AMD reveal next-gen ‘Nvidia killer’ graphics card at CES?

AMD can beat Nvidia now if they had more access to 7nm process tech, next year will be very interesting, especially if AMD releases earlier than expected their 7nm+ cards which will be an improvement to RNDA (how much is anyone's guess) plus a process improvement. This can be a rather big combine result. Not sure how ready Samsung 7nm EUV, seems to have been in developement forever. Nvidia 12nm huge chips does not equate to same performance with smaller chips, even shrinking down Nvidia Turing Arch would be very huge chips on a 7nm process. Amper will need some more magic I would say to get the size down for good enough yields - now for their Tesla line that would probably not matter. For big Ampere next year - restricted mostly to Tesla, small Amper gamers. I see AMD winning in performance for gamers.

Same fantasy land talk we hear right before any AMD GPU launch. Nvidia is just sitting on tech waiting for AMD to release anything.

Edit: remember when AMD was releasing 7nm and everyone was like OOOOHHHHHH SSSHHIIITTTTT NVIDIA WATCH OUT!"... and then it didn't even meet the performance of Nvidia cards AND THEN Nvidia released MORE cards they didn't even need to AND THEN they just calmly mention "lol, oh yeah, oops, totally forgot, we have 7nm GPUs too lol"?

Classic.
 
7nm AMD still has not gotten to 1080ti (16nm) levels.

Nm's don't matter much if the performance ain't there.

Absolutely nothing "wrong" with AMD GPU's. Just the performance has not rocked any boats in a while.
 
7nm AMD still has not gotten to 1080ti (16nm) levels.

Nm's don't matter much if the performance ain't there.

Absolutely nothing "wrong" with AMD GPU's. Just the performance has not rocked any boats in a while.

Pretty much. Their cards aren't bad, they just aren't top tier and they never meet expectations, whether that's AMD's or consumers.

We have some guys on the DC team here using VII's and doing some serious work.
 
AMD can beat Nvidia now if they had more access to 7nm process tech

I get serious entertainment from statements like this that imply that Nvidia is somehow not part of this equation -- that they don't have the very same access, including more clout at the very same foundries.

Put simply, AMD doesn't have the technology developed to challenge Nvidia. We have every reason to believe that AMD could, however, doing so would require a few releases at parity, and well, AMD is steadfastly maintaining their two year parity tail.
 
I get serious entertainment from statements like this that imply that Nvidia is somehow not part of this equation -- that they don't have the very same access, including more clout at the very same foundries.

Put simply, AMD doesn't have the technology developed to challenge Nvidia. We have every reason to believe that AMD could, however, doing so would require a few releases at parity, and well, AMD is steadfastly maintaining their two year parity tail.

Since Nvidia is using Samsung for their next gpu, perhaps they don't have the clout or the same tech as you think they do.
 
Since Nvidia is using Samsung for their next gpu, perhaps they don't have the clout or the same tech as you think they do.

Which?

They're likely to continue to manufacture products at both fabs, which gives them the manufacturing flexibility to address the significantly higher demand for their solutions.
 
Which?

They're likely to continue to manufacture products at both fabs, which gives them the manufacturing flexibility to address the significantly higher demand for their solutions.

They have not confirmed that Samsung is the sole provider, but it's likely as they would have to spend more time and resources for two different fab processes. TSMC has it's hands full with customers and likely Nvidia was not going to to get the production dedication they wanted.

https://www.extremetech.com/gaming/294441-nvidia-will-team-up-with-samsung-for-7nm-gpu-technology

https://www.sammobile.com/news/win-samsung-nvidia-7nm-process-next-gen-gpu/
 
Same fantasy land talk we hear right before any AMD GPU launch. Nvidia is just sitting on tech waiting for AMD to release anything.

