[VideoCardz]NVIDIA GP104 and first Polaris GPUs supposedly spotted on Zauba

[3DVagabond]

http://www.3dcenter.org/news/reihen...cht-gp100-gp102-gp104-gp106-gp107-gp10b-gv100

Nvidia already has top to bottom lineup of Pascal referenced in drivers. Competitor can barely manage 2 new GPUs.

https://www.zauba.com/import-graphics-processor-integrated-circuits-hs-code.html

GRAPHICS PROCESSOR INTEGRATED CIRCUITS BGLW5
GRAPHICS PROCESSOR INTEGRATED CIRCUITS NT0EH
GRAPHICS PROCESSOR INTEGRATED CIRCUITS 433X3
GRAPHICS PROCESSOR INTEGRATED CIRCUITS 2E73U
GRAPHICS PROCESSOR INTEGRATED CIRCUITS 3R08A
GRAPHICS PROCESSOR INTEGRATED CIRCUITS JM601

But people keep liking to slander and FUD Nvidia, the much more successful company because of their jealousy and need to defend their fail company.

Nice to see their software division is ready for when they finally have hardware too.

 

[Silverforce11]

But people keep liking to slander and FUD Nvidia, the much more successful company because of their jealousy and need to defend their fail company.

Did you get the memo from JHH to explain why he pretended to show "Pascal" to the world when he was clearly faking it?

He must have thought the wood-screwed Fermi went down great last time, so its a good stunt to pull out again when they are MIA for next-node chips?

How exactly is it slander or FUD when its the CEO of NV that's doing such outrageous things?

 

[nvgpu]

That is not even relevant, Drive PX2 isn't even sold to consumers.

Nvidia wants to show off their products at GTC 2016.

http://www.gputechconf.com/

APRIL 4-7, 2016
SILICON VALLEY

Keep on trying to FUD & slander though.

 

[ShintaiDK]

Market share will show the result as usual, and it all points the same way as always

 

[Techhog]

So, with the new information, we can finally move on from the monolithic die dream to reality, right? AMD has a high-end GPU planned, and the low-end one is confirmed to use GDDR5 and us this not viable for use with an interposer.

Yeah, I know it's wishful thinking to think that this speculation will stop before the official unveiling of the GPUs...

 

[Ghost of Cyrix]

I think this also should silence those who blabber on incessantly about AMD having no R&D budget so they can't compete. They were first out with a next gen graphics API, HBM, and it's looking now like 14nm FF.

You would think <insert extremely premature conclusion derived about speculative products> should silence people voicing very legitimate concerns based on factual R & D expenditure and AMD's market position?

I would think the 980 Ti mercilessly beating down the R9 Failure X in nearly every way should silence those who blabber on incessantly about buzzwords such as "next gen graphics API" that Nvidia currently win at, "HBM" that Nvidia are currently dominating without, and complaints about people bringing up AMD's factually low R & D, which clearly limited at least Fiji's ability to compete as well and resulted in a lack of new chips to take on Maxwell.


It's understandable; AMD are in a serious pickle and their most diehard supporters will confirmation bias their way into convincing everyone and themselves that it'll be alright. And heck, maybe it will be; maybe AMD are very smartly spending their limited funds on a strong 2016. But cut the crap: their funds ARE limited, their position at the top end IS weak, and they DO have to actually prove themselves with actual results before anyone is going to stop worrying about a factually in-peril company.

Results, not buzzwords and empty speculation, are necessary. Period.



Sincerely, a proud R9 290X owner who can see what an utter fail Fiji was against GM200.

 

[3DVagabond]

You would think <insert extremely premature conclusion derived about speculative products> should silence people voicing very legitimate concerns based on factual R & D expenditure and AMD's market position?

I would think the 980 Ti mercilessly beating down the R9 Failure X in nearly every way should silence those who blabber on incessantly about buzzwords such as "next gen graphics API" that Nvidia currently win at, "HBM" that Nvidia are currently dominating without, and complaints about people bringing up AMD's factually low R & D, which clearly limited at least Fiji's ability to compete as well and resulted in a lack of new chips to take on Maxwell.


It's understandable; AMD are in a serious pickle and their most diehard supporters will confirmation bias their way into convincing everyone and themselves that it'll be alright. And heck, maybe it will be; maybe AMD are very smartly spending their limited funds on a strong 2016. But cut the crap: their funds ARE limited, their position at the top end IS weak, and they DO have to actually prove themselves with actual results before anyone is going to stop worrying about a factually in-peril company.

Results, not buzzwords and empty speculation, are necessary. Period.



Sincerely, a proud R9 290X owner who can see what an utter fail Fiji was against GM200.

You always use so much hyperbole to jump to conclusions?

Pascal is MIA and using Maxwell for a stand in. nVidia is very good at old news ATM. You didn't disprove anything I said. Reads well, though. Two points for effective spinning.

 

[gamervivek]

AMD's problem has been clockspeeds and the chip demonstration they did at CES at 850Mhz isn't optimistic on that front.

