Discussion in 'Frontpage news' started by Hilbert Hagedoorn, Aug 5, 2019.
Even if Intel in 2-3 years puts out a $200 GPU, it doesn't mean anyone will want it...
You are pretty optimistic.I dont give Intel so much credit.
Optimistic about what? 14nm is fully detailed - it's density is known, it's D0 defect rate is known and we've seen it hit 5Ghz on ~200mm2 chips (4.3Ghz in AVX512 on a 350mm2 die). And then I even said I'm not sure whether or not that means they could build a decent GPU or get performance due to drivers.
Given the investment Intel has made in engineering, marketing, etc for their new GPU division....I'd expect they plan to follow through this time. Investors would likely be looking for the CEO's head if they fail to deliver.
Somebody will want it.... There's too many Intel fanboys for them not to get sales.
Yes it's detailed but using 14nm in 2020 when Nvidia and AMD will be on 7nm is not good for seling.Especially if they put some nonsense high price for dat gpu's.
This right here people keep repeating it ....tsmc 7nm is actually a bit less dence than intels 10 nm and their 14/16 nm is actually 20nm finfet because 20nm was expensive with minimal to no gains so none really used it ...so because 20nm finfet was vastly superior they renamed it 14/16 nm for marketing reasons , so for the love of god people stop talking about tsmc's nm like they are equal to intel's , that said tsmc seems to found a way to combat the clock speed regretion while intel is kind of stuck ....and except if they pull a miracle out of their sleeve i dubt we will see 5ghz on air in 10 nm.
HBM!! i dont know why i smell another GCN By intel
It's about time to see Intel 10nm cpu's i am curious how they compare to new zen cpu's.And if they reach 4.6-7ghz it will be even better for compare.
He clearly said they'll be releasing cards in all markets. Do you having reading comprehension problems or are you being intentionally obtuse in order to get that little dig in on Intel being greedy?
It's like some people have Intel Derangement Syndrome or something. If they cured cancer tomorrow, many of you would b*tch about all the doctors that would be out of work.
As for the topic, i hope they bring a very competitive card to the market in order to push innovation like we used to see years ago. Intel knows they can't half-ass this, and releasing a card that isn't as fast as what is available from their competitors would not be what it takes to peel away market share. Look how hard it is for an established GPU manufacturer, AMD, to gain market share. It's going to be that much harder for an unknown in this market like Intel to get people to jump ship. Personally, i think they have to beat both Red and Green in order to gain a foothold.
They released 10nm mobile chips all 4 core , while they do well on single core cause they can boost reletavely high on single core on multithreaded loads they are a bit faster than intels own 14nm mobile chips because on all core frequency they can not clock as high, and considering they released only 4 core ones i assume their yields are still really low forcing em to have to keep em as small as possible. And to answer jow they compare same old story core per core intel wins while amd wins on the gpu front eith the difference that the zen cores are really close ! Comparing to the pre-zen times
Couple things people need to remember here.
Intel has experience with GPUs, mainly iGPU but recent designs though lower performance compared to a dGPU, very respectable for what they can do. They have the capital for enough RnD to create something that's worthwhile, and they have a few names Raja Koduri being one on the team for the GPU.
Do I have faith it will be something that is cost effective at launch? Probably not, Intel parts usually have a higher price. Do I have confidence in Intel launching a decent product and growing on that? Definitely.
Nothing you said there has anything to do with my point. Do you have a reading comprehension problem?
Funny how that works both ways, especially considering I'm one of the people here who isn't pessimistic. Yet, you accuse me of being "intentionally obtuse". I said 1 thing you find "offensive", get all huffy about it, and then totally ignore the rest where I'm like "this is probably going to be a good GPU".
I am optimistic about the price performance ratio they will provide intel might want to gain some recognition in the gpu space ..so a 2060 /5700 like performance on 200-250 would be awesome ....i know i am dreaming but dreams are free!
That $200 figure was a typo - they forget the extra 0 at the end
I really don't care about the pricing nor the architecture, i just care
1) That it delivers decent performance, competitive with in Nvidia/AMD offerings, there is some that believe that 3 cant co exist on this market, would be interesting if that happens.
2) That its not a space heater nor need huge PSU to sustain them.
3) That has a good driver develop team, i assume drivers will suck for some time, but to put an effort into a team that its committed into fixing the issues.
Well, pricing still matters. GPUs are already getting way too expensive, so if Intel can drive down the prices of the competition then that's a big win for consumers. All 3 could definitely co-exist, and should.
Otherwise, I definitely agree with your other points.
S3, 3dfx,nvidia,ati,matrox,powerVR ....how awesome would have been if we still had all of em around?
Let's not forget to include Rendition on that list!
Only thing that is unrealistic is that there is nVidia on the list. There were many, but nVidia eliminated most of them by not following DX requirements and delivering higher performance.
Others foolishly delivered features that did cost transistors.
Even if their first real foray into this is 20% slower than AMD, they will catch up in less than five years, and they have the cash and the incentive to do so. I don't think that being disruptive this way (with pricing them at cost), is in the mentality of Intel as a company to do, but if AMD gains a tiny bit more momentum, Intel needs to do something radical to prove they're in the "game", so disrupting a major market might be the way to take the headlines off the terrible news from the CPU front.