AMD RDNA 4 GPU Memory and Infinity Cache Configurations Surface

[btarunr]

AMD's next generation RDNA 4 graphics architecture will see the company focus on the performance segment of the market. The company is rumored to not be making a successor to the enthusiast-segment "Navi 21" and "Navi 31" chips based on RDNA 4, and will instead focus on improving performance and efficiency in the most high-volume segments, just like the original RDNA-powered generation, the Radeon RX 5000 series. There are two chips in the new RDNA 4 generation that have hit the rumor mill, the "Navi 48" and the "Navi 44." The "Navi 48" is the faster of the two, powering the top SKUs in this generation, while the "Navi 44" is expected to be the mid-tier chip.

According to Kepler_L2, a reliable source with GPU leaks, and VideoCardz, which connected the tweet to the RDNA 4 generation, the top "Navi 48" silicon is expected to feature a 256-bit wide GDDR6 memory interface—so there's no upgrade to GDDR7. The top SKU based on this chip, the "Navi 48 XTX," will feature a memory speed of 20 Gbps, for 640 GB/s of memory bandwidth. The next-best SKU, codenamed "Navi 48 XT," will feature a slightly lower 18 Gbps memory speed at the same bus-width, for 576 GB/s of memory bandwidth. The "Navi 44" chip has a respectable 192-bit wide memory bus, and its top SKU will feature a 19 Gbps speed, for 456 GB/s of bandwidth on tap.



Another set of rumors from the same sources also point to the Infinity Cache sizes of these chips. "Navi 48" comes with 64 MB of it, which will be available on both the "Navi 48 XTX" and "Navi 48 XT," while the "Navi 44" silicon comes with 48 MB of it. We are hearing from multiple sources that the "Navi 4x" GPU family will stick to traditional monolithic silicon designs, and not venture out into chiplet disaggregation like the company did with the "Navi 31" and the "Navi 32."

Yet another set of rumors, these from Moore's Law is Dead, talk about how AMD's design focus with RDNA 4 will be to ace performance, performance-per-Watt, and performance cost of ray tracing, in the segments of the market that NVIDIA makes the most volumes in, if not the most margins in. MLID points to the likelihood of the ray tracing performance improvements riding on there being not one, but two ray accelerators per compute unit, with a greater degree of fixed-function acceleration for the ray tracing workflow (i.e. less of it will be delegated to the programmable shaders).

View at TechPowerUp Main Site | Source

 

[TheinsanegamerN]

It sure sounds to me like AMD is gonna stick to 7800xt performance and leave high end buyers out to dry.

Unfortunate but perhaps expected by this point.

 

[ARF]

What was the AMD's outlook last time when they tried this with the RX 580 and RX 5700 XT?

Maybe there is a major problem in AMD's management, and they have to sit around that Board of directors, and decide what actions are necessary to fix the abnormally poor execution of the graphics division.

The 6800 XT which is faster in some games has those 128 MB.

 

[64K]

That is what is needed from the next generation on the gaming GPU side of things. Something to match 5050 through 5080 in performance and price and much improved ray tracing performance. Solid drivers and ongoing improving of FSR.

Not intending to start up the debate about the worth of ray tracing but it is the future one step at a time. AMD knows this and that's why they are focusing on improving ray tracing performance. The bottom line is that to compete with Nvidia they have to compete with Nvidia so that is what they are going to do. I hope.

 

[Assimilator]

It sure sounds to me like AMD is gonna stick to 7800xt performance and leave high end buyers out to dry.

Unfortunate but perhaps expected by this point.

They don't have to capture the high end of the market. They tried that with RDNA3 and failed, which meant that their high-end products were stuck in that awkward valley of "not fast enough to be compelling to high-end buyers, not cheap enough to be compelling to midrange buyers". In contrast, AMD's midrange products have been selling relatively well; they've mostly been held back by pricing and late releases.

So in terms of RDNA4, AMD "just" has to build a midrange product that comes in earlier, cheaper and faster in ray-tracing than NVIDIA, while also having drivers that don't fail at basics like multi-monitor power draw. This would allow them to claw back marketshare (and mindshare) for a potential return to the high-end with RDNA5.

 

[Daven]

I’m guessing the 4nm die size of the hypothetical 8800XT will be the same size as the 6/5 nm 7900XTX but with less CUs. All the extra space will be taken up with additional ray tracing units.

Looks like external memory controllers are gone too which will also increase the size of the monolithic die.

 

[Dr. Dro]

It sure sounds to me like AMD is gonna stick to 7800xt performance and leave high end buyers out to dry.

Unfortunate but perhaps expected by this point.

High-end buyers who expect the power and feature set of a GPU like the RTX 4090 don't even look at AMD's laughably incompetent "flagships". That is why this needs to happen - make a competent product that lays no extraordinary claims to earn people's trust and mindshare again - while getting all their ducks in a row to come back to the race. This pit stop is very much needed.