Anyway, more AMD chips on the move from zauba,

 

[Techhog]

AMD's problem has been clockspeeds and the chip demonstration they did at CES at 850Mhz isn't optimistic on that front.

Anyway, more AMD chips on the move from zauba,

You mean the demonstration that obviously wasn't running anywhere near full speed in order to exaggerate the power efficiency? We can't extrapolate anything from that demo. In fact, you could even say that it's possible to come to the exact opposite conclusion and that the actual speed much be must higher (1000-1100MHz).

 

[gamervivek]

1000-1100 doesn't cut mustard at all with nvidia chips not breaking a sweat till 1.4Ghz. If AMD chips are still most efficient at 850Mhz, then they better hope that Pascal doesn't widen the clockspeed gap that Maxwell had and GCN4 is way more efficient arch. wise than GCN3 of Fiji. Or else they will again be second to nvidia this gen.

 

[3DVagabond]

1000-1100 doesn't cut mustard at all with nvidia chips not breaking a sweat till 1.4Ghz. If AMD chips are still most efficient at 850Mhz, then they better hope that Pascal doesn't widen the clockspeed gap that Maxwell had and GCN4 is way more efficient arch. wise than GCN3 of Fiji. Or else they will again be second to nvidia this gen.

How do they compare at the same clock speeds? If you are faster at the same clocks then you don't need to clock as high.

 

[Techhog]

1000-1100 doesn't cut mustard at all with nvidia chips not breaking a sweat till 1.4Ghz. If AMD chips are still most efficient at 850Mhz, then they better hope that Pascal doesn't widen the clockspeed gap that Maxwell had and GCN4 is way more efficient arch. wise than GCN3 of Fiji. Or else they will again be second to nvidia this gen.

So, you're saying that AMD should be aiming for stock speeds of 1.3-1.5GHz? Also, what makes you sure that's actually going to matter in the end? Clock speeds have never really mattered all that much before. If a 1.1GHZ AMD card can compete with a 1.3-1.4GHz Nvidia card, what's there to complain about?

 

[gamervivek]

Clockspeeds don't matter?

AMD and nvidia have exchanged roles where AMD would be leading on the clockspeed front with smaller dies competing with nvidia's bigger ones. Now nvidia's smaller dies compete with AMD's bigger chips with the exception that nvidia also make a >500mm2 behemoth and of course their software supremacy. The result is complete domination at the top once the clocks are turned to the max.

AMD have the transistor density on their side though and it will be further helped with the 14nm vs. 16nm that nvidia are on. Currently a Fury X keeps abreast with a 980Ti with around 100-150Mhz clockspeed deficit and around 1bln. transistor advantage.

 

[PPB]

Why people think nvidia will keep the same clockspeeds as maxwell and amd will keep the same clockspeeds as cgn 1.2/3? Both. Are new architectures, the clockspeed targets may very altogether. After all, remind me for how many generation nvidia gpus clocked to 1.4ghz? (dismissing the obvious hiperbole of "breaking a sweat")

 

[sontin]

Everything from G80 to Fermi.

 

[Ghost of Cyrix]

You always use so much hyperbole to jump to conclusions?

Pascal is MIA and using Maxwell for a stand in. nVidia is very good at old news ATM. You didn't disprove anything I said. Reads well, though. Two points for effective spinning.

Hyperbole? Jumping to conclusions? You're the one jumping to conclusions. What I stated was the reality of how AMD have been not competing with Nvidia for the duration of the past year or two despite the hyperbole people like you have been spewing for quite some time.

Good job at avoiding anything I actually said, building up a strawman, and sidestepping what's actually happening in the market with more of your own jumping to conclusions though.

You made a dumb statement trying to shame people worried about facts and how they bring into question AMD's competitive stance. I shot that down with reality. There's nothing more to that. I reached no conclusions, you just fail to comprehend what's actually being said.

Fun fact: Objective GPU buyers want actual competition, not buzzwords and baseless bickering.

 

[PPB]

Everything from G80 to Fermi.

Then intel had the first 8ghz cpu :awe:

Nonsense aside, it is maxwell. Dual pumping certain units inside the gpu isnt at all the same as having your shaders altogether running at 1.4ghz.

 

[railven]

You would think <insert extremely premature conclusion derived about speculative products> should silence people voicing very legitimate concerns based on factual R & D expenditure and AMD's market position?

I would think the 980 Ti mercilessly beating down the R9 Failure X in nearly every way should silence those who blabber on incessantly about buzzwords such as "next gen graphics API" that Nvidia currently win at, "HBM" that Nvidia are currently dominating without, and complaints about people bringing up AMD's factually low R & D, which clearly limited at least Fiji's ability to compete as well and resulted in a lack of new chips to take on Maxwell.


It's understandable; AMD are in a serious pickle and their most diehard supporters will confirmation bias their way into convincing everyone and themselves that it'll be alright. And heck, maybe it will be; maybe AMD are very smartly spending their limited funds on a strong 2016. But cut the crap: their funds ARE limited, their position at the top end IS weak, and they DO have to actually prove themselves with actual results before anyone is going to stop worrying about a factually in-peril company.

Results, not buzzwords and empty speculation, are necessary. Period.



