evernessinceWell I hope it's not just baby Vega (Navi should have improvements of it's own right?). If we assume AMD makes each die 1/3 the size of Vega, that makes each die 170mm2, which is quite a bit smaller than the RX 480 and even Ryzen's 213mm die size. Yields on such a die would be insanely high. It also places their basic 4 chip module (just like zen) above Vega 64 performance wise by 33% assuming there are no frequency improvements. What will likely happen though is Navi will bring some performance per watt improvements and AMD will target the sweet spot frequency, whatever that may be for Navi. They will likely also sell lower end products with a single die disabled. We know AMD can do an 8 Die solution but wouldn't that just be overkill? 266% of the current Vega 64 performance? That's enough power for 8K gaming or a very powerful rending chip. Just like Epic, AMD may reserve these chips for professional use.

There's a reason Nvidia also stated that MCM is the future and why they are also currently researching it. It's going to allow the continuation of moore's law for a little while longer.

The biggest assumption you made with that is perfect linear scaling. If it relys on driver support you can unfortunately cut that performance in half.

If these news are true, AMD should not let Asus, MSI and Gigabyte sell their motherboards

XaledIf these news are true, AMD should not let Asus, MSI and Gigabyte sell their motherboards

If only it was that simple.

TotallySorry to interject but he is correct when he said that they we're stuck with a chip that didn't scale. I don't know if you forgot or missed this key point of AMD's small chip strat. What the we're shooting for was a small chip that they depending on performance target package many dies on the same chip, at the time I imagined it to be something like crossfire on a chip, not too different from what they are doing with Ryzen. Then when Polaris finally rolled around it was too power hungry for them to realize such an ambition. So they we're technically stuck with a chip that didn't 'scale.' I'm certain that it is this effort that led to MCMs it is just that they are successful with implementing it on the CPU front. The original article can be found somewhere on Anandtech a piece about the Fury/7970 being the last monolithic die for AMD and their new focus we're smaller dies.

Um, did you even read the post your qouted. I specifically state that AMD had a small die size in mind from the start. Please read posts before commenting on them.

Vayra86- Vega perf/watt is fine but the overall performance per mm2 of die area is not. So what we get is a high clock to extract equal performance to what the competitor can do with a smaller die. And with that high clock, Vega perf/watt sucks. This is what happened with Vega 64, and a similar Vega 64 with GDDR5 would simply not have been possible. The bottom line is that Vega as a high end part simply isn't enough, no matter how you twist it or tweak it. Also, you handily omit the fact that the 'lower clocked efficient Vega' is also a silicon lottery. No guarantees.

- Nvidia's way of scaling their architecture is indeed great and what you are saying about how big a chip can be is exactly the problem GCN met during Hawaii and its the problem that Vega didn't manage to solve either. Meanwhile, there is a 35% performance gap at similar die sizes so now AMD has to resort to multiple dies, while Nvidia can postpone that for another full generation if they want to.

Or, perhaps we might even see a much bigger die area like what AMD did with Threadripper. Especially with HBM, the board has lots of space anyway. So there are tons of options, but ONLY if you have a highly efficient architecture that doesn't surpass TDP budgets for each segment. People simply will not accept a 400W GPU in this day and age, just as they won't accept hot and loud ones, let alone move all that hot air out of the ever smaller form factors on the marketplace. So efficiency is king for every single use case.

- You compare frequency limitations, but back when Maxwell was released, did you for one second consider the next gen would pass the 2 Ghz barrier for Nvidia? I for sure as hell didn't.

- I am not that worried about competition in the GPU space. I will say this, as I have said often; I think RTG would be much better served in the hands of a different company that can truly focus on its GPU effort instead of the happy marriage that is APU / custom chip design, because let's face it, for a true gaming GPU, those are all the wrong priorities and it shows. AMD is not doing anyone a service for the past few years and there is nothing on the horizon that is ready to dethrone Nvidia. Its up to AMD/RTG's priorities and management that we now have a high end abandoned for nearly two years...

