How Long Until Intel Competes at the Top End?

How Long Until Intel Competes at the Top End?

  • 1 Year

    Votes: 1 1.6%
  • 2-3 Years

    Votes: 8 12.9%
  • 3-5 Years

    Votes: 14 22.6%
  • 5+ Years

    Votes: 14 22.6%
  • NEVER!!!

    Votes: 25 40.3%

  • Total voters
    62

FRZ

2[H]4U
Joined
Jul 7, 2005
Messages
3,543
So now that the first retail cards are hitting the market, I'm curious when [H] Members believe Intel will have a GPU capable of competing with AMD/NVs higher end options.

Do you even think Intel will consider creating higher end GPUs, or will they focus on low-mid range cards?
 
Pending on them having the stones to stick with it for the long haul - I'll give them 3-5 years to be competitive at the high end (note: not leading, but trading blows).

I base this on being surprised their A7-series competing admirably at 1440p. Personally, I think if they can compete at the midrange market, that would be a success for them; Especially if they can translate this to wins in the mobile sector.
 
Depends on how much money they intend to put in to it.
If they really really commit to it then it could be as soon as 2-3 years.

In this case it one giant unknown.
 
I tend to agree that the outlook for a high end GPU anytime in the near future is bleak at best.

I will say though that I was plsantly surprised by the Arc performance for a first ever release. Perhaps we might see them make a run at the $700-900 range in the next couple years.
 
I don't think that they need to compete on the "top end", they just need to have a decently priced mid-range card with stable drivers and good performance relative to the price. The real goal being to decrease dependence on Nvidia and AMD for GPUs in Intel (CPU) gaming systems. If, in a few years, you can buy mid-level Intel Desktops and Laptops that have an Intel CPU and a discrete Intel GPU and get "good" gaming performance, then they will have succeeded. The low-end is already covered by systems with iGPU. High-end systems are too niche to really matter much.
 
I don't think that they need to compete on the "top end", they just need to have a decently priced mid-range card with stable drivers and good performance relative to the price. The real goal being to decrease dependence on Nvidia and AMD for GPUs in Intel (CPU) gaming systems. If, in a few years, you can buy mid-level Intel Desktops and Laptops that have an Intel CPU and a discrete Intel GPU and get "good" gaming performance, then they will have succeeded. The low-end is already covered by systems with iGPU. High-end systems are too niche to really matter much.

Competing in the top end influences purchases down the product stack/line, this is known.
 
I will say though that I was plsantly surprised by the Arc performance for a first ever release.

I was surprised that performance was good when averaged out, but I did not think that the drivers would be as bad as they ended up being. Intel has been developing GPU drivers for their iGPUs for a very very long time and I assumed that at least some of that would translate into being able to make decent drivers for their discrete GPUs. It didn't.

Competing in the top end influences purchases down the product stack/line, this is known.

It depends on who your target audience is. The majority of "Gaming" PCs are sold to people who don't even know what a GPU is. They just want "a computer that can play games".
 
Midrange and highend competition would be great, even if it isn't top of the line. But with how fast they were considering abandoning ship already... never?

NVIDIA has the lions share, so Intel would need to step it up in a way that draws from them rather than AMD for it to benefit consumers
 
I was surprised that performance was good when averaged out, but I did not think that the drivers would be as bad as they ended up being. Intel has been developing GPU drivers for their iGPUs for a very very long time and I assumed that at least some of that would translate into being able to make decent drivers for their discrete GPUs. It didn't.



It depends on who your target audience is. The majority of "Gaming" PCs are sold to people who don't even know what a GPU is. They just want "a computer that can play games".
yeah I agree there! They probably hired on a bunch of old ATI/AMD programmers lol...
 
Assuming Intel continues to try and compete in the market and hires the right people, my guess would be 3-5 optimistically. 5+ more than likely.
 
a decade before they have all the bugs worked out (see gcn, which took until rdna2 before they had bug-free cards on release day,) and 5 to ten years before they catch -up on performance
 
Competing in the top end influences purchases down the product stack/line, this is known.
If Intel can get a mid range GPU - they will make all their sales from OEM's cause you can bet Dell, HP and all of them will get epic pricing to sell all "Intel" machines, like they used to back in the day, Intel Chipset, CPU and Wifi!
High end is nice, but it is not where the money is anyways. Low/Mid range is pure volume profit margin heaven.
 
If Intel can get a mid range GPU - they will make all their sales from OEM's cause you can bet Dell, HP and all of them will get epic pricing to sell all "Intel" machines, like they used to back in the day, Intel Chipset, CPU and Wifi!
High end is nice, but it is not where the money is anyways. Low/Mid range is pure volume profit margin heaven.
You have a point but Nvidia didn't dominate the world by being a bargain...
 
