Intel is acting like it cares more about GPUs than CPUs, and that's a good thing

<snip>

2. While Ryzen did good. AMD made an error many years ago when they dropped 3DNow! support. Yeah it's only old games, but try to play a game which require it and say that Ryzen is good. AMD hasn't resolved the issue with 3DNow! support. Technically they put instruction call to MMX which is in same address space, but games don't support that to well or at all. Game calls for 3DNow (detected AMD CPU after all) and AMD CPU says... 'oh bollocks I don't have that'. And you get nice black screen or inverted negative textures or any other "charming" effects.
3Dnow support should have never been detected based on a vendor, and any software writer using it should have used the official recommended means of checking CPUID. If they didn't, then I have a hard time thinking their software was any good anyway (considering AMD plainly documents how to check CPUID at the beginning of the K6 documentation).
 
I disagree with the assertion. Just because Intel has done some work on its GPU driver doesn't imply that it cares more about GPUs than CPUs. Its CPUs have advanced a lot more than its GPUs in recent generations, and it's even releasing CPUs without GPUs (something it hadn't done for years). Intel also continues to work on future CPUs and discussed this at length a lot more than it discussed its GPUs.

So sure, Intel is giving its GPUs more attention, which is nice, but I don't think it's in any way more attention than it's giving its CPUs.
 
That's because they are about to have their CPU marketshare hit HARD for the next 3 years. It will take real time, effort, ingenuity, and luck to make a new architecture to take on Zen. 2022 is probably the earliest they will have a 7nm chip with a completely new design.

Until then they need to make it at least look like their GPU's can make up the difference on their bottom line.
 
It will be so overpriced(I can't imagine what kind of driver support they may provide) compared to AMD GPUs people may hardly buy
 
It will be so overpriced(I can't imagine what kind of driver support they may provide) compared to AMD GPUs people may hardly buy
Intel had no problem supplying drivers for their integrated graphics. Although they didnt have to support games a lot since many of those wouldnt work with playable fps.
 
It will take real time, effort, ingenuity, and luck to make a new architecture to take on Zen.

I'd be surprised if Intel doesn't have a competitive product with 10nm, as early as this year. And when I say competitive, I mean 'something which will run games better on average'.
 
Kick rocks Intel, no one asked for your discrete GPU entry b/c based on your unethical business practice patterns, w/ that nuisance being released, I'm sure it'll be overpriced, and probably won't work on anything but Intel hardware, lol
 
I'd be surprised if Intel doesn't have a competitive product with 10nm, as early as this year. And when I say competitive, I mean 'something which will run games better on average'.

You are going to be surprised then lol.

So answer this: Do you think Intel will launch a 125w 16-core at 5.5GHz for $400 this year? That's what it would take to merc Zen 2.
 
It will be so overpriced(I can't imagine what kind of driver support they may provide) compared to AMD GPUs people may hardly buy

I actually think Intel isn't going to price gouge with GPU's. They know they are not regarded well in the GPU world.

Furthermore while their 14nm+++ is about to be wholly inferior to 7nm for CPU's, I could actually see it being really good for GPU's.

In conclusion: I wouldn't be surprised if they launched a $399 card that takes on the 2080 and Navi effectively next year (Or whatever AMD/Nvidia has in 2020). It probably won't be the most efficient, but it might be around as efficient with really good 1080p performance. Also remember they can scale up to 700mm^2 with their fab.
 
Expect their usual evil marketing tactics or dirty trick to undermine and destroy competition anyway it can to try to dominate the market like they did with their intel cpu's. search it on Google. For example, just recently there was a benchmark for the i9-9900k and amd ryzen cpu. We knew the i9 was faster but what are the real numbers ? Yup, intel had to mess it up for some odd reason. Intel is dirty, please dont make GPU's if you hae an history of being deceipful and lying. Read more here: https://www.techspot.com/article/1722-misleading-core-i9-9900k-benchmarks/
 
That's because they are about to have their CPU marketshare hit HARD for the next 3 years. It will take real time, effort, ingenuity, and luck to make a new architecture to take on Zen. 2022 is probably the earliest they will have a 7nm chip with a completely new design.

Until then they need to make it at least look like their GPU's can make up the difference on their bottom line.