Yeah, the HBCC takes up allot of space on the die and it's a waste of space for gaming applications. I really don't get the point of it even being on gaming cards other than not having to make yet another SKU to save cash. That has to be it, after all AMD was pretty much out of money when they were making Vega. Lower clocked efficient Vega isn't silicon lottery, every Vega card has a sweet spot where you can trim significant power costs off the card. AMD pushed Vega 64 out the door to compete, they did not fine tune the BIOS to extract the best performance per watt. This is the only reason why so many Vega owners were able to shave so much power consumption by lowering voltage. This isn't the first time AMD has done this. They did it with the 290X as well to a lesser extent, although Vega is their more egregious offense.

Well I'd be fine with a 400W GPU if it provided performance worthy of that power consumption. AMD's MCM design has a much larger heat dissipation area, it is much easier to cool than a single big die. I'm pretty sure Nvidia is the one that originally proved that while noise and efficiency is great, performance is king. People bought Fermi by the bucketload despite it being much less efficient and noisy.

Pascal only achieves 2GHz or more through overclocking. The base clock for the 1080 Ti is 1480, which is only a moderate bump over the 980 Ti.

www.amazon.com/EVGA-GeForce-FOUNDERS-Support-11G-P4-6390-KR/dp/B06XH2P8DD

I am lucky enough to own a 1080 Ti SSC that goes over 2GHz but that's only because I won the silicon lottery. It's not really fair to compare overclocking results because they rely heavily on luck sometimes.

Well Rory Read (prior AMD CEO) was the one who came up with the APU idea and then Raja was the one who came up with Vega. If it wasn't for the APU, AMD would have never have landed the console deals and likely would have went bankrupt. Vega isn't good at gaming and is obviously targeted at professionals. It would have been great had ATI been bought by someone with more money but then again I don't know if AMD would have even been around today otherwise.

cdawallThe biggest assumption you made with that is perfect linear scaling. If it relys on driver support you can unfortunately cut that performance in half.

Infinitey fabric doesn't rely on drivers to work. Even still, we know that Infinity Fabric does have a very minor pentalty when transferring data from die to die. This can be mitigated through drivers and doesn't have a huge impact on performance.

So they are using "AREZ" for AMD, while "ARES" will hold to Nvidia?

If that's the case it the lack imagination from the people who was suppose think up this "branding" must be total dim-wads (yes I know the younger generation doesn't use the term and I'm showing my age) that went to college and sat around with their finger up somewhere.

I could've given 25 other good ways differentiating the brands rather than just what appear to be a miss-spelling.

Not sure people have settled on the spelling.

Digitimes: Asustek to release Ares brand for AMD gaming products, says paper

DigitimesIn response to Nvidia's GeForce Partner Program (GPP), Asustek Computer has planned a new gaming brand series, Ares, to market its AMD-based gaming products, while its Republic of Gamers (ROG) will be dedicated to machines with Nvidia-based hardware, according to a Chinese-language Apple Daily report, which said the vendor will start implement the new brand plan in mid-April.
Currently, AMD's and Nvidia's shares in the discrete graphics card market are around 30% and 70%, respectively, with the former's shipments picking up in the past few years and the latter hoping GPP will help prevent AMD's advance, the paper noted.
Micro-Star International (MSI) and Gigabyte Technology so far have not yet made firm plans for Nvidia's GPP, the paper added.

CasecutterSo they are using "AREZ" for AMD, while "ARES" will hold to Nvidia?

If that's the case it the lack imagination from the people who was suppose think up this "branding" must be total dim-wads (yes I know the younger generation doesn't use the term and I'm showing my age) that went to college and sat around with their finger up somewhere.

I could've given 25 other good ways differentiating the brands rather than just what appear to be a miss-spelling.

I don't remember any Ares video card with nvidia gpu. Corresponding nvidia would be Mars, which were all dual gpus(remember this Pos).

jabbadapI don't remember any Ares video card with nvidia gpu. Corresponding nvidia would be Mars, which were all dual gpus(remember this Pos).

