• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Intel about to expand into Graphics, AMD on the rise, are the tables turning on nVidia?

Caporegime
Joined
17 Mar 2012
Posts
50,840
Location
ARC-L1, Stanton System
I think there is a little hyperbole in the video linked below but there is some truth to it.

nVidia haven't been able to grow their partner portfolio much recently, even losing some major partners like Tesla.

Compare that to AMD who are already in every Console bar one, the nVidia / ARM powered Switch.

This year AMD have announced partnering with Google for its Stadia game streaming service who will be using AMD's GPU's.

Cray will be using entirely AMD based systems for its upcoming 1.5 ExaFlop Super Computer, which is 50% more powerful than the one Intel's building, its using Vega GPU's and Zen cores.

Samsung just announced they will be using AMD's graphics IP For Mobile Phones / Tablets

They are steadily growing their market share for Bread and Butter servers.

Add to that at least one Motherboard manufacture have said they are now selling more AMD boards than they do Intel.

One major retailer who published statistics for such things report that for more than a year they have been selling more AMD CPU's than Intel, in the last 6 moths or so 2 to 3 times as many.

Top 3 best selling CPU's on Rain Forest in the past year have been almost exclusively Ryzen, often Ryzen dominates the top row of 6.

OEM's are using more and more Ryzen products in their products.

The up-coming Ryzen 3000 looks like it will have higher IPC than Coffeelake and with 50% more cores and a fat 16 core 32 thread 'still mainstream' sitting waiting for an attempted strike back from Intel.

Navi will have better price to performance ratio to RTX but i don't think nVidia care about that, they dominate and will continue to dominate PC gaming. but out side of that they are not winning the partnerships that AMD are.

That moves us onto Intel, who will have the dominant player in the GPU space firmly in their sights, once they launch their GPU's they will start strutting around on nVidia's hallowed ground and while AMD are in a position to persuade some of nVidia's natural customers, as i have just explained, Intel have the much bigger carrot and they have a huge stick.

I don't "FEAR for nVidia" but some cracks are not just beginning to show, but are opening up with just AMD becoming more brazen than they have been able to in recent years, just think what Intel will do when they invite themselves into nVidia's home.

Also, Intel are starting to get eaten by AMD in their backyard.


FUN FUN times.
 
Last edited:
From a purely economics perspective, Nvidia can simply buy out AMD.
Its doesn't work like that , For one AMD would have to agree to a Buyout, nVidia can't just say to AMD here is some cash we own you now, if AMD don't want to sell then nVidia can go and stick it up them.

The other problem nVidia have is their current worth is $87bn and falling to AMD's $32bn and rising, its going to be extremely difficult to convince investors to stump up what will probably be near $40bn by the end of this year, half what nVidia are worth when they have lost half their value in less than a year, nVidia are not doing well, not well at all. Gigantic debts like that would kill them off.

SeEabPN.png
 
Last edited:
What part of "From a purely economics perspective," don't you understand?
Of course AMD would have to agree, and there would be huge legal hoops to jump through.


You own number simply backup my point that Nvidia is significantly larger and healthier than AMD, except you mistakenly think the change in nvidia's share's have particular relevance to the discussion. Thier share value raised far too quickly on promises of short term automotive sales as well as bitcoin hype. The market simply corrected itself. Actual economic health of Nvidia is excellent, with increasing revenue, increasing profits, increasing market share and rapidly increasing revenue in new markets.

What part of "From a purely economics perspective," don't you understand?

Completely moot then given its not going to happen...

You own number simply backup my point that Nvidia is significantly larger and healthier than AMD

Well, they are currently about twice the size in monetary value, but they are in free fall while AMD goes from strength to strength, if things keep going the way they are AMD will over take nVidia in a couple of years.

nVidia are not in a good place right now, they are not landing the big money contracts, AMD are, nVidia have lost half their monetary size in a year, AMD have doubled theirs.
 
it is amazing how much false information can appear in one post. Did you ever both to research any of the humpf you have just posted.

E.g., it would take you seconds to realize how wrong you https://marketrealist.com/2019/04/how-nvidia-increased-data-center-revenue-ninefold-in-three-years/



https://www.inverse.com/article/547...oser-as-elon-musk-says-a-i-chip-in-production


https://www.fool.com/investing/2019/03/11/amds-graphics-market-share-nears-decade-low.aspx


Is that the best you can do DG?

Collectively what nVidia have achieved in the last 3 years does nothing to counter my point that they have gained nothing of big money contracts this year, AMD have those, that's one reason why nVidia are down 50% and AMD are up 100% in this same last year.

Yes nVidia will continue to be dominant in retail gaming GPU's, they have been that for years and little has changed there, its also not important when you can't land new big money contracts outside of retail gaming, that's what AMD are doing.