Edit: remember when AMD was releasing 7nm and everyone was like OOOOHHHHHH SSSHHIIITTTTT NVIDIA WATCH OUT!"... and then it didn't even meet the performance of Nvidia cards AND THEN Nvidia released MORE cards they didn't even need to AND THEN they just calmly mention "lol, oh yeah, oops, totally forgot, we have 7nm GPUs too lol"?

Classic.
Same fantasy about Nvidia and their RTX misleading marketing that fan boys sucked up. We will see one way another in the end. Funny you get so excited when Nvidia is possibly threaten.

I get serious entertainment from statements like this that imply that Nvidia is somehow not part of this equation -- that they don't have the very same access, including more clout at the very same foundries.

Put simply, AMD doesn't have the technology developed to challenge Nvidia. We have every reason to believe that AMD could, however, doing so would require a few releases at parity, and well, AMD is steadfastly maintaining their two year parity tail.

Put it simply, AMD does have the technologies and success at the smaller nodes and will soon be on 7nm++. Nvidia has not shown anything yet, you don't know neither do I. As it deals with Nvidia it is more a guess then a conclusion. Shrinking down Turing would be some rather big chips - hopefully Nvidia has more magic then Turing stupid pricing, big chips and poor implementation of RT which is somewhat useless.

As for not catching up to a 1080Ti, Look at benchmarks of TD2 Vulkan, size between the two chips and power requirements. New games and API RNDA will do better than Pascal or the 1080 Ti. Older games the 1080 Ti is indeed ahead but who buys a high end card for old games?

The Nvidia killer is coming. . . :)
 
So that's an honest no, thanks.

Nvidia produces many different products (as does AMD), so it absolutely makes sense for them to use different fabs according to strengths and availability. Same as AMD, or even Intel, and definitely the same as Apple.
 
Same fantasy about Nvidia and their RTX misleading marketing that fan boys sucked up.

No, but it'll be interesting to see The Faithful suck up AMDs marketing when they get around to releasing RT ;)

We will see one way another in the end. Funny you get so excited when Nvidia is possibly threaten.

This is trolling.

Put it simply, AMD does have the technologies and success at the smaller nodes and will soon be on 7nm++. Nvidia has not shown anything yet, you don't know neither do I. As it deals with Nvidia it is more a guess then a conclusion. Shrinking down Turing would be some rather big chips - hopefully Nvidia has more magic then Turing stupid pricing, big chips and poor implementation of RT which is somewhat useless.

You're arguing against AMDs entire existence as a GPU producer, basing your point entirely on some assumption that Nvidia has suddenly just decided to stop developing GPUs.

Do you really buy your argument?

As for not catching up to a 1080Ti, Look at benchmarks of TD2 Vulkan, size between the two chips and power requirements. New games and API RNDA will do better than Pascal or the 1080 Ti. Older games the 1080 Ti is indeed ahead but who buys a high end card for old games?

That's some wild extrapolation you've got going there :ROFLMAO:

The Nvidia killer is coming. . . :)

Perhaps in a decade... or three.
 
Which?

They're likely to continue to manufacture products at both fabs, which gives them the manufacturing flexibility to address the significantly higher demand for their solutions.
Likely, think, hope -> zero indication Nvidia is using TSMC smaller nodes while AMD has disclosed down to 5nm. You or I do not know what Nvidia has, every indication, brief or news has Nvidia with Samsung on the 7nm EUV for Ampere, which is struggling and delayed as in Samsung.
 
Likely, think, hope -> zero indication Nvidia is using TSMC smaller nodes while AMD has disclosed down to 5nm. You or I do not know what Nvidia has, every indication, brief or news has Nvidia with Samsung on the 7nm EUV for Ampere, which is struggling and delayed as in Samsung.