The original ROG gpus were Nvidia under the Sub-brand Matrix

XzibitThe original ROG gpus were Nvidia under the Sub-brand Matrix

Well, look at that puny little 9600gt. But ain't Ares cards have been all dual gpus as is mars?

Edit: for the spelling. just look at the how ares have been wrote by Asus. That letter E is Sigma(Σ) from Greek alphabet, someone might confuse it with latin Z.

jabbadapWell, look at that puny little 9600gt. But ain't Ares cards have been all dual gpus as is mars?

I believe so

Ares = Dual chip AMD
Mars = Dual chip Nvidia

Asus also used Matrix sub-brand on AMD cards too. Just recently it stopped using Matrix for both (980 Ti & 290X).

XzibitI believe so

Ares = Dual chip AMD
Mars = Dual chip Nvidia

Asus also used Matrix sub-brand on AMD cards too. Just recently it stopped using Matrix for both (980 Ti & 290X).

Speaking of sub-brands were there any AMD Poseidons. Quick googling just gives four Nvidia cards, so I guess not.

You can read more here
www.techspot.com/news/73661-nvidia-gets-anti-competitive-unsavory-geforce-partner-program.html

Reason why i sold my two watercooled zotacs 1080 long ago, and went full tilt AMD (32 core epyc, and 2 frontier edition watercooled vegas)
I knew nvidia is up to no good.

The multi-billion dollar AMD corp thank you for your support.

jabbadapI don't remember any Ares video card with nvidia gpu. Corresponding nvidia would be Mars, which were all dual gpus(remember this Pos).

I miss the time period when nvidia allowed freedom to glue two midrange GPU's together to an AIB. Call that whatever you want, but Asus was able to sell two 760's glued together no issues, now they can't even offer a 1070Ti with clockspeeds over default.

evernessinceUm, did you even read the post your qouted. I specifically state that AMD had a small die size in mind from the start. Please read posts before commenting on them.

I did read that but I was pointing out that you were incorrect in saying bug was incorrect when he said to that their chip didn't scale up when in a matter of fact that was what AMD was aiming for and the entire purpose their small chip strat but the chip was to power hungry to scale up so they were stuck with a small chip. I know I probably didn't make that point clear enough but it was 3am and care to edit the post.

hmm... does Zotac & Galax signed up for the controversial GPP? If so, not gonna be surprised. bigger names like ASUS is gonna get a big no-no from me. I only buy their motherboards, NOT the GPUs coz it's just too friggin' expensive.

Tsukiyomi91hmm... does Zotac & Galax signed up for the controversial GPP? If so, not gonna be surprised. bigger names like ASUS is gonna get a big no-no from me. I only buy their motherboards, NOT the GPUs coz it's just too friggin' expensive.

Asus GPU's are no more expensive than anyone else...

I truly don't see what the big deal is about. Asus calls their nVidia cards Asus Broccoli and call their AMD cards Asus Asparagus ... so what ? MSi calls their nVidia cards MSI Cheetah / their AMD cards AMD Jaguar ... so what ? Gigglebyte calls their nVidia cards Europa / AMD cards Ganymede, and a future Intel card Callisto ... who cares ? What's the concern ? I can't imagine why they were not different in the 1st place... In fact Im kinda surprised that Intel didn't call the layers when AMD did the X370 after Intel's Z370.

Is there an issue that if you are an AMD fan the nVidia name might sound cooler or visa versa ? Are you worried what your friend's will think ? Wher's the beef here ? Personally I want different names so as to easily be able to distinguish between model lines. If the worry is that one name might not sound as impressive as the competitors ? Well at the upper end, the card's aren't as impressive but it's not th ename that's responsible for that it's the power usage and performance. That's where the judgment's should be made.

