AMD Retreating from Enthusiast Graphics Segment with RDNA4?

[R0H1T]

Because Nintendo doesn't sell games based on their visual quality.

 

[HD64G]

My thought is that they have found hard for drivers to get the maximum performance out of the new arch and they will delay its launch until N41 and drivers balance is good enough. I don't think they are abandoning that. Just a thought though. Time will tell.

 

[persondb]

My thought is that they have found hard for drivers to get the maximum performance out of the new arch and they will delay its launch until N41 and drivers balance is good enough. I don't think they are abandoning that. Just a thought though. Time will tell.

Their new arch(RDNA 3) is basically broken. The only performance uplift they got is from higher clocks, while the architecture improvements seem to have been almost nothing(likely because of broken features...).

RDNA 3 is really a repeat of Vega, where there was a lot of marketing about feature that would revolutionize everything(packed math, hbcc, primitive shaders, etc etc) and they were really, never used or not even enable in the drivers.

 

[ToTTenTranz]

I think this is what people miss, because it's constantly repeated that "AMD shouldnjust sell GPUs cheaply to get marketshare". How do you explain that to investors?

An investor who isn't capable of recognizing what it takes for a company to increase their 12% marketshare shouldn't be an investor.

If that's true, then why Nintendo choose Nvidia?

Can you explain me that?

Because Nvidia had warehouses full of those Tegra chips for tablets that had failed miserably against their Snapdragon and Apple competition, which they were willing to sell for cheap.
Tegra X1 was such a failure that Nvidia never even tried to enter the Android market again.

 

[R0H1T]

A common excuse, but a flawed argument. AMD had investors who backed them during their dark Bulldozer days.

Not really, not sure if any of the Gulf states(sovereign funds?) were major investors at that time or not but GF was pretty much all AMD ~ if they went belly up Mubadala would lose a heck of a lot more! It ain't as cut & dry as you're making it out to be.

 

[thurnis.silva]

i think they scared of intel

 

[thunderingroar]

AMD attempted to pivot from being a high-volume low-margin producer with Zen 1/2/3 and RDNA 1/2, to a low-volume high-margin one with Zen 4

Thats entirely false?? If anything zen3 was the high margin high volume product. Are people so quick to forget all the community outrage when AMD raised prices on zen3 (built on same 7nm+12nm as zen2) while also deciding to exclude the box coolers?

Meanwhile Zen 4 launched with 50$ lower MSRP on 7700x (5800x MSRP was 450$) and 100$ lower MSRP on 7950x (5950x MSRP was 800$) and the non X parts now also include box coolers. Currently ryzen 7000 prices are very reasonable.

Your claims on zen4 being "high margin" are even less true once you realize that Zen3 was still using that dogshit dirt cheap Global Foundry 12nm for IO die which made all of their cpus have awful idle power draw compared to intel. Now zen4 uses much better and more efficient (also a lot more expensive) TSMC 6nm IO die. May i also remind you that IO die is actually bigger than the core complex die and it heavily factors in the total BoM cost

Its crazy how you speak with such confidence about company's startegy, profit margins, leadership, greed and even future "bankruptcy pies", when the underlying facts that formed your opinion are untrue.

Intel who wised up and decided to compete on value (retaining DDR4 support was a masterstroke)

I'd argue that zen4 launch timing into ddr5 market when the memory prices started to plummet is more impressive, especially considering AM5 should last to 2026 and people will reuse their memory kits

And while both Intel and NVIDIA have other, massive, sources of revenue to fall back on if their particular marketing strategies fail... AMD simply does not.

Server market? Consoles and handhelds?

 

[TheinsanegamerN]

This move worked with RDNA 1, I don't see why it couldn't work now.

It didnt though. It didnt work with GCN either.