We have Nvidia's history and AMDs history as GPU manufacturers. Your hope is that these are going to completely reverse, and you're doubling down on that argument :ROFLMAO:
 
They have not confirmed that Samsung is the sole provider, but it's likely as they would have to spend more time and resources for two different fab processes. TSMC has it's hands full with customers and likely Nvidia was not going to to get the production dedication they wanted.

https://www.extremetech.com/gaming/294441-nvidia-will-team-up-with-samsung-for-7nm-gpu-technology

https://www.sammobile.com/news/win-samsung-nvidia-7nm-process-next-gen-gpu/

NVidia has been a premier TSMC customer for over a decade, so I doubt they could not get well planned for capacity they need, and we will really have to wait and see on where NVidia fabs it's 7nm parts:

https://pcper.com/2019/07/nvidia-working-with-both-samsung-and-tsmc-for-ampere/
... they commented on the rumors and speculation with this statement from Debora Shoquist, Executive VP of Operations: “Recent reports are incorrect – NVIDIA’s next-generation GPU will continue to be produced at TSMC. NVIDIA already uses both TSMC and Samsung for manufacturing, and we plan to use both foundries for our next-generation GPU products.”​
 
We have Nvidia's history and AMDs history as GPU manufacturers. Your hope is that these are going to completely reverse, and you're doubling down on that argument :ROFLMAO:
lol, trolling? future does not have to repeat, AMD cards are somewhat laid out in front of us and they look rather nice. Nvidia on the other hand, at least to me are very questionable. Eventually they will release Ampere, when is the question.

As for AMD and Nvidia, there has been times when AMD was clearly beating Nvidia so not even sure of your logic.
 
NVidia has been a premier TSMC customer for over a decade, so I doubt they could not get well planned for capacity they need, and we will really have to wait and see on where NVidia fabs it's 7nm parts:

https://pcper.com/2019/07/nvidia-working-with-both-samsung-and-tsmc-for-ampere/
... they commented on the rumors and speculation with this statement from Debora Shoquist, Executive VP of Operations: “Recent reports are incorrect – NVIDIA’s next-generation GPU will continue to be produced at TSMC. NVIDIA already uses both TSMC and Samsung for manufacturing, and we plan to use both foundries for our next-generation GPU products.”​
I remember that but was that not reflecting current state as in 12nm at TSMC and 14nm at Samsung? Has any indication from anywhere that Nvidia is using 7nm+ at TSMC? Your talking about two separate tape outs, revisions etc. if the same GPU, different GPU's I can see one or another but not both.
 
AMD cards are somewhat laid out in front of us and they look rather nice.

Two years behind in features and efficiency with mid-level performance on a newer node 'looks rather nice'?

As for AMD and Nvidia, there has been times when AMD was clearly beating Nvidia so not even sure of your logic.

Release cycles lined up for a month or so where AMD had a product that was competitive. That's as close as AMD has ever gotten, ever.
 
I remember that but was that not reflecting current state as in 12nm at TSMC and 14nm at Samsung? Has any indication from anywhere that Nvidia is using 7nm+ at TSMC? Your talking about two separate tape outs, revisions etc. if the same GPU, different GPU's I can see one or another but not both.

You could read the link and check.

But no, this was referring specifically to where next generation (7nm) parts are being fabbed. It was a specific comment to all the rumors that Samsung would be producing 7nm GPUs for NVidia.
 
I seriously wanted the R7 to beat the 1080ti for $500. Some here might remember that I did.

Like others, I'd love to see a 2080 Super contender from AMD at $600. Even without RT it would be cool and get Nv's attention.

And most likely big Navi will compete well with a 2080 S. And probably priced at $50 lower. $100 if were lucky. With no RT hardware.

$675?
 
As for not catching up to a 1080Ti, Look at benchmarks of TD2 Vulkan, size between the two chips and power requirements. New games and API RNDA will do better than Pascal or the 1080 Ti. Older games the 1080 Ti is indeed ahead but who buys a high end card for old games?
Yes true... just kind of sad that an almost 3 year old card is still being compared to AMDs latest. Where Navi does very well (ie, Vulkan), there are other instances where its still behind the 1080ti. Consistency seems more the issue.

The Nvidia killer is coming. . . :)
Poor Ampere.
 