John NaylorI truly don't see what the big deal is about. Asus calls their nVidia cards Asus Broccoli and call their AMD cards Asus Asparagus ... so what ? MSi calls their nVidia cards MSI Cheetah / their AMD cards AMD Jaguar ... so what ? Gigglebyte calls their nVidia cards Europa / AMD cards Ganymede, and a future Intel card Callisto ... who cares ? What's the concern ? I can't imagine why they were not different in the 1st place... In fact Im kinda surprised that Intel didn't call the layers when AMD did the X370 after Intel's Z370.

Is there an issue that if you are an AMD fan the nVidia name might sound cooler or visa versa ? Are you worried what your friend's will think ? Wher's the beef here ? Personally I want different names so as to easily be able to distinguish between model lines. If the worry is that one name might not sound as impressive as the competitors ? Well at the upper end, the card's aren't as impressive but it's not th ename that's responsible for that it's the power usage and performance. That's where the judgment's should be made.

That is not the issue...the issue is Nvidia wants claim for the already established branding of these companies to only be used for them.

bugYes, because initially Vega was launched for mobiles. Only then we got the compute and AI cards with gaming parts to follow much, much later.
It's really funny (and worrisome at the same time) to watch people making stuff up and rewriting history rather than admit they're sometimes wrong.

For bonus points, here's how Vega competes with Pascal irl: www.phoronix.com/scan.php?page=article&item=12-opencl-98&num=1

They launched on gaming first because that would be the lowest price point, and frankly this was the sector AMD was neglecting the most. Duh.

What's really worrisome is how oblivious so many people (like you) around here are to how the market works. Like you didn't even bother to address the facts I mentioned because it would mean changing your opinion.

1. FACT: A 14nm 471mm^2 Vega card has about the same level of compute as an 815mm^2 12nm Volta card.
2. FACT: Low-powered Vega gpu's outperform M150X Nvidia cards while using less energy (While having the same die size!).

So are you going to dispute the two facts above? I am very excited for the mental gymnastics to come, but maybe you will surprise me. We can always hope ;)

John NaylorI truly don't see what the big deal is about. Asus calls their nVidia cards Asus Broccoli and call their AMD cards Asus Asparagus ... so what ? MSi calls their nVidia cards MSI Cheetah / their AMD cards AMD Jaguar ... so what ? Gigglebyte calls their nVidia cards Europa / AMD cards Ganymede, and a future Intel card Callisto ... who cares ? What's the concern ?

And this is why you are not a lawyer, and hopefully not in any positions of power.

cdawallI miss the time period when nvidia allowed freedom to glue two midrange GPU's together to an AIB. Call that whatever you want, but Asus was able to sell two 760's glued together no issues, now they can't even offer a 1070Ti with clockspeeds over default.

Haha this actually a pretty funny point. Tells you how much more competitive Maxwell was compared to Kepler...

Just call them series RadeonZ en call it a day.

ASUS just lost by business.
This GPP by Nvidia is causing more damage by the day.

evernessinceIt makes zero sense for ASUS to create this cheap sounding brand instead of using the ARES brand they already have unless they are taking all the gaming brands. The GPP seems to be going beyond even what was previously reported. Nvidia is going to have all the main brands and AMD is going to be relegated to cheap knock-off brand names.

I hope Nvidia is sued for this program and fast.

I hope so too, Nvidia cannot get away with this BS.

Fret not AMD fans, Navi will win out in the end!

HYPE MODE ON

wccftech.com/rumor-amd-navi-mainstream-gpu-to-have-gtx-1080-class-performance/

the54thvoidIf AMD make a fantastic gfx card, and it has an awesome cooler and still carries the 'Frozr' name, it won't make it less appealing.

In other words "marketing doesn't work". You can't be serious.
It won't work on tech savvy users, but that's a minority of customers.

FluffmeisterNavi will win out in the end!

Win out what? #downtripoffyourcustomers award?
AMD is winning that annually since forever.

medi01Win out what? #downtripoffyourcustomers award?
AMD is winning that annually since forever.

Yes, down trip off your customers.