Post reply

The message has the following error or errors that must be corrected before continuing:: Warning - while you were reading a new reply has been posted. You may wish to review your post.

Name
Email
Subject
Message icon

Other options

Return to this topic
Don't use smileys

Verification:

Please leave this box empty:

Shortcuts: ALT+S post or ALT+P preview

Topic summary

Posted by Exynos 2600

- Today at 16:09:00

Quote from: RDNA 4 iGPU on Today at 12:48:07Estimated integrated mobile RDNA4 960M / 980M / 990M launch: January 23, 2026 (probably at the CES 2026) to April 23, 2026
Launch at CES 2026 is expected, but the rumor is that it's gonna be just a rename (with a maybe higher clock), like Phoenix to Hawk Point.

The question is: What made AMD use RDNA3 in integrated graphics and why would AMD not use RDNA4 in 2026?

RDNA 5 based on the desktop launches:
January 16, 2027 (using average) (CES 2027?)
May 27, 2027 (using pattern)
+ adding the 6 to 7 months from above
Medusa Halo: July 19 2027 to August 18, 2027

I really don't see AMD launching RDNA4 or RDNA5 igpus that early. Whenever AMD launch a new gfx architecture, it always comes to their dgpus first then it takes atleast a couple of years to later come to their consumer apus. I don't know the reason for this delay, other than they don't care to prioritise them more or they don't have the resources to do it any faster but that's always how it's been. That said, I expect PS5 to be one of the first to get RDNA5 and I guess you could call that an igpu/APU.

If RDNA4 was being announced this CES 2026 or in next 6 months I'm pretty sure we would have had heard leaks about it by now as that is pretty big news.

So far it seems the first rdna4 igpu won't be an x86 APU but Samsung's arm SoC, the Exynos 2600.

@256-bit vs 384-bit,

I must admit, have been very impressed by glm 4.7 when it comes to coding tasks. Every other free cloud LLM I tried that far couldn't really code. This thing actually can, provided you stick to simpler tasks and are specific about it. Anything more complex and it starts making mistakes. But that's why they're letting people use it for free I guess, they're trying to quickly improve it. Too bad it has massive memory requirements so I would never run it locally. Maybe in future we will have an llm of this caliber that can run on 32 GB with how fast things are progressing.

Posted by RDNA 4 iGPU

- Today at 12:48:07

QuoteAfter all, didn't they skip rdna1 by going directly from vega to rdna2?

Indeed.

RDNA 1 en.wikipedia.org/wiki/RDNA_(microarchitecture):
desktop launch: Juli 7, 2019
+ 555 days or 18 months and 5 days
Ryzen 5000 integrated mobile launch: Jan 12 2021

RDNA 2 en.wikipedia.org/wiki/RDNA_2:
earliest desktop launch: Navi 21: Nov 18, 2020
+ 412 days or 13 months and 17 days
Ryzen 6000 integrated mobile 660M and 680M launch: Jan 4, 2022

en.wikipedia.org/wiki/RDNA_3:
RDNA 3
desktop launch: December 13, 2022
+ 22 days to ~60-90 days
Ryzen 7040 "Phoenix" integrated mobile 760M and 780M announcement: January 4, 2023 (delayed to April?)

RDNA 3.5
(no desktop)
880M and 890M: Jul 2024
+ 184 to 214 days or 6 to 7 months
Strix Halo: Jan 2025

RDNA 4 en.wikipedia.org/wiki/RDNA_4:
desktop: February 28, 2025
+ 329 days to 419 days (average of last 3 integrated mobile launches)
Estimated integrated mobile RDNA4 960M / 980M / 990M launch: January 23, 2026 (probably at the CES 2026) to April 23, 2026
Launch at CES 2026 is expected, but the rumor is that it's gonna be just a rename (with a maybe higher clock), like Phoenix to Hawk Point.

The question is: What made AMD use RDNA3 in integrated graphics and why would AMD not use RDNA4 in 2026?
RDNA4 in integrated graphics would be nice:
RDNA3 780M improves by 10-20% over RDNA2 680M, probably limited by memory bandwidth, which aligns with the memory bandwidth increase of 6400 MT/s (RDNA2) to 7500 MT/s (RDNA3).
RDNA3.5 uses 8000 MT/s. For RDNA4, AMD would need to use LPDDR5X-9533 MT/s memory. 9533 MT/s memory is available. Hardware-accelerated FP8 for ML upscaling seems to especially make sense in laptops, where one could then upscale 720p to 1080p.

RDNA 5 based on the desktop launches:
January 16, 2027 (using average) (CES 2027?)
May 27, 2027 (using pattern)
+ adding the 6 to 7 months from above
Medusa Halo: July 19 2027 to August 18, 2027

Posted by 256-bit vs 384-bit

- December 29, 2025, 10:28:33

Quotea 384-bit LPDDR6 or a 256-bit LPDDR5X memory controller.