  • Like
Reactions: noko
like this
I don't really care who has more clout with who. I want a 50-80% faster card than my Titan X Pascal for $750-800 from one of the 3 GPU manufacturers. I don't have any loyalty to them, none give me free shit so as a consumer I want the best value for my $.
 
I seriously wanted the R7 to beat the 1080ti for $500. Some here might remember that I did.

Like others, I'd love to see a 2080 Super contender from AMD at $600. Even without RT it would be cool and get Nv's attention.

And most likely big Navi will compete well with a 2080 S. And probably priced at $50 lower. $100 if were lucky. With no RT hardware.

$675?


As long as they release a flashable cut down card for $500 and reach within 3% of the performance of the flagship. I'm good.
 
7nm AMD still has not gotten to 1080ti (16nm) levels.

Nm's don't matter much if the performance ain't there.

Absolutely nothing "wrong" with AMD GPU's. Just the performance has not rocked any boats in a while.

????

The 5700xt = 1080ti at 1080p and 1440p

The radeon vii beats it at 4k.

Navi is a mid tier chip.

All of you guys are silly.

Is a 2080ti killer coming? Probably not. As it stands navi appears to be pretty bandwidth starved which is an easy problem to fix. Die is also tiny so they could easily add more CUs. Is 2080 super performance feasible if they add some CUs and bandwidth? Appears that way. If they did a 56cu navi (40% increase) it seems entirely possible that they could hit 2080 super performance, and still be in a 250-275 watt tdp envelop. They just need way more bandwidth as evidenced by the drop off in performance as resolution increases.

relative-performance_1920-1080.png

relative-performance_2560-1440.png
relative-performance_3840-2160.png
 
Last edited:
????

The 5700xt = 1080ti at 1080p and 1440p

The radeon vii beats it at 4k.

Navi is a mid tier chip.

All of you guys are silly.

Is a 2080ti killer coming? Probably not. As it stands navi appears to be pretty bandwidth starved which is an easy problem to fix. Die is also tiny so they could easily add more CUs. Is 2080 super performance feasible if they add some CUs and bandwidth? Appears that way. If they did a 56cu navi (40% increase) it seems entirely possible that they could hit 2080 super performance, and still be in a 250-275 watt tdp envelop. They just need way more bandwidth as evidenced by the drop off in performance as resolution increases.

View attachment 206511
View attachment 206512View attachment 206513

I stand corrected, AMD had a big win here.

My apologies.
 
A NVidia (2080Ti) killer, is a question of AMD willingness to go big, and timing.

Timing: IMO, it has to land before Ampere. This seems likely.

Now, it's down to willingness to go big.

A 20 Billion Transistor, 500 mm^2 Navi die, with 80 CUs (5120 SPs), 512 bit memory would leave the 2080 Ti behind. That's a true NVidia killer.

IMO, AMD is not willing to go that big. It's bloody expensive, and sales are slow at the top end, so making it's investment back is questionable.

Alternatively, Big Navi might aim instead get close to 2080Ti, but undercut pricing. So 400mm^2 die with 64 CU, but priced at something like $899.

IMO the latter, is a lot more likely what Big Navi looks like, and the killer part is a lot more debatable.
 
5700 efficiency is about on par with nvidia and the resident nvda long guy says they're two years behind, aka 'just caught up with skylake' round two, then calls people 'trolling'? The absolute gall.
 
because intel owns the laptop and oem market right now. When they push out their discrete graphics cards ..they'll have an easy way to undercut costs to system builders on both the PC and laptop markets (both of which are basically nvidia only market and they primarily live in intel machines).

If nvidia is pushed out of the laptop and oem pc market to a significant degree, they'll quickly see their value drop ...their cash dwindle.. which plays into their ability to invest in R&D and if they fall to 2nd place they immediately become irrelevant. They wont be able to compete on price (which intel and amd will volley over) and they wont be able to fund performance without volume sales ...