All the young people selling intel short. Intel has been here before. I remember buying 2 or 3 generations of AMD CPUs because they were both cheaper *and* faster than anything intel could drop. But then the core architecture arrived and all that changed.

Intel is quite good at riding on their coattails for years, getting eclipsed, and then returning with a vengence. They have massive amounts of money and still own their own fabs and the vast majority of the enterprise market (and sectors outside CPUs). Intel might just surprise everyone (again).

I for one hope they pull a giant rabbit out of their hat and put Nvidia on the defensive. People complain about Intels practices but what Nvidia has been doing for the past few years is just ugly. Nvidia needs some real competition and the sooner the better.
 
So answer this: Do you think Intel will launch a 125w 16-core at 5.5GHz for $400 this year? That's what it would take to merc Zen 2.

No, but neither will AMD. And no, that's not what it would take to beat Ryzen 3000. An 8 core that's 10% faster on average in games will be enough for most people who aren't AMD fans (I.e., most of the market) to buy Intel over AMD. I'd say even less, but I think that 10% is reasonable for Intel to achieve. Frankly I'm still waiting to see if Ryzen 3000 will beat the 9900K at games.

And yes, AMD might have the better product for productivity, like it did with Ryzen 1000, but that still didn't make consumers rush to buy Ryzen CPUs.

That's not to say that Ryzen 3000 will fail at getting some people away from Intel. I'm sure it will help AMD's market share. I just think that any assumptions that it will be something that Intel finds hard to get over are wrong.

(BTW, I find it annoying that when writing 'I.e.' the form converts the 'I' to uppercase. I tried to change it several times.)
 
Last edited:
What the hell do ppl do with all these cores ? does any game , beyond AOTS, take advantage of more than four crores? extra cores can be nice , but for most, not needed.
Until the consoles start multi threading I suppose ....ps5 xbox scarlet Nin San
 
I wouldn't be surprised if they launched a $399 card that takes on the 2080 and Navi effectively next year (Or whatever AMD/Nvidia has in 2020). It probably won't be the most efficient, but it might be around as efficient with really good 1080p performance. Also remember they can scale up to 700mm^2 with their fab.
Bartender, I'll have what he's drinking!

Tell me, exactly what is it in the past history of Intel product development, pricing, or corporate margin targets that makes this seem so likely to you? If you "wouldn't be surprised" by that, then it's safe to say you are probably incapable of being surprised of anything. Which is not a positive thing, since it probably makes it pretty hard for you to figure out what is actually likely to happen.

As for the article itself, the author seems to be regarding as potential signs of competitive *strength* things that actually indicate the opposite, they have to do these things to have even a hope of getting traction in this market. For example, Nvidia has taken its time in embracing FreeSync simply because there is no one in position to force them to. THAT is a sign of real strength, not having to accept things if it is not to your advantage.
 
Bartender, I'll have what he's drinking!

Tell me, exactly what is it in the past history of Intel product development, pricing, or corporate margin targets that makes this seem so likely to you? If you "wouldn't be surprised" by that, then it's safe to say you are probably incapable of being surprised of anything. Which is not a positive thing, since it probably makes it pretty hard for you to figure out what is actually likely to happen.

As for the article itself, the author seems to be regarding as potential signs of competitive *strength* things that actually indicate the opposite, they have to do these things to have even a hope of getting traction in this market. For example, Nvidia has taken its time in embracing FreeSync simply because there is no one in position to force them to. THAT is a sign of real strength, not having to accept things if it is not to your advantage.

Interesting how you insult me while you have absolutely zero supporting details to your argument lol. It is a fact that Intel can make 700mm^2 cards. It is a fact that Intel actually had incredibly efficient GPU's in the Ivy Bridge - Broadwell Era, but they hit a scaling wall at ~35w. All Intel needs to do is bring in some professionals (like Raja!) who can help them update and scale up their already competitive GPU architecture. It remains to be seen if they will manage to do so effectively, but no one should be surprised.

I mean really - you would be surprised if in 2020 Intel matched a 2080 with much better efficiency? The 2080 will be 2 years old by then genius.

God I love when I am accused of being an Intel fanboy - it just shows you how dogmatic some Nvidia and AMD fanboys are...
 
Back