(AMD's Strix Halo has a 256-bit wide memory interface and supports up to 128 GB RAM. Strix Halo: 8000 MT/s * 256-bit / 1000 / 8 = 256 GB/s.)

AMD specifically advertises Strix Halo for running AI / LLMs and if the memory density wouldn't change with LPDDR5X in 2027+, then the 256-bit rumor does only make sense, if the "Halo" word is defined to stay at 256-bit / at basically the same performance as Strix Halo (just like the "Point" word in e.g. Phoenix Point indicates the same tier). This [using LPDDR5X] would make Medusa Halo a refresh/rename of Strix Halo, just like Phoenix Point to Hawk Point. Hawk Point came one year after Phoenix Point, but Medusa Halo would come out 2+ years after Strix Halo and to be at the same tier / performance and memory size (let's say it will be 10000 MT/s vs 8000 MT/s, but still at the same up to 128 GB RAM), is almost unexpected?

In 2027+, what would really be needed, is at least a 384-bit wide chip (preferable a 512-bit one): All the big and smart generalizer MoE LLMs require at least a 384-bit wide memory interface and the 192 GB RAM that would come with it to fit into the RAM.

If DDR6 / LPDDR6 would (almost) double the memory bandwidth and double the memory density at the same bit width, then a 256-bit width would be enough:
14400 MT/s * 256-bit / 1000 / 8 = 460.8 GB/s and up to 256 GB RAM. Improvements in the MoE LLM architecture, would also reduce the requirement for the memory bandwidth: 8000 MT/s vs 14400 MT/s is not double, but then it wouldn't need to be either. But the doubling in memory size would be required, as this is how MoE scales (fewer active parameters (=less memory bandwidth required), but more in size to compensate for it).

[1]:

huggingface.co/unsloth/GLM-4.7-GGUF (the Q3_K_M quant is 171 GB, so one would at least need 171 GB RAM + additional 10-20 GB for the OS and the context)
huggingface.co/unsloth/MiniMax-M2.1-GGUF (at least a Q4_K_M quant (138 GB) is recommended according to the latest user comments)
huggingface.co/unsloth/MiMo-V2-Flash-GGUF (Q3_K_M (147 GB) to Q4_K_M (187 GB))

Posted by RXAX

- August 24, 2025, 14:45:39

Quote from: Hotz on August 23, 2025, 10:55:55AMD: RDNA 4 iGPUs haven't even arrived

I don't think rdna4 igpus are ever coming. All the leaked timelines for it suggest very late release dates, like 2027. By that time, they might as well just go rdna5/udna. It's not unusual for amd to skip generations. After all, didn't they skip rdna1 by going directly from vega to rdna2?

Also, wasn't there an interview from someone pretty high up at amd recently, where someone asked about amds plans for rdna4 outside of desktop? And the response was pretty much, rdna4 is mainly for desktop only, that there aren't any plans currently in the making and not to expect much else.

Months ago there were some leaks about rdna4 mobile dgpu being worked on but it wouldn't surprise me if they've been cancelled. Nvidia is just too dominant when it comes to laptop dgpus with OEMs, so it seems AMD is just doubling down on the APU/igpu strategy.

Posted by Alphonsokurukuchu

- August 24, 2025, 05:43:28

ofcourse it's MILD

Posted by Hotz

- August 23, 2025, 10:55:55

AMD: RDNA 4 iGPUs haven't even arrived in the bulk of mini-PCs and laptops (only RDNA 3 or 3.5 in the case of Strix Halo), and yet they wave the flag about RDNA 5....

INTEL: Xe2 iGPUs haven't even arrived in the bulk of mini-PCs and laptops (only as niche products, whereas it should have been in Arrow Lake or Refresh), and yet they wave the flag about Xe3....

It's disgusting how both of them are jumping from one iteration to the next within such a short time, without delivering in large quantities, and make you feel whatever you buy is old §hit, and cause the driver support being be stopped earlier, as only the new GPU iteration will be served.

Posted by Drakkan

- August 23, 2025, 05:51:35

Quote from: Papua on August 23, 2025, 02:30:05People haven't even been able to get their hands on strix halo chips yet.

Eh? I got one months ago with no effort.

Posted by Papua

- August 23, 2025, 02:30:05

People haven't even been able to get their hands on strix halo chips yet.

Posted by Redaktion

- August 22, 2025, 20:35:55

In a comprehensive leak covering AMD Zen 6 APUs, including Medusa Point, serial leaker Moore's Law Is Dead has revealed a ton of details regarding the Medusa Halo and the Medusa Halo Mini APUs. Both APUs feature a mix of Zen 6 CPU cores and an RDNA 5 iGPU.

https://www.notebookcheck.net/Detailed-AMD-Medusa-Halo-and-Medusa-Halo-Mini-APUs-leak-claims-up-to-26x-Zen-6-cores-and-next-gen-RDNA-5-iGPUs.1093912.0.html