!
that's the point, raja wanted to focus on making vega work, but lisa forced him to give more attention to navi instead
Why isn't your shitty country making any GPUs?
>newfags too young to remember Larrabee
It was inferior to Nvidia and Radeon offerings at that time
Even if they got AyyMD's poo designer it will be slightly worse or equal to Pooega
Because all of the GPU intelectual property is trapped in the hell called 'United States copyright law'
It doesn't need to be as good. Look at Vega and Polaris. You just need to price it right.
Intel has most of the gpu market so it would be really really bad if they got all of it.
Why didn't your county just come up with it first then?
>Intel is going to release dedicated GPU units in 2020 off setting the market
See DRAM makers price fixing to give you a idea of how bad things can turn.
Discrete GPU market is about to undergo massive demand destruction in the 2020s. Just like what happened to discrete audio cards in the 2000s.
AMD RTG is going to make decent iGPU and semi-integrated GPU solutions that make mid-tier GPU obsolete. Intel is getting off its arse and is going to match them.
Nvidia is practically locked and they are so scared from losing marketshare that they attempted to do a preemptive marketing/strong-arming BS with GPP before whistle-blowers caught them red-handed.
Discrete GPUs are going to end-up being niche just how discrete audio cards exist today.
>told to put more resources on one project than the other
>HURR HOW HANDICAP?
They handicapped his work on Vega.
Sony had nothing to do with GCN gen 1. They started working with AMD after the first generation which is plainly visible because Sony's Mark Cerny played a large role in the expansion of ACE. Raja left AMD about 2 years before GCN released. GPUs take like 5 years to bring to production with the last 2 years being sampling and bugfixes, so Raja was definitely in charge of GCN.
Shills have been saying this since Llano, APUs are still worthless garbage. The ONLY way integrated graphics are ever good is if they use high bandwidth memory like GDDR5 in the PS4 or Xbox One X, but desktops obviously don't use this for system RAM due to latency constraints. Even DDR5 won't be good enough to make desktop APUs not shit.
I think you autists are missing the big point here. It doesn't matter if intel will release shit dedicated GPU products. What matters is that they'll break the duopoly or at least force Nvidia and AMD to reduce prices to compete assuming that Intel will undercut both. Hopefully they do.
>AMD RTG is going to make decent iGPU and semi-integrated GPU solutions that make mid-tier GPU obsolete. Intel is getting off its arse and is going to match them.
That's nonsense. That's basically a pseudo-console...dGPU's are going to last for a while until tech isn't stranglehold at infancy at the moment!
/thread
Nope, it is reality. The masses don't care about meme-tier gayming non-sense like 4K gayming, AF/AA + EXTREME DETAILS!, 200FPS.
It is the consequence of miniaturization. Almost everything on a modern motherboard used to be on separate discrete cards. Discrete GPUs are the next victims.
It will not be long until iGPU solution start yielding 1080p/60-100FPS gayming in majority of stuff. The masses will little to no reason to get a mid-range discrete GPU. Protip: Mid-range discrete GPU make up the lion's share of revenue in the entire discrete GPU market.
Ubiquitous iGPU already killed off the basic 2D/3D discrete GPU market. The only units that exist are either old hardware units or overpriced low-end Quadro/FirePro SKUs.
Intel makes this announcement every year or two. Nothing ever comes of their R&D in the consumer market. Larrabee ended up becoming Xeon Phi which apparently is only used for very specialized academic applications, render houses, governments, "deep learning", etc. Whatever R&D goes into the next iteration is just going to feed that market since Intel already has customers there.
What do you mean by the masses. Average Joe consumer never cared about discrete GPUs. Today they do all their computing on mobile devices and gaming on consoles. Discrete GPU cards are and always have been a hobbyist market and hobbyists want those periodic upgrades. The periodic upgrades are still giving something better every time, which didn't really happen with sound cards and I/O port cards.
This man will release nothing good.
>It will not be long until iGPU solution start yielding 1080p/60-100FPS gayming in majority of stuff
Are you aware that the 2200G and 2400G can barely match the Xbox One console, which came out in 2013? That Vega 8 is less powerful than the meme 750Ti, which is widely considered to be the bare minimum for 1080p gaming? Integrated graphics are a dead end. Intel wants to make dGPU's, the Intel/Vega Frankenstein SoC is actually a CPU + dGPU connected together, and laptops are moving towards using dGPU's in midrange and high end configurations. The industry is literally moving in the OPPOSITE direction of what you claim.
Nope, it is moving towards intergrated GPUs since Clarkdale/Sandy Bridge.
Intergrated audio didn't take over discrete audio overnight. It took years but it eventually happened.
The pieces for next generation of integrated are already in place. The 2020-2021 is the beginning of the end.
By the time 2025 rolls around. Mid-range tier discrete GPUs are going to be difficult to come by and justify. Discrete GPUs are pretty high-end niche tier stuff for towards who are willing to the extra mile.
Actually, discrete GPUs are mostly bought by kiddies (High-School/College-age). Hobbyist have always been a small, but vocal minority.
Even most of those hobbyist will eventually find it difficult to justify those "periodic" upgrades if they continue to give smaller and smaller returns. They just say "fuck it" when integrated GPU can take of their needs without none of the hassle and strings attached.
I have seen the consequence of miniaturization on the PC front since 1980s.
There's nothing special about discrete GPUs that make them immune to it.
Death to GPUs and heterogeneous architectures, be they from nVidia, AMD, Intel, ARM or anywhere else. All hail RISC-V and the vector extension which will bring both 3D rendering and professional vector processing back into unified, homogeneous software.
>99% of the dedicated GPU market is owned by Nvidia
Stopped reading there. This is such a blatant advertising post that I can't believe the mods are leaving it up.
You think an all-in-one cpu can beat a GPU? How amazing is this extension?
it's 5 in the morning and i already know that this is the most retard shit i will read all day
>all-in-one cpu
It doesn't have to be all-in-one. There's no intrinsic problem in having two CPUs in a system that are homogeneous in architecture but different in performance attributes. The point just being that they can both run the same operating system image, using the same paging structures. You'd basically just ask for your rendering threads to be scheduled on the CPU that performs vector operations better.
>How amazing is this extension?
It is pretty amazing, though. The key points being:
1. The vector unit has configurable registers, allowing more advanced implementations to effectively give wider vectors if a program requests fewer registers. If some rendering algorithm only needs five or six registers, it might get 128-wide vectors since there's room for it in the vector register file.
2. It has a vector-length register, similar to Cray systems, meaning there's none of the traditional SIMD clean-up ugliness for uneven elements. Even if you just loop over two elements, you still gain from using the vector instructions for it.
I do believe the vector extension may actually become the killer feature for driving RISC-V adoption. For professional applications (eg. things using CUDA today) it's potentially much better than GPUs, since it is fundamentally a CPU, so it can do all the standard nice scalar flow control while not losing anything in the vector department, and from there it'll likely trickle down to 3D graphics. The people working on it really seem to know what they're doing.
GPU are Trojan horse.
GPU begin compute high parallel unit +ASIC for graphics operations and high bandwidth memories.
Nvidia infiltrate engineer and science universities with CUDA and hey this graphics card for gaming too could run compute loads.
Nvidia want laptop discrete market begin biggest in sales by 2020.
Intel using dedicarte as Intel + Vega processor replace Vega with own GPU to attack Nvidia markets.