As for Tesla, you have to read the whole thing, your link confirms Tesla are no longer using nVidia's GPU's.
A polite thank you note is par for course when you don't renew a large contract.
And still working with Tesla, sure, an adviser or consulting position is also par for course at least when parted from said contract on friendly terms, they still need someone to make the boss a coffee.
 
Its doesn't work like that , For one AMD would have to agree to a Buyout, nVidia can't just say to AMD here is some cash we own you now, if AMD don't want to sell then nVidia can go and stick it up them.

The other problem nVidia have is their current worth is $87bn and falling to AMD's $32bn and rising, its going to be extremely difficult to convince investors to stump up what will probably be near $40bn by the end of this year, half what nVidia are worth when they have lost half their value in less than a year, nVidia are not doing well, not well at all. Gigantic debts like that would kill them off.

SeEabPN.png

AMD now worth $34.5bn and rising....
 
Exactly this, according to some NVidia don't do well in the low end so AMD could very well be hit much harder in this segment. depending on how well received the INTEL GPU's are received of course.

Myself I think that both will take a hit as I think that NVidia actually do quite well in the low end. Of course it is easy to forget that the people on here are mainly the enthusiasts, and we don't look at the market the same way that Joe average does.

There is little value in making sub $150 Discrete Graphics, there are no margins in them, The PCB and all its components, the Heat Sync, the Fans. Nvidia only do it to keep high OEM market presence.

Both AMD and Intel are moving to on Chip Graphics and memory, AMD's 5nm APU's 'Ryzen 5000 series' are slated to have RDNA2 integrated graphics with onboard HBM, Intel already have a facsimile of that with Integrated Radeon graphics but they are large and expensive, its the cost effectiveness of them that's been holding that idea back so far but AMD at least are getting a stage where its completely viable with the size and efficiency of RDNA 2, smaller, more efficient console style APU's.

Nvidia's long tern problems remains the lack of X86 and AMD64.
 
Last edited:
BTW, Intel have a real fight on their hands with AMD in the X86_64 space right now, AMD are vastly outselling Intel in Retail and their new servers CPU's make Intel look generations old, AMD's "Rome" is 230% the performance of Intel's best Platinum Line while using less power and costing less.

AMD's top of the range 16 core 'Mainstream' waiting in the wings is already leaked, it convincingly beats Intel's soon to be released top on the line 18 core HEDT making that whole line DOA. this before AMD release their HEDT line.

Intel right now are more worried about AMD than Nvidia. an astonishing comeback and Intel look decidedly caught in the headlights.
 
Last edited:
That's not true. What should happen is that the OEMs will build the machines intel CPU + intel GPU rather than intel CPU + nvidia GPU.

Or AMD CPU + AMD GPU. you're right but you've quoted the wrong text.

There is little value in making sub $150 Discrete Graphics, there are no margins in them, The PCB and all its components, the Heat Sync, the Fans. Nvidia only do it to keep high OEM market presence.

Both AMD and Intel are moving to on Chip Graphics and memory, AMD's 5nm APU's 'Ryzen 5000 series' are slated to have RDNA2 integrated graphics with onboard HBM, Intel already have a facsimile of that with Integrated Radeon graphics but they are large and expensive, its the cost effectiveness of them that's been holding that idea back so far but AMD at least are getting a stage where its completely viable with the size and efficiency of RDNA 2, smaller, more efficient console style APU's.

Nvidia's long tern problems remains the lack of X86 and AMD64.

This is the whole point really, if Intel and AMD start making Discrete performance level integrated graphics in a small efficient singular package OEM's are going to save a lot of money not just in Discrete GPU costs but also design, manufacturing, internal space, battery life... if low to mid tier level graphics performance is all in a neat CPU sized all in one package.

Imaging in two years an RX 570 to 5700 performance level APU with 4 to 8GB of onborad video RAM.
 
Last edited:
Like this?

Useless End:
GT 710
GT 730
GT 1030

Low End:
GT 1050
GTX 1050TI
GTX 1650
GTX 1060 6GB

Mid Range:
RX 570
RX 580/590
GTX 1660
GTX 1660TI
GTX 1070
Vega 56
GTX 1070TI
Vega 64
RTX 2060
RTX 2060S
RX 5700

High End:
RTX 2070
RX 5700XT
RTX 2070S
Radeon VII
RTX 2080
RTX 2080S

Enthusiast:
RTX 2080TI
RTX Titan
 
AMD will have made their money from being commissioned to make a semi-custom design in the first place. But they also won't be making a loss selling the actual product either. Same with the new Ryzen Surface; it's the partnership and commission from Microsoft that carries the bulk of the money, not the selling of the resulting APU.

