Discussion RDNA 5 / UDNA (CDNA Next) speculation

[adroc_thurston]

Well the AT article is long gone

i hate Future PLC so much

 

[Heartbreaker]

What fans? DF is as far away from "AMD fan" as humanly possible.

I assume the "we" DF meant is the community at large, not just themselves.

 

[adroc_thurston]

I assume the "we" DF meant is the community at large, not just themselves.

The community has no idea what to expect from gfx13 whatsoever.

 

[Saylick]

It's not that they hate AMD; it's just that them (especially Battaglia) love NV so much they just do infinite NV shill pieces every (even the smallest) product launch.

Ultimately, DF doesn't want to bite the hand that feeds them, which does make them a useful tool for Nvidia to disseminate marketing talking points when the proper time comes.

 

[gdansk]

Believe it or not DF does less Nvidia-glazing than gamers. DLSS has what gamers want.

 

[blackangus]

Digital Foundry's take on is RDNA 5 being overhyped

One of the best signs is that major console platforms have promised to use RDNA5 for their next-generation hardware. That choice shows a lot of faith in the underlying infrastructure and is a big commitment. From our point of view, this alone suggests that RDNA 5 will be able to handle tough long-term needs.​

AMD RDNA 5 Architecture: What the Current Information Really Suggests | NoobFeed

There has been talk about whether AMD's RDNA5 design is being praised too much. People are often afraid of what might happen if new AMD GPUs get too much att...

www.noobfeed.com

I mean I think historically the answer is clearly yes. So I dont see it being any different this time.
The community always dreams bigger than reality =)

 

[ETI4711]

AMD earning call:

From a product standpoint, Valve is on track to begin shipping its AMD-powered Steam Machine early this year, and development of Microsoft’s next-gen Xbox featuring an AMD semi-custom SoC is progressing well to support a launch in 2027.

 

[Z O X]

I mean I think historically the answer is clearly yes. So I dont see it being any different this time.
The community always dreams bigger than reality =)

True, but dreams are based on reality: 9800/4870/5870/7970.
We all hope AMD performs a miracle since this nvidia dominated era is a hell for gamers ...

 

[adroc_thurston]

We all hope AMD performs a miracle since this nvidia dominated era is a hell for gamers ...

Lisa hates the gamers even more.

 

[marees]

Lisa hates the gamers even more.

ps4 saved amd from bankruptcy

 

[adroc_thurston]

ps4 saved amd from bankruptcy

Not really, no.
It did keep GPU people employed for sure.

 

[Z O X]

Lisa hates the gamers even more.

Saw you make this statement several times, but didn't find a good explanation.
She really seemed genuine from few public appearances that I saw.
But if you mean that not going all-in each generation is a bad strategy, that is IMHO not true.
Sorry, enough derailing from me ...

 

[adroc_thurston]

Saw you make this statement several times, but didn't find a good explanation.

Client gfx programs are essentially gambling and Lisa hates gambling.

She really seemed genuine from few public appearances that I saw.

She's a very cool CEO.

But if you mean that not going all-in each generation is a bad strategy, that is IMHO not true.

No, it's the only strategy. Either you win, or you don't.
Problem is, winning is expensive and who knows what the market impact will be.

 

[Thunder 57]

Lisa hates the gamers even more.

I know you like to make controversial statements but that one is just wrong. Nvidia is giving gamers the middle finger by postpoing or outright canceling the Super refresh and if the rumored production cuts are true. Have you seen 5070 Ti prices?

The worst you could say about Lisa Su is that she is apathetic towards gaming. Outright hating them? That's just making a statement to get a reaction. And look, it worked.

 

[adroc_thurston]

Outright hating them?

She's the reason Radeon guys can never truly compete against NV.

 

[Win2012R2]

Lisa hates the gamers even more.

Liar.

 

[adroc_thurston]

Liar.

Denial's not gonna change anything.

 

[Joe NYC]

Client gfx programs are essentially gambling and Lisa hates gambling.

Which is one of the reason AMD (Lisa) are not good at leading. Just good at following, when they realize "Oh shhht... we missed another trend / market, we better catch up", after others already validated it, and made killing on it.

She's a very cool CEO.

No, it's the only strategy. Either you win, or you don't.
Problem is, winning is expensive and who knows what the market impact will be.

Others are making big bets on markets they are not even in.
AMD is not investing enough in the same markets even when AMD is already in those markets.

There was a good excuse for it when AMD did not have the resources, but it is less and less excusable when the resources are there for decent product / refresh cycle, but only sub-par effort is invested in / executed.

Even in a rare occasion when AMD introduces a leadership product, such as Strix Halo, AMD still managed to drop the ball, not push it hard enough and now AMD is just standing around, waiting for others to catch up and overtake AMD, before they join the race as a follower, rather than a leader.

If survival is the lowest level of success, Lisa mastered it fine.
Then comes prosperity, and Lisa is doing fine there as well there.
But leadership, such as industry leadership - that's where she falls short.

 

[adroc_thurston]

Which is one of the reason AMD (Lisa) are not good at leading. Just good at following, when they realize "Oh shhht... we missed another trend / market, we better catch up, aft3er others already validated it, and made killing on it."

They rule server CPUs with an iron fist.

Others are making big bets on markets they are not even it.