 

[64K]

High-end buyers who expect the power and feature set of a GPU like the RTX 4090 don't even look at AMD's laughably incompetent "flagships". That is why this needs to happen - make a competent product that lays no extraordinary claims to earn people's trust and mindshare again - while getting all their ducks in a row to come back to the race. This pit stop is very much needed.

Should AMD focus on the over 99% of the gaming market with whatever wafer allotment they have or gamble some of that on the xx90 buyers? imo AMD is making the best choice with their resources.

 

[ARF]

Should AMD focus on the over 99% of the gaming market with whatever wafer allotment they have or gamble some of that on the xx90 buyers? imo AMD is making the best choice with their resources.

Look at German Mindfactory. RX 7900 XTX outsold RX 7600 10 to 1 or even more.

https://www.mindfactory.de/Hardware/Grafikkarten+(VGA)/Radeon+RX+Serie/RX+7900+XTX.html

https://www.mindfactory.de/Hardware/Grafikkarten+(VGA)/Radeon+RX+Serie/RX+7600.html

 

[mkppo]

I think they're trying to substantially reduce die area and compete against the 50/60/70 series. Probably not trying to use up too much fab allocation on these chips when they can sell every instinct they can make at a margin far, far greater than gaming GPU's or pretty much any other chip.

 

[Dr. Dro]

Look at German Mindfactory. RX 7900 XTX outsold RX 7600 10 to 1 or even more.

https://www.mindfactory.de/Hardware/Grafikkarten+(VGA)/Radeon+RX+Serie/RX+7900+XTX.html

https://www.mindfactory.de/Hardware/Grafikkarten+(VGA)/Radeon+RX+Serie/RX+7600.html

1. Mindfactory is a single store
2. The EU is the most receptive market to AMD, by far
3. The 7900 XTX's 24 GB is attractive for AI use. There's a high chance many if not most of these cards sold for this reason

I think they're trying to substantially reduce die area and compete against the 50/60/70 series. Probably not trying to use up too much fab allocation on these chips when they can sell every instinct they can make at a margin far, far greater than gaming GPU's or pretty much any other chip.

Fab allocation which can be used to manufacture more profitable, high-margin Epyc processors indeed.

 

[ARF]

I think they're trying to substantially reduce die area and compete against the 50/60/70 series. Probably not trying to use up too much fab allocation on these chips when they can sell every instinct they can make at a margin far, far greater than gaming GPU's or pretty much any other chip.

Fab allocation which can be used to manufacture more profitable, high-margin Epyc processors indeed.

They should exit the markets, then:
1. Consumer Ryzen;
2. Consumer Radeon;
3. Semi-custom Playstation and Xbox chips.

 

[Vya Domus]

What was the AMD's outlook last time when they tried this with the RX 580

Polaris based cards were one of their most successful products in this last decade lol.

 

[TheinsanegamerN]

They don't have to capture the high end of the market. They tried that with RDNA3 and failed, which meant that their high-end products were stuck in that awkward valley of "not fast enough to be compelling to high-end buyers, not cheap enough to be compelling to midrange buyers". In contrast, AMD's midrange products have been selling relatively well; they've mostly been held back by pricing and late releases.

So in terms of RDNA4, AMD "just" has to build a midrange product that comes in earlier, cheaper and faster in ray-tracing than NVIDIA, while also having drivers that don't fail at basics like multi-monitor power draw. This would allow them to claw back marketshare (and mindshare) for a potential return to the high-end with RDNA5.

The 7900xtx sold relatively well, much of its early life it was out of stock.

1. Mindfactory is a single store
2. The EU is the most receptive market to AMD, by far
3. The 7900 XTX's 24 GB is attractive for AI use. There's a high chance many if not most of these cards sold for this reason

1) Agreed, but its one of the few that has actual sales numbers
2) I'd love to see the numbers to back this up.
3) Given the issues Tiny Build had with 7900s for AI builds, I highly doubt that "most" 7900xtx or Xt sales went to AI use. Running those commercial loads on nvidia is a lot easier.

High-end buyers who expect the power and feature set of a GPU like the RTX 4090 don't even look at AMD's laughably incompetent "flagships". That is why this needs to happen - make a competent product that lays no extraordinary claims to earn people's trust and mindshare again - while getting all their ducks in a row to come back to the race. This pit stop is very much needed.

Should AMD focus on the over 99% of the gaming market with whatever wafer allotment they have or gamble some of that on the xx90 buyers? imo AMD is making the best choice with their resources.

It's funny, we've heard this argument before, the last time AMD snubbed the high end and focused on mid range chips.

When AMD did this with polaris, what was the result? Oh yeah, nvidia grabbing up huge chunks of the market as high end AMD buyers migrated to nvidia as AMD stagnated on 290x performance until the disaster that was Vega. Meanwhile, Nvidia leverages that position to sell record numbers of GPUs at higher prices then anything AMD made. Then AMD released the RX 5000 series, still failing to compete at the high end, and failing to make an impact.

