• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

[wccf] NVIDIA Maxwell Based GeForce Graphic Cards Reportedly Shipping In March 2014

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
I wouldn't mind if they came out with a 28nm Maxwell part that includes the ARM core and unified memory just to get all the bugs worked out before the higher performance 20nm parts arrive later in the year.
 
KompuKare said:
Surely, the reason that Hawaii's power consumption might benefit more from aftermarket cooler is because like the classic IDC posts show, leakage at 95°C is a lot worse than at 83° or so where the reference GTX 780TI runs at. So yes, in theory both would probably benefit from aftermarket coolers but the 290X should get a bit more from it.

In theory that is: but as blackend23 pointed out, non-ref models tend to have fancier power delivery, run at different voltages and get to charge premium for being factory overclocked. Now in theory AIBs could bin parts for their OC's but most don't (still have a sore point with my Gigabyte Windforce 7950 forcing 1.25V mostly because Gigabyte can't be bothered to bin their chips - mine runs LTC fine at 1022/1500 with 0.95V for instance).

But anyway, 28nm TSMC is pretty much equal for GPUs isn't it? And once you take into account a few factors (GCN's higher DP performance vs GK104, die size of GK110 vs Hawaii) they two architectures have fairly similar perf/watt. Even the TPU chart shows a max of 70/100 spread which is hardly massive considering the chart has everything from 650/7770 to 290x/780Ti.

Still, since the temperature/leakage/power consumption thing should be very well known to anyone working in the industry, that makes AMD's decision to run with the old blower reference cooler even stupider (although I suspect a beancounter made the decision not an engineer).
Click to expand...
Excellent post, the bolded part especially, so fanboys take note. You have two very competitive companies working with a now old process with two years+ of production parts; they have tweaked and tweaked this to get every drop out of it, and the net difference is negligible. Any variance you see it going to come from board design, individual coolers, RAM, etc. and especially what specific software (and therefore drivers) is being tested. The fact that they are so close is the most impressive thing here, not the pissing contest some of you are trying to instigate.
 
blackened23 said:
So what happens when you put an aftermarket cooler on a Kepler card? Same thing, right? Like I said Hawaii is a fast chip in the right conditions. But the conditional "but aftermarket will make it efficient!" clause is ridiculous because if that were true, it would apply to kepler also
Click to expand...

It *does* apply to Kepler though. After market cards have a better performance per watt than the reference design.
 
80nm: early 2007
65nm: late 2007 (< 1 year)
55nm: early 2008 (< 1 year)
40nm: mid 2009 (1 year)
28nm: late 2011 (2 years)
20nm: mid 2014 (wtf happened?)
 
sm625 said:
80nm: early 2007
65nm: late 2007 (< 1 year)
55nm: early 2008 (< 1 year)
40nm: mid 2009 (1 year)
28nm: late 2011 (2 years)
20nm: mid 2014 (wtf happened?)
Click to expand...

R+D became exponentially more expensive, basically. Also, TSMC can't keep up with intel - they do not have FinFETs yet and their 20nm will be roughly equal to intel's 22nm.

Basically, intel has a vast warchest for R+D while TSMC doesn't. That R+D costs billions over the course of years, IIRC intel did R+D for FinFET over the course of 8 years. TSMC's 20nm is still in research due to cost and time factors, they don't have the resources that intel has.
 
blackened23 said:
Basically, intel has a vast warchest for R+D while TSMC doesn't. That R+D costs billions over the course of years, IIRC intel did R+D for FinFET over the course of 8 years. TSMC's 20nm is still in research due to cost and time factors, they don't have the resources that intel has.
Click to expand...

This.

TSMC's "14nm" process is the process that their 20nm process should have been. It's the same density, but with FINFET transistors. They had to abandon the 20nm planar HP node because they couldn't get it to perform better than their SoC node, and now they're rushing to try and get out 20nm with FINFETs and magic marketing to make it seem equivalent to Intel's 14nm.
 
blackened23 said:
R+D became exponentially more expensive, basically. Also, TSMC can't keep up with intel - they do not have FinFETs yet and their 20nm will be roughly equal to intel's 22nm.

Basically, intel has a vast warchest for R+D while TSMC doesn't. That R+D costs billions over the course of years, IIRC intel did R+D for FinFET over the course of 8 years. TSMC's 20nm is still in research due to cost and time factors, they don't have the resources that intel has.
Click to expand...

TSMC leverages R&D expenses from their customers. In fact TSMC's R&D + their top 10 customers R&D expenses is more than Intel's and Samsung's process R&D combined.

All these fabs have been working on FinFET's for years, more than a decade in all likelihood. Fact is, there was no need for FinFET's at 22nm - Intel just decided to go early.
 
