Wednesday, April 24th 2024
AMD's RDNA 4 GPUs Could Stick with 18 Gbps GDDR6 Memory
Today, we have the latest round of leaks that suggest that AMD's upcoming RDNA 4 graphics cards, codenamed the "RX 8000-series," might continue to rely on GDDR6 memory modules. According to Kepler on X, the next-generation GPUs from AMD are expected to feature 18 Gbps GDDR6 memory, marking the fourth consecutive RDNA architecture to employ this memory standard. While GDDR6 may not offer the same bandwidth capabilities as the newer GDDR7 standard, this decision does not necessarily imply that RDNA 4 GPUs will be slow performers. AMD's choice to stick with GDDR6 is likely driven by factors such as meeting specific memory bandwidth requirements and cost optimization for PCB designs. However, if the rumor of 18 Gbps GDDR6 memory proves accurate, it would represent a slight step back from the 18-20 Gbps GDDR6 memory used in AMD's current RDNA 3 offerings, such as the RX 7900 XT and RX 7900 XTX GPUs.
AMD's first generation RDNA used GDDR6 with 12-14 Gbps speeds, RDNA 2 came with GDDR6 at 14-18 Gbps, and the current RDNA 3 used 18-20 Gbps GDDR6. Without an increment in memory generation, speeds should stay the same at 18 Gbps. However, it is crucial to remember that leaks should be treated with skepticism, as AMD's final memory choices for RDNA 4 could change before the official launch. The decision to use GDDR6 versus GDDR7 could have significant implications in the upcoming battle between AMD, NVIDIA, and Intel's next-generation GPU architectures. If AMD indeed opts for GDDR6 while NVIDIA pivots to GDDR7 for its "Blackwell" GPUs, it could create a disparity in memory bandwidth performance between the competing products. All three major GPU manufacturers—AMD, NVIDIA, and Intel with its "Battlemage" architecture—are expected to unveil their next-generation offerings in the fall of this year. As we approach these highly anticipated releases, more concrete details on specifications and performance capabilities will emerge, providing a clearer picture of the competitive landscape.
Sources:
@Kepler_L2 (on X), via Tom's Hardware
AMD's first generation RDNA used GDDR6 with 12-14 Gbps speeds, RDNA 2 came with GDDR6 at 14-18 Gbps, and the current RDNA 3 used 18-20 Gbps GDDR6. Without an increment in memory generation, speeds should stay the same at 18 Gbps. However, it is crucial to remember that leaks should be treated with skepticism, as AMD's final memory choices for RDNA 4 could change before the official launch. The decision to use GDDR6 versus GDDR7 could have significant implications in the upcoming battle between AMD, NVIDIA, and Intel's next-generation GPU architectures. If AMD indeed opts for GDDR6 while NVIDIA pivots to GDDR7 for its "Blackwell" GPUs, it could create a disparity in memory bandwidth performance between the competing products. All three major GPU manufacturers—AMD, NVIDIA, and Intel with its "Battlemage" architecture—are expected to unveil their next-generation offerings in the fall of this year. As we approach these highly anticipated releases, more concrete details on specifications and performance capabilities will emerge, providing a clearer picture of the competitive landscape.
114 Comments on AMD's RDNA 4 GPUs Could Stick with 18 Gbps GDDR6 Memory
When I play Helldivers2 and loving the performance at 4K am I thinking about Nvidia.
None of these products are here and just about everything I have read is conjecture and projection of opinions.
This explains why the last launch of Radeon was in Germany. Nvidia has strong mind share when people lament that it can't do DLSS to do 4K when the secret is that it does not need any upscaling to play 4K just fine at 144Hz.
The biggest issue for RDNA3 users is not performance but low computing power draw like Video playback.
Or something like that, Idk. I don't think that AMD will abandon radeon, if they wanted to, they would've done so long ago. They'll walk it off, there's a reason as to why they're looking to target the most popular market segment, instead of the halo marketing crown.
Not that halo products are pure marketing, but most consumers will be looking at the mid-low price range, let's be real here.