If Intel can get a mid range GPU - they will make all their sales from OEM's cause you can bet Dell, HP and all of them will get epic pricing to sell all "Intel" machines, like they used to back in the day, Intel Chipset, CPU and Wifi!
High end is nice, but it is not where the money is anyways. Low/Mid range is pure volume profit margin heaven.


the problem with mid-range: you cant subsidize the entire chip lineup unless you have a halo product

See Vega 64, where AMD was a year late after Polaris barely match NVIDIA's 1060, and if they hadn't had an Ether rush, it would have gotten totally nuked by 1070 Ti!
 

wELL,SURE THEY ARE -THANKS TO CUDA PROCESSIONAL SOFTWARE, -IT TOOK AMD HALF A DECADE OF GIVING AWAY FREE UNLOCKED COMPUTE CORES ON GCN BEFORE THEY MADE ANY PROGRESS IN THE SERVER MARKET! you hAV E TO BUILD AN ENTIRE API TO COMPETE IF YOU WANT TO BE TAKEN SERIOUSLY ENOUGH TO GET REPEAT SALES

Without functional consumer stuff COMBINED WITH SOFTWARE SUPPORT, Intel will be limited to whoever pays THem to build a supercomputer (only doe so far)
 
are you having a stroke?


no, i suffer from HD disease, and when you add the fact that I'm also running on three hours sleep, its way too much work for me to open a tool like ultra edit to quick fix caps mistakes

What makes this even worse: i cant touch type, so I'm stuck editing it after-then-fact as well as I can manage
 
no, i suffer from HD disease, and when you add the fact that I'm also running on three hours sleep, its way too much work for me to open a tool like ultra edit to quick fix caps mistakes

What makes this even worse: i cant touch type, so I'm stuck editing it after-then-fact as well as I can manage

So you are saying you are having a stroke.. got it.
 
Screenshot_20221124-134634.png
Battlemage
Celestial
Druid

Battlemage might be good depends when it's released.
 
If Intel can actually effectively produce the GPU in house on a good process node and give a good price/performance ratio, that would be good. Just having a top card that is competitive is not enough. Availability, versatility for users ability to use, OEM support, price/performance and options that make the card desirable all count.
 
Never. They will be competitive with the top end integrated stuff though. They have the money. They could hire the talent. They could be top tier, but they will not stick with it. They will take it as far as they need to, for the server and workstation market.
 
If you are referring to total revenue of this in dollars, the data center/enterprise segment has large TAM (total available market) and is growing faster than the consumer market, so I think intel will want to be a part of this growth segment. I'm guessing Intel, AMD and Nvidia will compete aggressively (each could be close to a third of that market share) within the next 5 years 😇
 
Intel has a lot of battles they're fighting right now. I kinda doubt they have the capacity to manage a video card division without a major shakeup.

They can burn money trying to get to the top, and that is effective to a point, because it turns out engineers when given money and left alone will generally be able to hit their target so long as they're in an otherwise cooperative non-toxic environment. But, I don't have much hope of that with Raja at the helm.

The general problem, which is alluded to by others, is that the top is also a moving target. Which was/is the problem with ATi/AMD trying to catch up to nVidia. ATi predates nVidia (1985 vs 1993), but they have always been a step or 5 behind - from TNT to the first GeForce to the present (yes, with notable exceptions such as the 9700/9800 etc). And catching up isn't just reaching parity with this generation, it's parity with the next generation. And as we're seeing that "reach" is another 30% or so, or incredibly significant and incredibly hard to do. If AMD is having trouble with really great management and engineers, then I'm not sure that "just" Intel money can tackle the problem.

So, I voted never, because that's the only option that's "closest". I think it's possible for them to catch up, but it will take a lot of concerted effort over a long period of time, with excellent management, and scalped engineers. Right now most of the parts of the equation aren't there. We'll see if they actually have the temerity to push through MANY financial quarters of losses in their "video card division" when they have to answer to shareholders or not. While it's wise from a long term perspective, so much of the stock market is short-term, short-sighted. If they can't eat it for at least 10 years, then never really will be the answer. And if they can't get the people like I said, then also the same.
 
Intel has a lot of battles they're fighting right now. I kinda doubt they have the capacity to manage a video card division without a major shakeup.

They can burn money trying to get to the top, and that is effective to a point, because it turns out engineers when given money and left alone will generally be able to hit their target so long as they're in an otherwise cooperative non-toxic environment. But, I don't have much hope of that with Raja at the helm.

The general problem, which is alluded to by others, is that the top is also a moving target. Which was/is the problem with ATi/AMD trying to catch up to nVidia. ATi predates nVidia (1985 vs 1993), but they have always been a step or 5 behind - from TNT to the first GeForce to the present (yes, with notable exceptions such as the 9700/9800 etc). And catching up isn't just reaching parity with this generation, it's parity with the next generation. And as we're seeing that "reach" is another 30% or so, or incredibly significant and incredibly hard to do. If AMD is having trouble with really great management and engineers, then I'm not sure that "just" Intel money can tackle the problem.

So, I voted never, because that's the only option that's "closest". I think it's possible for them to catch up, but it will take a lot of concerted effort over a long period of time, with excellent management, and scalped engineers. Right now most of the parts of the equation aren't there. We'll see if they actually have the temerity to push through MANY financial quarters of losses in their "video card division" when they have to answer to shareholders or not. While it's wise from a long term perspective, so much of the stock market is short-term, short-sighted. If they can't eat it for at least 10 years, then never really will be the answer. And if they can't get the people like I said, then also the same.

That's incredibly insightful.
 
AMD could not produce enough APUs for the consoles. Holding back availability of the Xbox and Playstation sales. Intel if they can get their nodes down maybe a much better partner for like Sony or Microsoft. Just another avenue which Intel maybe able to capitalize on for mindshare, developer support.

For future data centers, super computing, I don't see how Intel cannot push development of GPUs. Visual processing, parallel processing is much needed.

Intel I see need to address their manufacturing capability to start out competing the competition again.
 
Back
Top