When AMD abandoned the high end market for GPUs with GCN, this gave nvidia's maxwell the top tier on a silver platter. That immediately led to GCN 4.0, AKA polaris, where AMD didnt have anything above the low tier. The result was nvidia's 1070 making more money per unit then any GPU, and outselling all of AMD COMBINED. As did the 1080. And the 1060 sold 4x as all of polaris. rDNA1 had a decent run, with AMD fixing drivers and fixing their reputation, but at the same time high end buyers had no options. Anyone who had a vega card had no upgrade path, meanwhile the 1080ti/2080/2080ti sat right there with a whole market to themselves.

All of this funneled money into nvidia, allowing them to pull further ahead with things like RT. The last thing AMD needs to do now is abandon the high yield cards. They need to get a competent stack of cards with a true halo product out there, like nvidia has consistently done for 15 years. The first time AMD's marketshare increased since 2014 was with rDNA2, the first full stack since 2014.

Or just sell RTG to intel with a clause that they can continue to license the GPU IP for their APUs in the future. Or maybe samsung, they would have their own fabs.

 

[R0H1T]

AMD's shipping more (high end) x86 CPU's than ever so any claim that it's low volume is laughable at best.

Sure most of them probably end up in servers but again not low volume!

 

[bug]

Not really, not sure if any of the Gulf states(sovereign funds?) were major investors at that time or not but GF was pretty much all AMD ~ if they went belly up Mubadala would lose a heck of a lot more! It ain't as cut & dry as you're making it out to be.

You're just muddying the waters. The assertion/allegation was that investors will pack up shop after one disappointing quarter. I'm just saying short term investors aren't the only ones around, long term investors also exist. Seems pretty cut&dry to me.

 

[TheinsanegamerN]

I think you're wrong there, Intel won't end the duopoly, they'll just switch places with AMD and leave it completely intact....Nvidia seemingly has an unbreakable hold over the consumer base and unfortunately, it cannot simply be broken with an empirically better product and here's why:

words words words words

When, exactly, was the last time AMD offered that? Between 2014 and 2018 AMD couldnt come close to the performance of nvidia's cards, only able to compete via lower prices and crap drivers. 2019 saw the 5700xt, which offered comparable raster performance and fixed drivers, but had no RT compared to the year old RTX 2000 series and no raster answer to the 2070 super/2080/2080super/2080ti.

RTX 3000 once again had superior RT performance. rDNA2 had a niche for itself, it was superior in raster at 1080p and 1440p, and had larger framebuffers. rDNA2 also sold very well, held back by AMD themselves whom prioritized ryzen production, meaning that for over a year you couldnt physically buy a rDNA2 card. I know, I was one on them trying to find one.

Then with rDNA3 we're back to not competing with the top dog, with now vastly inferior RT performance, propelled on by an arch that seemingly bring little improvement clock for clock to rDNA2. AMD themselves have struggled with it. We, once gain, have to compete on price.

Sure seems to me that for over a decade AMD has played second fiddle to nvidia, which worked fine until AMD decided it wanted to be a "PrEeEeMiUm BrAnD" and jacked up prices. Everyone bashed nvidia for having a flat perf/$ increase over the last gen, but AMD did the exact same thing with the 6700xt and 6600xt.

If AMD couldnt actually make a superior product, like they did with the HD 5870, or the 7970, or the 290x, they would have no problem selling cards. But they havent really done that. They have a good card, like the RX 480 8GB, and that card does well, but the whole stack doesnt.

 

[TheoneandonlyMrK]

Another read of this is that the high end radeon will be built from multiple smaller dies.

Eventually, they will.

What is high end next generation..



Ask Nvidia and it would be a 5060ti on price, a 5070 on specs.

Ask Intel and it will be A870

Ask AMD and they're not bothering.

Interesting times ahead.

I personally think they'll work on multi die GPU by perfecting that GCD die FIRST.

All while hopefully releasing decent APUs for entry level.

And having very viable low to mid entrant GPUs.

Shit times ahead though AI is spanking wallets all round.