No? Let me help you then Blackened.

If FinFET's were needed at 22nm, why don't TSMC, Samsung or GF's 20nm processes need them?

And the answer is? They weren't needed at 22nm, Intel just went early with them. LOL!

Not that it's helping them at 14nm mind you.
 
Of course TSMC and samsung want FinFETs, they just don't have the time or money investment that intel has. Intel started their R+D well before either samsung or TSMC did, we're talking YEARS AND YEARS AGO, while TSMC/samsung are just now trying to catch up. TSMC is now scrambling to make it work.

That's just...the biggest load of crap i've read here in a while.
 
blackened23 said:
Of course TSMC and samsung want FinFETs, they just don't have the time or money investment that intel has. Intel started their R+D well before either samsung or TSMC did, we're talking YEARS AND YEARS AGO, while TSMC/samsung are just now trying to catch up. TSMC is now scrambling to make it work.

That's just...the biggest load of crap i've read here in a while.
Click to expand...

If you're talking about your own posts then yes I agree. It's not about "want" it's about timing. Do you think FinFET's are some kind of magic that only Intel can do? Do you know that TSMC demonstrated 25nm FinFET's in 2002?

2002 - http://www.tsmc.com/tsmcdotcom/PRLi...tion=detail&&newsid=1357&&newsdate=2002/12/11

GF - http://www.globalfoundries.com/technology/14XM-FAQ.aspx

We have more than 10 years of FinFET R&D to build on as we prepare to bring this technology to production&#8211;far more than any other pure-play foundry. In fact, through our partnership in the Common Platform Alliance, we own more than 3/4 of the industry patents on FinFET technology. We are confident that this heritage of deep R&D will allow us to lead the foundry volume ramp of FinFETs as we did with HKMG
Click to expand...
 
Last edited:
SiliconWars said:
If you're talking about your own posts then yes I agree. It's not about "want" it's about timing. Do you think FinFET's are some kind of magic that only Intel can do? Do you know that TSMC demonstrated 25nm FinFET's in 2002?

2002 - http://www.tsmc.com/tsmcdotcom/PRLi...tion=detail&&newsid=1357&&newsdate=2002/12/11

GF - http://www.globalfoundries.com/technology/14XM-FAQ.aspx
Click to expand...


Lol...

"According to at roadmap, the industry should introduce 25nm transistors in 2007. The next step will be to refine the manufacturing techniques for producing these devices in high volume."

Looks looks they had a small schedule slip getting to the second part of that statement.

Using press releases as a citation = 🙄
 
Last edited:
Irrelevant. All that matters is putting Blackened's never-ending nonsense to rest. TSMC not having Intel's resources or FinFET experience, really? If you're gonna make statements like that you'd at least attempt to make sure they were even half-way correct instead of it being the total opposite. :hmm:

Not like he'd ever let anything as important as the facts get in the way though.

Infraction issued for callout. One week off for accumulation of points.
-- stahlhart
 
Last edited by a moderator:
Lets see. Intel shipped 22nm with Finfet for almost 2 years now.

TSMC ships 20^H^H..14nm with Finfets in maybe 2 years time.

GloFo...so far behind I cant even be bothered to estimate it.
 
96Firebird said:
I wouldn't mind if they came out with a 28nm Maxwell part that includes the ARM core and unified memory just to get all the bugs worked out before the higher performance 20nm parts arrive later in the year.
Click to expand...

There was another rumor that suggested that the Maxwell 28 wouldn't include the ARM cores. This makes me think that there won't even be any 20 nm GPUs in 2014.
 
MrK6 said:
Things must be pretty bad for nvidia to go on the record saying such things. Unfortunately for them, as the going gets tough it helps to have closer ties to a fab like AMD does (although I don't think things at GloFo are rosy either). The market is being strained from all sides, including technology, demand, and monetary angles, so things are getting interesting.
Click to expand...

To be fair, that article is over a year and a half old
 
Teizo said:
To be fair, that article is over a year and a half old
Click to expand...

Can you find anything disproving it? Anything more recent? I've been trying to find a newer source but people have gotten more tight-lipped about this topic... I remember that the conference people weren't happy that these slides leaked, NV didn't want to comment, either.
 
blastingcap said:
Can you find anything disproving it? Anything more recent? I've been trying to find a newer source but people have gotten more tight-lipped about this topic... I remember that the conference people weren't happy that these slides leaked, NV didn't want to comment, either.
Click to expand...

No. I haven't looked. Things may have changed, or may have not.

I imagine talking publicly about it is not good business either, which is likely why nothing more recent has come to surface.

/Don't bite the hand that feeds you.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top