AMD finally got back on track with the 6000s, and the 7000s became a repeat of the RX 200s where only the high end saw improvement and prices stagnated. Now we're doing RX 5000 again, dropping the high end to favor only mid range and lower.

Imma predict this is going to go just as well as it did last time, with complete stagnation and consumers moving on to nvidia to find some semblance of improvement.

 

[64K]

Look at German Mindfactory. RX 7900 XTX outsold RX 7600 10 to 1 or even more.

https://www.mindfactory.de/Hardware/Grafikkarten+(VGA)/Radeon+RX+Serie/RX+7900+XTX.html

https://www.mindfactory.de/Hardware/Grafikkarten+(VGA)/Radeon+RX+Serie/RX+7600.html

Not sure if this is more trolling but I was referring to competing with the 4090. The 7900 XTX was the competitor to the 4080. Over 99% of the gaming market is for below the xx90 class which includes 100% of AMD's GPUs.

Imma predict this is going to go just as well as it did last time, with complete stagnation and consumers moving on to nvidia to find some semblance of improvement.

There will be improvement from AMD in ray tracing performance.

 

[londiste]

tl;dr
Exactly the same cache configuration as RDNA3 cards with same memory bus width.

 

[GhostRyder]

At the end of the day this is probably the smarter move. The very top end maybe something thats talked about, but with pricing the way they are right now its an area where most people cannot afford/Dont want to spend the money. Upper mid and midrange cards are where the market is and hopefully the changes mean the focus is going to be these areas and improvements across the board. Hopefully this will shake up the market especially if this stuff comes first at prices people can be happy with. GPU's have gotten so outrageous lately...

 

[TheinsanegamerN]

There will be improvement from AMD in ray tracing performance.

rDNA3 was supposed to improve ray tracing, but it failed to do so, with RT perf per CU being almost identical to rDNA2.

Now we're seeing rDNA4 will have the same cache and memory config as rDNA3, with a similar core count.

Where is this improvement coming from?

 

[wNotyarD]

Where is this improvement coming from?

If rumours are to be believed, either increased ray accelerator count per CU or a rework of their RT engine.

 

[R0H1T]

High-end buyers who expect the power and feature set of a GPU like the RTX 4090 don't even look at AMD's laughably incompetent "flagships".

What feature set? Running local LLM with 8GB VRAM or fake frames with "fake" reflections

 

[ARF]

trolling... I was...

Not surprising.

The 7900 XTX was the competitor to the 4080.

Let's see:

Navi 31: Die size: 529 mm^2, Transistors: 57.7 Billion, VRAM: 24 GB, Pixel rate: 480 GPix/s, Texture rate: 960 GTex/s
AD103: Die size: 379 mm^2, Transistors: 45.9 Billion, VRAM: 16 GB, Pixel rate: 280 GPix/s, Texture rate: 760 GTex/s

So, again, which is a competitor to which?

1. Mindfactory is a single store

Mindfactory is not a "single store". It is the largest retailer in Europe, and sells tons and tons of inventory. Be respectful.

 

[R0H1T]

Where did you get the die size for Navi, it's an MCM right?

Ok so from TPU, but the meaty part is just ~
304.35 mm² (GCD Die) on 5nm TSMC.

 

[64K]

Let's see:

Navi 31: Die size: 529 mm^2, Transistors: 57.7 Billion, VRAM: 24 GB, Pixel rate: 480 GPix/s, Texture rate: 960 GTex/s
AD103: Die size: 379 mm^2, Transistors: 45.9 Billion, VRAM: 16 GB, Pixel rate: 280 GPix/s, Texture rate: 760 GTex/s

So, again, which is a competitor to which?

You can't compare specs from Nvidia with AMD to understand a performance level comparison. The internet is full of benchmarks that make it clear that the 7900 XTX is the competitor to the 4080. You are going to have to do some research or you can't have a grasp of what is going on in the tech world and you just waste everyone's time trying to explain obvious things to you.

 

[DemonicRyzen666]

7900 xtx in rasterization matches +/- 5% 4090
7900 xtx enabled raytracing +/-5% to a 3090

Adding another raytracing unit does not seem wise as RDNA3 has a hardtime filling up its improved RT units with BHV transversal additions. Beside that it also doesn't even both using it 2 issue per-clock addition either unless specifically coded for it.

 

[mkppo]

rDNA3 was supposed to improve ray tracing, but it failed to do so, with RT perf per CU being almost identical to rDNA2.

Now we're seeing rDNA4 will have the same cache and memory config as rDNA3, with a similar core count.

Where is this improvement coming from?

RDNA3 had pretty decent uplift over the 6900xt in RT, where are you seeing the lack of improvement?

RDNA3 largely matched RTX 3090 in RT, while RDNA2 lagged behind it a fair bit.

Core counts and cache isn't everything that affects RT performance..