This, its an AMD design but they don't sell chips, they get IP fees and MS manufacture the chips, $100 to $120 per console sold.
 
To go back to the question in the title, as a household with 4 recent (Pascal and Turing) Nvidia cards... I really hope so.

They are great products, but competition is always good. The DGPU space has effectively been owned by Green and Red since forever, with Green really dominating in recent years. Red are fighting back, and that's awesome. With a third serious player this is only going to heat up, and maybe, just maybe, we'll see some serious price action.

Here here...
 
Intel Xe GPU Architecture Detailed – Ponte Vecchio Xe HPC Exascale GPU With 1000s of EUs, Massive HBM Memory, Rambo Cache

“At the heart of Xe architecture we have a new fabric called XEMF. It is the heart of the performance of these machines. We called it the Rambo Cache. It is a unified cache that is accessible to CPU and GPU memory.” - Raja Koduri
  • 2x density scaling vs 10nm
  • Planned intra-node optimizations
  • 4x reduction in design rules
  • EUV
  • Next-Gen Foveros & EMIB Packaging
Blue Team's First HPC GPU, The 7nm Ponte Vecchio - Landing in The Aurora Supercomputer in 2021
  • Leadership Performance (For HPC, Data Analytics, AI)
  • Unified Memory Architecture (Across CPU & GPU)
  • All-To-All Connectivity Within Node (Low Latency, High Bandwidth)
  • Unparalleled I/O Scalability Across Nodes (8 Fabric Endpoints per node, DAOS)
Xe-intel.png

https://wccftech.com/intel-ponte-vecchio-xe-hpc-gpu-detailed-1000-eus-hbm2-rambo-cache-clx/

Also known as HSA, Heterogeneous System Architecture.

I'm interested to see what happens because when AMD pioneered it no one was interested and with that it feel into obscurity.



zcVht6n.png


z2bBZzK.png


qi1yKiH.png


tPM1Qli.png
 
Last edited:
Oh... now i get it. Never mind.... i have been at my Desk too long... need a break to take my stupid head off :o
 
As has been hotly debated many times on this forum over the last couple of years Intel were mad for throwing a lifebuoy to Mr Koduri as he walked the plank during his exit from the sinking ship ISS RTG which he ran for AMD, It beggars belief that somehow a billion dollar company like Intel missed all the warning signs when they hired the fired Koduri to steer their ship into the choppy waters of the GPU market, It now looks like things are going wrong over at Intel's GPU department as they did for RTG under Koduri's reign, If there was one thing they should have learned from watching Mr Koduri's time at AMD it's that while he may be good at working on a gpu's architecture when it comes to heading up a division he's completely incompetent. Let's hope that this time the captain goes down with his ship.



I agree with AdoerdTV on Koduri completely, Under his premiership AMD's GPU's tried to do everything and because of that ended up doing nothing well and with high power consumption, exactly like the complaints Intel insiders now apparently have about their new GPU's, Intel's GPU's are Vega '''''before Lisa reformulated them for mobile and increased the gaming IPC by 50%.

And yes, Intel are arrogant and incompetent.
 
Which Vega failures? Vega was an absolute compute beast, not only the MI cards but even the humble V64/56. For 1/5th the money had 80% the performance of Volta based cards on computing.

Gaming performance wise was in par to GTX1080 in DX11 and outperformed the GTX1080 on DX12 and Vulkan. That happens even today on benchmarks using current drivers. And we talking about the Reference Vega 64. Not some better cooled ones like the Nitro which is 20% faster out of the box.

Sure you might say all these are irrelevant today since the 5600XT Red Dragon at $280 beats both the above.



A presense at which segment? The Intel gGPU is just faster than the Ryzen 4800U IGP, which is 15W laptop APU!!!!!!!
They do not dare to pit it against the 4800H laptop CPU!!!!! :D

It was a compute monster, but for its size and 300 watts it should have been.

https://www.techpowerup.com/review/amd-radeon-rx-vega-64/29.html

4096 Shaders and in gaming it was no faster than a much smaller 180 watt 2560 shader GTX 1080.

Vega was reformulated after Kudori left, in Renoir Mobile APU's its 59% faster per CU than how Kudori left it. 59% its like Ryzen 1000 vs Bulldozer.

https://www.notebookcheck.net/This-...-CU-despite-still-being-on-Vega.449849.0.html

Imaging how fast Vega 64 would have been if Lisa, not Kudori had been in charge.

PS: i can't find it now but there is an early benchmark of AMD's Renoir Mobile APU iGPU performance comparison to Intel's new iGPU and on the same TDP AMD's is faster. Significantly faster.
 
Back
Top Bottom