 

[jesdals]

Well it would be a shame - we need more competition in the PC GPU segment not less

 

[Chomiq]

Presumable they mean the Steam Deck and ROG Ally, which together have increased the Linux marketshare from 2% to 3%. Meanwhile everyone else is making do with their GTX 1060 from 7 years ago, or they've just bought a PS5 instead.

And at the same time they chose to ignore Tegra being present in Switch for years now.

 

[TheinsanegamerN]

And at the same time they chose to ignore Tegra being present in Switch for years now.

The tegra is a completely different product from ryzen APUs tho?

 

[Tek-Check]

A nasty summer tech gossip... due to lack of other important topics to talk about.

 

[R0H1T]

You're just muddying the waters. The assertion/allegation was that investors will pack up shop after one disappointing quarter. I'm just saying short term investors aren't the only ones around, long term investors also exist. Seems pretty cut&dry to me.

Two things ~ while long term investors most probably stayed with them through thick & thin but also the likes of Mubadala & GF would've lost billions had AMD gone under. Now how much influence they had in AMD staying afloat in that span no one really knows, but you can bet people would've lost a lot more if AMD went under specially GF & anyone depending on AMD's business.

 

[Chomiq]

The tegra is a completely different product from ryzen APUs tho?

But with 122+ million Switch units sold it IS dominating the handheld gaming market.

 

[TheoneandonlyMrK]

This threads kinda funny.

AMD high end thread, all debate APU.

AMD low end thread, the high end is shit, 4090 rule's.

 

[pressing on]

But with 122+ million Switch units sold it IS dominating the handheld gaming market.

It's not just a case of unit sales, it also the revenue that it brings.

For example the graphic below shows TechEpiphany's tweet on the latest (week 31) GPU sales at Mindfactory, Germany.

So AMD achieves 51.56% in unit sales. But the AMD ASP(Average Selling Price) is lower. As a result AMD gets only 41.63% of the revenue, with NVIDIA on 58.25%. So unit sales (and % market share numbers based on them) are not the whole story.

 

[Denver]

You don't have contact with AMD to at least try to confirm this information?

 

[xrli]

Considering the existence of the MI250 and MI300 family, I have a feeling that the next Radeon flagship could be 2 smaller dies linked together by infinity fabric. But GPU MCM are not as simple as CPU MCM and the cost saving may not scale very well.

 

[Vayra86]

Even if true that doesn't appear to be happening until RDNA5, so 3 years at least and I wouldn't bet on there ever being a RDNA5. AMD will struggle next year against Intel's Battlemage IMO and drop to third in a few years. They won't even bother trying to compete other than with APU's.

Right, so Intel is going for that fabled 250~300% performance boost by 2025? Noice! And even then they're just looking at AMD's last gen offering pretty much.

I'm in popcorn/wait and see mode here in every possible way. Just glad I bought a 7900XT as it is because the coming gens aren't looking to get much better and the lower you go in current gen stacks, the worse it gets. That still strikes me as a novelty many fail to recognize. High end purchases are equally effective cost/frame, and therefore better as they resell much more easily too and hold value.

What's not a novelty here though is AMD's 'consistency' wrt their GPU updates. Mother of god, what a mess.

 

[pressing on]

You don't have contact with AMD to at least try to confirm this information?

TechEpiphany as far as I know is releasing the figures supplied to him by Mindfactory. They reflect the local market for that retailer in Germany that may not be representative of what's happening elsewhere. Mindfactory publish (to date) sales numbers - in units of 5 or 10 - for individual products on their website. So there is some rounding down/rounding up going on that is also reflected in the figures supplied by TechEpiphany.

 

[Punkenjoy]

A theory that haven't been explored, but could be that AMD is on a dead end with the RNDA architecture. They didn't had what they expected from RNDA 3, and their adjusted simulations for RDNA4 were not that much better.

They looked at where the market is heading and they probably see no huge gain for RT and AI without doing a new clean sheet architecture. So instead of spending for 3-4 chips for RNDA 4, they would just do one while they work on their next gen architecture.



This is pure speculation like almost everything under this post. But with the recent release, it could be something probable.