Do not underestimate the position intel will be in when it's discrete cards are ready to goto market in the near future. They will slaughter not only nvidia's but to some extent, amd's low end and laptop market share and probably eat into the cash cow mid-end market. and it'll happen quickly. Nvidia has no where to pivot, no alternative markets to grow into. It's living on borrowed time until intel drops their hammer.

Never Forget!!!
upload_2019-12-13_0-19-1.jpeg

Larrabee.
It was going to revolutionize too!
 
Are some forgetting RNDA+ and another node advantage? Efficiency and performance can be greatly increased here. RDNA was late and I see it one of the reasons why Big RNDA did not make it besides having 7nm process availability. AMD can't even keep up with Zen 2 chips, delayed their lower line of GPU's etc. AMD already had a bigger 7nm GPU, Vega II, 4096 shader chip so having a bigger RNDA chip was most likely mapped out, taped out etc. Just no room for it and probably just not worth it for them to cannibalize their CPU's or other GPU's over it. If AMD has almost exclusive access to 7nm+ or sufficient availability then maybe AMD will launch RNDA+ chips earlier than expected, supposenly TSMC yields are very good at this point for production and they are also in full production - 7nm+ would be a different process and not restricted by 7nm availability is my thought here. It would be nice if Nvidia also used this option and Ampere comes out as well.

As for RT, I am not certain if AMD will have separate dedicated area of the chip to help process matrix math for RT or include it in the shader arrays for matrix math speedup having faster access. Global Foundary, Samsung, Hynix are all making HBM2E a faster version of HBM2. At least someone, most likely AMD will have HBM2E, Nvidia I would predict will use on their next Tesla line whenever that will come out - I expected Nvidia to have their next Tesla line out already using Ampere - seems they are delayed.

I also think some are forgetting the massive speedup Turing had over Pascal with Vulkan and DX12 (myself included here), looking at those games to previous generation does make Turing look better and not just using older API games to compare them. Turing biggest problem, at least for me, is the pricing, initial hardware issues and overated RT (Nvidia marketing) did not help much. Hopefully AMD has a very solid solution, competitive to check Nvidia prices and to promote continued rapid advancements.
 
Last edited:
Wrong. They are three years behind. 1080Ti still owns them, and nvidia is releasing Ampere in a matter of months.

You mean the 5700xt has reached performance parity with the 1080ti at a mid range price point.

The 5700xt is within spitting distance of the 2070 super and it is within 10-20 watts of it depending on which review you read. Thanks to the process node advantage they are at parity in terms of efficiency (2070 super being roughly equal to the 5700xt in that they both do the same amount of work at the same power consumption).

Source on ampere coming out in a matter of months?
 
Are some forgetting RNDA+ and another node advantage? Efficiency and performance can be greatly increased here. RDNA was late and I see it one of the reasons why Big RNDA did not make it besides having 7nm process availability. AMD can't even keep up with Zen 2 chips, delayed their lower line of GPU's etc. AMD already had a bigger 7nm GPU, Vega II, 4096 shader chip so having a bigger RNDA chip was most likely mapped out, taped out etc. Just no room for it and probably just not worth it for them to cannibalize their CPU's or other GPU's over it. If AMD has almost exclusive access to 7nm+ or sufficient availability then maybe AMD will launch RNDA+ chips earlier than expected, supposenly TSMC yields are very good at this point for production and they are also in full production - 7nm+ would be a different process and not restricted by 7nm availability is my thought here. It would be nice if Nvidia also used this option and Ampere comes out as well.

As for RT, I am not certain if AMD will have separate dedicated area of the chip to help process matrix math for RT or include it in the shader arrays for matrix math speedup having faster access. Global Foundary, Samsung, Hynix are all making HBM2E a faster version of HBM2. At least someone, most likely AMD will have HBM2E, Nvidia I would predict will use on their next Tesla line whenever that will come out - I expected Nvidia to have their next Tesla line out already using Ampere - seems they are delayed.

