Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
AMD Discussion
- Thread starter koyoot
- Start date
- Sort by reaction score
That is not the question.
The question is what do developers do when you use an AMD card.
Proprietary APIs restrict portability and customer choice.
I could swear you really must have something against AMD.
I never used their stuff and you don't see me (or other people here) constantly putting them down.
Did they hit you as a child?
You don't have to use anything from their portfolio you know?
But at least don't keep bashing, no need.
We all know by now that you buy loads of Titans and all. Good for you.
But please stop picking on what you seem to hate or something.
Have you realized that the only times you post is to make some stupid remark regarding AMD or saying how great NVidia is?
Man up already!! Wrong gender I guess.
I never used their stuff and you don't see me (or other people here) constantly putting them down.
Did they hit you as a child?
You don't have to use anything from their portfolio you know?
But at least don't keep bashing, no need.
We all know by now that you buy loads of Titans and all. Good for you.
But please stop picking on what you seem to hate or something.
Have you realized that the only times you post is to make some stupid remark regarding AMD or saying how great NVidia is?
Man up already!! Wrong gender I guess.
ATI's poor drivers and quick abandonment of recent hardware soured me on them at an early age. ("You mean that I have to buy new graphics cards to upgrade from NT 4 to Win2K - cause you're not updating the driver?" I did buy new cards - from Nvidia.)
Somewhat true - but often with laptops the most you get a choice of integrated GPU or separate GPU. You seldom get to choose the manufacturer of the separate GPU.
Malfunctioning funny bone?
Apology accepted.
Last edited:
Accepted, and I deleted my reply to that line. And my message was low-key because I recognize the language and culture differences.
Younger people in the US are moving to reject phrases with inherent gender bias. There's probably no clearer example than the phrase you used.
No hard feelings, case closed.
https://twitter.com/Daniel_Bowers/status/768270633125806081?s=09
So it appears that this really has hit the performance target. Now the core clocks is all we need to know. FP, Integer designs are similar/slightly bigger than Broadwell, so only thing that would let down this architecture is Cache.
Heh,: https://twitter.com/BitsAndChipsEng/status/769959012632825856
So it appears that this really has hit the performance target. Now the core clocks is all we need to know. FP, Integer designs are similar/slightly bigger than Broadwell, so only thing that would let down this architecture is Cache.
Heh,: https://twitter.com/BitsAndChipsEng/status/769959012632825856
http://www.linleygroup.com/index.php
Linley Group has released analysis on Zen, unfortunately behind a paywall.
Linley Group has released analysis on Zen, unfortunately behind a paywall.
http://semiaccurate.com/2016/08/31/amd-goes-7nm-new-global-foundries-agreement/
And to add to all this, today is the day, when 14nm HP process production started. This is the process on which Zen, end product, will be produced. It is hybrid between FinFET and SOI technologies. It is 20% better than LPP. This is also process on which can be Polaris respin produced, and named as RX 4X5 GPUs.
It is the same process on which IBM Power9 CPUs are produced
.
And to add to all this, today is the day, when 14nm HP process production started. This is the process on which Zen, end product, will be produced. It is hybrid between FinFET and SOI technologies. It is 20% better than LPP. This is also process on which can be Polaris respin produced, and named as RX 4X5 GPUs.
It is the same process on which IBM Power9 CPUs are produced
.
Last edited:
This is very similar to the slippery slope that helped get AMD in trouble in the first place. When they spun out the Fabs (Global Foundries) they had agreements in place to consume the wafers from the fabs ( so they had a fixed base line business ). This looks like similar context where GF bought the IBM fab and needs volume to "fill it up". (IBM's Power can't do that all by itself.).
This works fine as long as AMD has successful product which generate at least as much wafer production as they are precommitted to. Not so great if AMD can't get design wins to generate that demand. ( that 'deep-pockets' oil money is more of an upside mechanism for the foundry. I suspect they will run for the hills if AMD gets into trouble. )
"Go straight to 7nm" probably means they'll likely be on 14nm longer for most of the their line up. .
Better at top end performance (e.g., high clocks and willing to bleed a bit more leakage.). May not be better at Perf/Watt.
Well, this is just part of the story. When AMD spun off their fabs, they had to sign the WSA(Wafer Supply Agreement) with GloFo. And it was signed to the end of... 2024 year! AMD with WSA HAD TO buy certain capacity, whether it will go out to shops, or not, and that was driving their last losses in previous months, because until Polaris, everything AMD made was made on TSMC fabs, because they were cheaper to run, and... better. So AMD had two deals: with GloFo from amendment of WSA, and TSMC from technological perspective, and business perspective.
14 LPP process from Samsung, that GloFo used was good enough for Polaris as a pipe cleaner for the WSA.
From business perspective it was right move. Right now, yesterday, or two days ago mass production started of new 14 nm HP process, that Zen CPUs will resemble.
Holly crap:
http://wccftech.com/starship-amd-processor-48-cores-7nm/
This could be the one on the nnMP
Or not, Apple will not provide that many cores for sure. But cut down dies maybe.
http://wccftech.com/starship-amd-processor-48-cores-7nm/
This could be the one on the nnMP
Or not, Apple will not provide that many cores for sure. But cut down dies maybe.
We will see when 7 nm process will come out. 28 nm was with us for last 4 years, because of the 20 nm Failure. Similar thing will happen with 10 nm. So we can expect 7 nm around... 2020 year.
So better be prepared for marginal breakthroughs.
So better be prepared for marginal breakthroughs
.
Am I missing something? That puts Naples at half the speed of an Iphone 6 in single core, and only four times faster in multi-core!
Chapter 11 for AMD if that's really true.
Let's hope they get those speeds up. But, I think its a good sign that they're already testing them.
Well said. In the absence of competition, the behaviour of these corporation becomes worse and worse. You can set your clock by that. I hope AMD can make a comeback.
Do you refuse to allow Google apps on your systems, and use other search engines?
Did we not learn anything about mono-culture from the Microsoft monopoly trials?