I was on Nvidia for over 10 years, moved to AMD now and I can't say I'm looking at a better product. Its not worse either. But certainly not better. It does lack features the competitor does have. So maybe it is worse, but I've told myself I don't need that featureset. Might be some cognitive dissonance there... ;) There are a tiny handful of things I like about my 7900XT as well, but there is no killer feature that makes me never go back to green. And for the rest of it... its a GPU that does GPU things fine.
Its really that simple in the end. If you're not making products people want, people won't buy them.
1. For the life of the PS5/PS5 Pro and Xbox 1 games will be created on Ryzen/Radeon platform. As they age, programming on those will improve that will mean an advantage for console ports using AMD PCs. It is already happening.
2. AMD are making crazy money on their APUs. The Steam Deck is in the top 10 in Global sales on the Steam platform consistently. The release of the MSI Claw (even if it is for future proofing) is evidence of how far AMD has come in the APU space. This will also mean more programming for Ryzen/Radeon as Games start to get ported (likely from Mobile) onto these platforms. In fact I am confident that someday soon on Amazon you will be able to buy a Ryzen based handheld with those retro Roms like PS/PS2/Dreamcast and Arcade. I have already built one with my 5600G (desktop).
3. Outside of the US, AMD Radeon is actually doing pretty good in terms of sales
4. Where I think people are right about AMD leaving the high end is pricing. The funny thing about the 6500XT was after it was used to compare the 86/8700G in reviews the sales increased to the point where the card has increased by $70 where I live in price. In fact AMD is already the complete stack of current GPUs vs Nvidia. As an example even the lowly 6500XT supports HDMI 2.1 and DP 1.4 for nice support for your 4K 120Hz TV. So even from the 6400 (I think) those kinds of features are more compelling (to me) than other features that GPUs come with.
5. The China/AI effect has coloured the market. People are quick to qoute market share with out rememebering that in the 3rd quarter Nvidia pumped their Asian Partners with 4090s to sell as many to China before the Embargo date. Even today I am sure I saw a headline on TPU about this topic. Those cards are considered GPUs and applied to GPU sales. AI is the new buzz word in tech and the narrative from Nvidia is they are not only the best but the only kid in town.
Because Nvidia is the master at that as well as loyal fanboys spreading rumors that AMD drivers suck, constantly sneeringly mentioning products that were failures (Fury X) and the independent media not even acknowledging Radeon's existence.....well that group of people which potentially would be happy with AMD gpu's gets slimmer and slimmer and slimmer.
To the point that investing in that is not even something worth doing.
Has FSR, that works for everyone, including abandoned Nvidia owners (GTX1080 etc) done anything to change that perception and give them more sales? nope.avi.
So why continue to invest?, they wont win that battle, the market/people have decided on that already.
AMD usually has: a rocky launch, a feature isn't quite ready, selling points are delivered somewhere down the line. Features aren't available sometimes. Products are late to market. Etc.
None of that is marketing. Its company performance.
Recent example:
FSR hasn't made waves because its not quite on the level of DLSS and it mostly serves, ironically, Nvidia users on Pascal. FSR3 is too late for RDNA3. Those are strategic blunders.
Nvidia is still suffering woe from the 12VHPWR adoption. AMD had high idle power draw at launch.
I did not ask for (but appreciate) Freesync, Crossfire or FSR
Late to market as compared to what exactly? There are only 2.5 now.
DLSS 1.0 remember that stuff? still sold the idea of that gamers need it.
Nvidia being late to market is hard when they decide the market...
And yes FSR serves Nvidia users on Pascal, the idea being to create some goodwill with people who currently use Nvidia to maybe go for AMD in the future, as I said, has it made waves? has that effort paid off? nope, just a bunch of negativity (again look at Digital Foundry).
But whatever, nothing changes the fact that over 80% of the market is in Nvidia's hands, you might say its AMD's fault its like that, but that does not change the fact that it is like that and so only more and more logical for AMD to dip out of that segment entirely, maybe keeping the midrange which sells a teeny tiny bit.