I also think some are forgetting the massive speedup Turing had over Pascal with Vulkan and DX12 (myself included here), looking at those games to previous generation does make Turing look better and not just using older API games to compare them. Turing biggest problem, at least for me, is the pricing, initial hardware issues and overated RT (Nvidia marketing) did not help much. Hopefully AMD has a very solid solution, competitive to check Nvidia prices and to promote continued rapid advancements.

No one is forgetting RDNA2. It's expected for Big Navi. It's the main reason for the delay of Big Navi. Because RDNA2 is where you get HW RT support, and launching a top end chip without HW RT support is simply a non starter going forward.

But RDNA2 is also where AMD starts paying the RT tax, so they will have to give die space to dedicated RT HW.

Please skip the delusional comedy about AMD getting near exclusive access to a TSMC process, after using the Fab for a year, when TSMC has decade+ long relations with important customers like NVidia, and Apple.
 
You mean the 5700xt has reached performance parity with the 1080ti at a mid range price point.

The 5700xt is within spitting distance of the 2070 super and it is within 10-20 watts of it depending on which review you read. Thanks to the process node advantage they are at parity in terms of efficiency (2070 super being roughly equal to the 5700xt in that they both do the same amount of work at the same power consumption).

Thanks, I didn't feel like having to repeat that for the 50 millionth time but that's exactly what I meant. RDNA is basically at parity in performance, power use and price in the tiers it competes at, no matter how many NVDA shares you have. 5500 appears to have bucked that trend a little, maybe they cut it down too much? Not sure. Little cheaper and it'd be perfectly positioned.
Saying it's three years behind is utter fanboy bullshit, same thing when 'someone' claimed repeatedly AMD was 6 years behind on CPUs this month, guess that makes the 9900k the same? Lol. Mid range - performance range cards don't RT very well either (I don't consider 1080p an acceptable compromise in 2019).

It's always funny when they do catch up or do something as well, because 'insert x is just around the corner, it doesn't count'. Those perpetually moving goalposts again...
 
Thanks, I didn't feel like having to repeat that for the 50 millionth time but that's exactly what I meant. RDNA is basically at parity in performance, power use and price in the tiers it competes at, no matter how many NVDA shares you have. 5500 appears to have bucked that trend a little, maybe they cut it down too much? Not sure. Little cheaper and it'd be perfectly positioned.
Saying it's three years behind is utter fanboy bullshit, same thing when 'someone' claimed repeatedly AMD was 6 years behind on CPUs this month, guess that makes the 9900k the same? Lol. Mid range - performance range cards don't RT very well either (I don't consider 1080p an acceptable compromise in 2019).

It's always funny when they do catch up or do something as well, because 'insert x is just around the corner, it doesn't count'. Those perpetually moving goalposts again...

Agree. Navi closed the gap, they (Turing vs Navi) are now close enough in technology, transistor to transistor for the difference to be essentially irrelevant, there is no big performance/transistor or power/performance difference right now. (5500 is memory BW starved IMO).

The only factor (mentioned in a previous post) is how Big a chip AMD is willing to build to beat NVidia. Will it be ~500mm^2 to clearly beat 2080Ti, or will it be ~400mm^2 to do more like 5700Xt and come close to 2080 Ti, but undercut on price. I think it will be the latter.

I hope it really does gets announced at CES.
 
Agree. Navi closed the gap, they (Turing vs Navi) are now close enough in technology, transistor to transistor for the difference to be essentially irrelevant, there is no big performance/transistor or power/performance difference right now. (5500 is memory BW starved IMO).

The only factor (mentioned in a previous post) is how Big a chip AMD is willing to build to beat NVidia. Will it be ~500mm^2 to clearly beat 2080Ti, or will it be ~400mm^2 to do more like 5700Xt and come close to 2080 Ti, but undercut on price. I think it will be the latter.