Like what.

but it is less and less excusable when the resources are there for decent product / refresh cycle, but only sub-par effort is invested in / executed.

They make the best server CPUs by a country mile.
(You) being too poor for them is a (You) problem.

Even in a rare occasion when AMD introduces a leadership product, such as Strix Halo, AMD still managed to drop the ball, not push it hard enough and now AMD is just standing around, waiting for others to catch up and overtake AMD, before they join the race as a follower, rather than a leader.

Push what, where?
Do you even understand how PC as a market works?

 

[marees]

Which is one of the reason AMD (Lisa) are not good at leading. Just good at following, when they realize "Oh shhht... we missed another trend / market, we better catch up", after others already validated it, and made killing on it.



Others are making big bets on markets they are not even in.
AMD is not investing enough in the same markets even when AMD is already in those markets.

There was a good excuse for it when AMD did not have the resources, but it is less and less excusable when the resources are there for decent product / refresh cycle, but only sub-par effort is invested in / executed.

Even in a rare occasion when AMD introduces a leadership product, such as Strix Halo, AMD still managed to drop the ball, not push it hard enough and now AMD is just standing around, waiting for others to catch up and overtake AMD, before they join the race as a follower, rather than a leader.

If survival is the lowest level of success, Lisa mastered it fine.
Then comes prosperity, and Lisa is doing fine there as well there.
But leadership, such as industry leadership - that's where she falls short.

I was thinking that going for profit rather than users is like going impressions/views rather than unique users

I would look at increasing the overall customer base (unique users) & not just profitability & share value. s***w share holders

 

[Joe NYC]

They rule server CPUs with an iron fist.

Because there was not a great deal of leadership needed to just execute on many "threads going brrr"

It was a known problem to be solved. Not a new use case.

Like what.

Like Qualcomm, Nvidia investing in laptop SoCs, AMD not investing, falling behind.

NVidia investing in gaming GPUs and leveraging them to get into workstation GPU, crypto, AI. AMD missing out, falling behind.

Intel investing in decent iGPU, AMD standing around, falling behind.

The 3 year cycle on Strix Halo blew AMD out of the water, and by the time the next iteration ships, AMD slipped from 1st place to 5th place, behind:
- Intel
- NVidia
- Apple
- Qualcomm

Push what, where?
Do you even understand how PC as a market works?

When Qualcomm manages to push their product harder than AMD, you can tell that AMD is not doing enough.

Not doing enough also opens the doors to new entrants, such as Qualcomm and NVidia, even while they have a huge disadvantage of swimming upstream, lacking x86 compatibility.

 

[adroc_thurston]

Because there was not a great deal of leadership needed to just execute on many "threads going brrr"

It was a known problem to be solved. Not a new use case.

You really, really need to stop talking stuff shit outta your depth.
Server multiprocessors are some of the hardest things to design and build

Like Qualcomm, Nvidia investing in laptop SoCs, AMD not investing, falling behind.

NVidia investing in gaming GPUs and leveraging them to get into workstation GPU, crypto, AI. AMD missing out, falling behind.

Intel investing in decent iGPU, AMD standing around, falling behind.

The 3 year cycle on Strix Halo blew AMD out of the water, and by the time the next iteration ships, AMD slipped from 1st place to 5th place, behind:
- Intel
- NVidia
- Apple
- Qualcomm

Ah, you're buttmad AMD is not giving you treats.
Sorry pal, it's a server kingdom.
They don't care about you. They will never care about you.
You will get server scraps and you WILL enjoy them.

When Qualcomm manages to push their product harder than AMD, you can tell that AMD is not doing enough.

?
We're in Year 8 of QCOM fumbling the bag.

 

[Joe NYC]

I was thinking that going for profit rather than users is like going impressions/views rather than unique users

I would look at increasing the overall customer base (unique users) & not just profitability & share value. s***w share holders

Just having a quality product in the market place can achieve both.

If you don't, you keep introducing new [products, until you do. That's how Intel managed to go from Raptor Lake - Meteor Lake -> Lunar Lake -> Arrow Lake -> Panther Lake while AMD introduced a Raptor Lake competitor and went on to take a multi-year nap.

And when you have a mis-step, such as RDNA3, you try to recover from it. Instead, AMD is going to be on this unfortunate generation until ~2029 in APU.

This, while AMD has the tools (RDNA4, RDNA5) and talent in place to correct it, but does not.

 

[adroc_thurston]

Just having a quality product in the market place can achieve both.

If you don't, you keep introducing new [products, until you do. That's how Intel managed to go from Raptor Lake -> Lunar Lake -> Arrow Lake -> Panther Lake while AMD introduced a Raptor Lake competitor and went on to take a multi-year nap.

And when you have a mis-step, such as RDNA3, you try to recover from it. Instead, AMD is going to be on this unfortunate generation until ~2029 in APU.

This, while AMD has the tools (RDNA4, RDNA5) and talent in place to correct it, but does not.

I get it, you're buttmad about client iGPUs.
It's okay pal, you're free to buy Intel.
Oh and it's 3.5 until 2031.

 

[Tachyonism]

This, while AMD has the tools (RDNA4, RDNA5) and talent in place to correct it, but does not.

Wasnt there something about RDNA4 not fitted to be used as iGpu. Iirc it was something along the line of RDNA4 was an architecture designed to push the limit and doesnt really perform well at lower wattage.