Sincerely, a proud R9 290X owner who can see what an utter fail Fiji was against GM200.

This post, woof. That is how it is. This is why I couldn't buy Fury X, and took off my red shades.

AMD does everything first, yet they seem to be last where it matters - generating revenues/profits.

EDIT:

I should have kept reading, effin nailed it:

Fun fact: Objective GPU buyers want actual competition, not buzzwords and baseless bickering.

Once the red-tint comes off, you start to realize just how poorly AMD competes.

 

[Ma_Deuce]

AMD does everything first, yet they seem to be last where it matters - generating revenues/profits.

Does manufacturer profit really matter to you more than performance?

EDIT:

I should have kept reading, effin nailed it:


Once the red-tint comes off, you start to realize just how poorly AMD competes.

You have got some heavy tint alright, but I think you just changed your shades.

 

[ShintaiDK]

Does manufacturer profit really matter to you more than performance?

These are linked. To get profit you need performance metrics.

 

[tviceman]

That's a fake slide, they just made that up...

All the info is in the main article:

http://venturebeat.com/2016/01/15/a...o-full-graphics-immersion-with-16k-screens/2/

This the the important point:



For them to go public and say that they are ahead of NV at least several months is big, it means they are CERTAIN and their lead could be even longer. This is why JHH demoed "Pascal" with 980M chips...

"OVERCLOCKERS DREAM"

AMD talks a great game, when do they deliver in a timely fashion? Nvidia hardly ever talks, they just deliver competent products out of the gate that are balanced and don't need 1+ years of maturity. Everyone making a big deal about a prototype unit JHH displayed during CES is a nonissue. They said PX2 wouldn't be available until Q3 of this year. Anyways, I hope AMD delivers first; they need to. AMD is currently losing on all fronts: perf/w, perf/mm2, and perf/transistor. They can't afford to lose on time to market too.

 

[3DVagabond]

Clockspeeds don't matter?

AMD and nvidia have exchanged roles where AMD would be leading on the clockspeed front with smaller dies competing with nvidia's bigger ones. Now nvidia's smaller dies compete with AMD's bigger chips with the exception that nvidia also make a >500mm2 behemoth and of course their software supremacy. The result is complete domination at the top once the clocks are turned to the max.

AMD have the transistor density on their side though and it will be further helped with the 14nm vs. 16nm that nvidia are on. Currently a Fury X keeps abreast with a 980Ti with around 100-150Mhz clockspeed deficit and around 1bln. transistor advantage.

They have one GPU that's faster and that requires hard O/C'ing as the workload shifts to GPU limited (higher res). I don't call that complete domination. I also think the "software supremacy" is just an extension of "AMD drivers suck" rhetoric. Boost is typically higher than 100MHz-150MHz clock advantage for nVidia. It's more like 1GHz Fiji vs. 1.3GHz GM 200 in real life.

 

[3DVagabond]

Hyperbole? Jumping to conclusions? You're the one jumping to conclusions. What I stated was the reality of how AMD have been not competing with Nvidia for the duration of the past year or two despite the hyperbole people like you have been spewing for quite some time.

Good job at avoiding anything I actually said, building up a strawman, and sidestepping what's actually happening in the market with more of your own jumping to conclusions though.

You made a dumb statement trying to shame people worried about facts and how they bring into question AMD's competitive stance. I shot that down with reality. There's nothing more to that. I reached no conclusions, you just fail to comprehend what's actually being said.

Fun fact: Objective GPU buyers want actual competition, not buzzwords and baseless bickering.

All I said is it's obvious that AMD is pushing ahead with R&D. They had low level API 1st, HBM 1st, and now appear ahead on the next node (I say appear only because nVidia could theoretically be playing possum). That's not lagging in R&D. You've brought all of these other factors into the discussion that have nothing to do with what I said and in no way disproves it.

We all know nVidia sells more. That's not the point. It's just the only thing people have to fall back on when anything positive is pointed out about AMD.

 

[Vesku]

Once the red-tint comes off, you start to realize just how poorly AMD competes.

It's just about what is available at the time. AMD was the better buy when it was 7970/7950 vs 680/670/770/760, 780s were in their own category until aftermarket 290s became available, then 970s (if you can get past being misled on memory config) and 980Ti.

If Nvidia doesn't launch Pascal at the same time then big Polaris will be the obvious buy for a while.

 

[GodisanAtheist]

While I generally agree with guys like Sontin and ShintaiDK's "brutal realism" about the relative positions of Nvidia vs. AMD from a metrics perspective, I can't help but admire where AMD is with how little they have.

Despite being outclassed in almost every technical category, AMD still manages to offer up fairly compelling solutions vs NV. I don't think anyone could say AMD is as bad off as they were in the 2900XT/HD3870 vs 8800GTX massacre days of yore.

While lean times for AMD might mean being broken up and sold for scrap, I personally hope it means we get a competitor who is HUNGRY for some wins. They have progressively less and less to lose, so they may try more and more hail mary solutions like Multi-GPU on Interposer.

Sorry for the off-topic-ish-semi-thread-crap. My 2c.