I hope it really does gets announced at CES.
Well said, the transistor counts and even 'cores' (coincidentally but irrelevant of course) are so close it's quite uncanny, as if they both made the same thing via completely different engineering paths, never seen that before.
I think they'll go smaller as you expect, 7nm defects are a little higher etc so would make sense from that pov alone. Heard anything as to if it's EUV or not?

RT will be a very interesting part too, if they bother or not. Hints are that they have and I'm curious if their approach will be more efficient, it quite possibly could be without wasting as much silicon doing raster alone, so we might see it closer to 2080ti than expected. Later patents would suggest they missed the boat but who knows what tricks they have if consoles are close to being finished too..
Memory bandwidth is a pretty good point, it's about the only thing I can see AMD still being a little behind in (compression) but the overall effect when looking at 5700 is negligible as shown by the results, so not really a major factor there either strangely enough.
Hopefully they can throw some of that cpu division dosh at GPU r&d and pull a zen2 for mcm gpus in future... Fingers crossed - it would be nice to see competition along the stack in a more prompt manner.

That said hats off to Raja for pulling this one out of his ass. I certainly didn't see navi being this good and owe him an apology.. I know he lurks ;)
 
Last edited:
  • Like
Reactions: Auer
like this
RT will be a very interesting part too, if they bother or not. Hints are that they have and I'm curious if their approach will be more efficient, it quite possibly could be without wasting as much silicon doing raster alone, so we might see it closer to 2080ti than expected. Later patents would suggest they missed the boat but who knows what tricks they have if consoles are close to being finished too..

The RT HW "hints" are rather overwhelming at this point, between the patents, and both next generation consoles touting HW RT, I think Hardware RT in Big Navi is a given at this point.

The biggest potential for HW savings vs NVidia RT, is eliminating the superfluous Tensor Cores. They are used for DLSS Scaling (which has been ineffective), and theoretically for denoising, but nearly every RT game so far seems to instead be doing their own shader based denoising. So the Tensor cores are doing nothing necessary for RT.
 
You mean the 5700xt has reached performance parity with the 1080ti at a mid range price point.

The 5700xt is within spitting distance of the 2070 super and it is within 10-20 watts of it depending on which review you read. Thanks to the process node advantage they are at parity in terms of efficiency (2070 super being roughly equal to the 5700xt in that they both do the same amount of work at the same power consumption).

Source on ampere coming out in a matter of months?
Just imitating the AMDrones as far as how something better is always coming ;). Ampere is scheduled for h1 2020. 2019 ends in a few weeks. Nvidia isn't even on the 7nm process yet amd needed that just to come close. Once nvidia hits 7nm amd is done, over, FINISHED! :ROFLMAO: (/only half sarcastic)
 
Not sure it is hype stage, guessing or putting the puzzle together. Kinda like what this thread is about, AMD's Nvidia Graphics Card Killer at CES. While Nvidia large market has to be supported reliably with two foundaries is my take, how that will play out is anyone's guess here:
https://www.ocacnews.net/overseascommunity/article/article_story.jsp?id=246946

Nvidia Samsung move, which kinda hard to say that since they were already using Samsung may have been very prudent due to previous inpeding Tariffs between China and the US. So if AMD is not using Samsung, for whatever reason, makes Nvidia look a little bit smatter. My GUESS, is AMD will also use Samsung. Interesting is the South Korean President pledge to make South Korea the leader in high end processors, maybe Samsung is controlled by the government. Anyways high end as in what is in HPC, Datacenters are to me GPU's and CPU's. Or is it high end phones he was talking about :D. No clear idea what he really meant or if any real control over Samsung and a typical politician trying to take credit for something he has no clue about for election purposes.
https://www.patentlyapple.com/paten...round-for-3nm-plant-reveals-future-plans.html

Now those automatically assuming Nvidia will just blow away AMD may need to check if they are delusional. Until the pedal hits the metal, as in released tested projects it is just a guessing game. AMD has been on a very rapid pace staped, usually companies doing this do this with all their products - wait and see while of course debates from little second hand reports go on endlessly.
 
Back
Top