I certainly wont fault them for it, the market has spoken for a while now.
And yeah I remember DLSS1.0. When did AMD start on FSR? They've had their sweet time.
But there are bigger sausages on the green end.
Probably they will saturate the market with 16GB VRAM GPU's more efficient powerdraw at better prices (since everyone wants mooooree VRAM ), and leaving High-End performance ( for the time being) to Nvidia.
and launch after a revised SKU with GDDR7 ?
Something is going on at AMD strategy...
There have been rumors of top end RDNA4 having 7900xtx performance with far less power used for $400. That sounds too good to be true but it won't be long before we find out. That is certainly the goal. AMD has a steeper hill to climb to catch up with Nvidia than they did to catch up to Intel. RDNA4 will be 2 generations behind Nvidia 5000.
$400 sounds abit unrealistic
store.steampowered.com/hwsurvey/videocard/ Two come to my mind:
1. 16-pin 12VHPWR which melts.
2. DP 1.4 which brings only 25.92 Gbit/s signal throughput from the card to the displays. Good enough for 8K@30 8-bit colour, or 4K@120 8-bit colour.
But if you look at it proper, look at that, the most popular cards are all found within the mid-low range.
Though it may be influenced by OEMs, it is still undeniably the most popular segment.
As for Intel, their ray tracing engine is great, but otherwise, I don't have any fears of them catching up within the next 2-3 generations. I hope they prove me wrong, though, as we desperately need a price war.
www.mindfactory.de/Hardware/Grafikkarten+(VGA)/Radeon+RX+Serie/RX+7900+XTX.html
I very much doubt Nvidia will use GDDR7 for ALL their RTX 5000 series cards either... One bad product is not enough to sink a company this big. AMD made dogshit CPU's for tears and it took nearly a decade for thing to get so bad that they were on the verge of closing. One product in a two year life cycle is hardly meaningful. Also you assume RDNA4 is bad because the lack of GDDR7? That's bizarre reasoning. I what way? Because by this metric everyone is "missing out on AI" aside from Nvidia.
Also most people dont want better AI in their GPU's. They want better performance at same or lower prices. Oh you can count on Intel to mess something up. Besides from what i've seen from the leaks Battlemage is also targeting the same performance tier as RDNA4. More is better, but not if the cost is double, but the gain is not. GDDR6X is Nvidia exclusive just as GDDR5X was. Both mode by Micron. Also there are no GDDR7 cards released yet so at the moment it's strange to say that AMD is somehow two generations behind.
Also first gen GDDR7 will only run at 28Gbps speeds. Not that big of an improvement over standard 18-20Gbps GDDR6 or up to 22Gbps GDDR6X.
And at least the top N48 based die will certainly not cost 199 for 7900XT like performance. 399-450 is expected. Indeed. Most people want better performance at same or lower prices. Not ultra expensive $2000 GPU's that need DLSS FG to barely run 60fps at 4K. And it took 10 years to get to that 6-12 months point. And RDNA will only have roughly 12 months on the market before RDNA5. It's stopgap solution to have something out there. Intel will take Radeon to town? The same Intel who missed their Initial Arch launch by a country mile. The same Intel who plans to target the same performance/price as RDNA4 later this year? And the same Intel who has not even managed half or Radeon's market share thus far? Well if you call struggling by losing only to 4090 outright (in both raster and RT), then id say AMD is not doing so bad at 4080 performance.
Counter-Strike 2.
It is a problem. Because the halo product sells all the other siblings.
I would prefer to buy Radeon RX 9900 XT when it appears with RDNA 5, than the low-end-mid-range RX 8800 XT which succeeds RX 5700 XT - RX 6600 XT - RX 7600 XT.
2. Ray tracing doesn't run properly on anything except for the 4090. The post I quoted proves otherwise.
The 8800 XT is positioned to be at roughly 7900 XT level by rumours. Where you get that it's a 7600 XT successor is beyond me.