Thursday, February 19th 2026

AMD Ryzen 500-Series "Medusa Point" APU Brings RDNA 4m iGPU With FSR 4 Support

AMD's next-generation Ryzen 500 series of APUs, which include the "Medusa Point" and "Medusa Halo" SoCs, are slowly taking shape, and we are learning more about the integrated graphics of "Medusa Point." According to the latest compiler patches, "Medusa Point" will feature a variant of RDNA 4 called RDNA 4m, likely a mobile version designed for laptops and mini PCs. In contrast, the "Medusa Halo" will utilize AMD's next-generation RDNA 5 / UDNA GPU microarchitecture. "Medusa Point" will introduce a desktop-exclusive RDNA 4 with the new RDNA 4m variant. Although we initially lack comparisons between the two, some instruction set extensions, such as WMMA and SWMMAC instructions, indicate support in the new "GFX1170" GPU, which should be associated with the GFX11 generation, also known as RDNA 3. Currently, this is believed to be an upgraded RDNA 3 with many RDNA 4 modules, enabling FSR 4 support even on the less powerful "Medusa Point" APU.

When it comes to the overall system configuration, AMD targets "Zen 6" CPU cores for both "Medusa" variants. Both "Point" and "Halo" versions will likely differentiate with the same factors we have in the current generation APUs, which is a mix of AMD's "Zen 6" and "Zen 6c" CPU cores for the "Medusa Point," while the more powerful "Medusa Halo" will feature a uniform "Zen 6" core cluster for maximum CPU performance on that side as well. Recently, we learned that the platform will adopt LPDDR6 memory, which will increase the memory bandwidth by 50% over the standard LPDDR5X memory we see implemented today.
Interestingly, this is AMD's first generation that will step away from the RDNA 3.5-based iGPU design in the coming years. AMD plans to make its RDNA 3.5 iGPU long-lived until 2029, so most consumers buying laptops in the coming years will experience the same integrated graphics capabilities as those found in today's Ryzen AI 300 and 400 series processors. This means that the RDNA 4m, with all its cutbacks, will be a welcome addition to the integrated graphics as "Redstone" features like ML-based upscaling and rendering technologies rely on the RDNA 4 IP to make it work.
Sources: LLVM Compiler Patches, Phoronix, via VideoCardz
Add your own comment

13 Comments on AMD Ryzen 500-Series "Medusa Point" APU Brings RDNA 4m iGPU With FSR 4 Support

#1
TheinsanegamerN
Wow. That took, what, a week from panther lake launching for AMD to backtrack on the whole "rDNA 3.5 until 2029" thing? :roll:
Posted on Reply
#2
Apocalypsee
What's with the segmentation? RDNA 3, 3.5, 4, 4m, 5 don't tell me some of them gonna be on legacy driver? Or worse, only some will get support to latest FSR, which to me is a big thing for integrated GPU because limited performance and power.

I give credit to Intel to actually give Xess 3 support to all Alchemist, Battlemage and Core Ultra iGPU, no exceptions.
Posted on Reply
#3
dgianstefani
TPU Proofreader
Finally. Non RDNA 2/3 APUs.

What stopped them from doing this a year ago? Instead of the Strix refresh that bumped clocks by 100 MHz?

Ofc it's still not full RDNA 4, gotta do it by half measures...
ApocalypseeWhat's with the segmentation? RDNA 3, 3.5, 4, 4m, 5 don't tell me some of them gonna be on legacy driver? Or worse, only some will get support to latest FSR, which to me is a big thing for integrated GPU because limited performance and power.

I give credit to Intel to actually give Xess 3 support to all Alchemist, Battlemage and Core Ultra iGPU, no exceptions.
Only 4 and newer get FSR 4, and likely future revisions.

Agreed, iGPU where upscaling matters most.

Actual competition for handhelds on the horizon instead of RDNA 3.5 refreshes getting blown out of the water by Lunar Lake, with the downside being LNL is 20% more expensive.

Panther Lake vs Medusa Point.
Posted on Reply
#5
jesdals
Lets get to commercial use og DDR6 as fast as possible so we can buy DDR5 again 👍
Posted on Reply
#6
AntiX3DGlazer
jesdalsLets get to commercial use og DDR6 as fast as possible so we can buy DDR5 again 👍
that's... not how this works
Posted on Reply
#7
ymdhis
RDNA4m is a marketing name, going by the internal codename this would be more like RDNA3.7 (GFX11.7, RDNA3.5 is GFX11.5, RDNA3 is GFX11, and RDNA4 is GFX12).

Also this is still 1 years from release, and 3 years from being ported to AM5.
ApocalypseeWhat's with the segmentation? RDNA 3, 3.5, 4, 4m, 5 don't tell me some of them gonna be on legacy driver? Or worse, only some will get support to latest FSR, which to me is a big thing for integrated GPU because limited performance and power.

I give credit to Intel to actually give Xess 3 support to all Alchemist, Battlemage and Core Ultra iGPU, no exceptions.
RDNA4 did not perform to expectations so they didn't bother developing it further; AMD put his R&D into RDNA5 instead. Same as RDNA1, that one didn't have high end nor igpu versions either, iGPUs went from Vega to RDNA2.

Radeons and APUs are both considered third rate backup products by AMD compared to the Zen gravy train, so they don't update iGPUs unless the stars align. It's really very annoying.
Posted on Reply
#8
jesdals
AntiX3DGlazerthat's... not how this works
No but one can only dream and hope in 2026
Posted on Reply
#9
TheinsanegamerN
dgianstefaniFinally. Non RDNA 2/3 APUs.

What stopped them from doing this a year ago? Instead of the Strix refresh that bumped clocks by 100 MHz?

Ofc it's still not full RDNA 4, gotta do it by half measures...


Only 4 and newer get FSR 4, and likely future revisions.
What stopped them was that intel couldn’t yet compete with rDNA 3.5. Now that panther lake and Xe3 are here AMD has to get off the couch and do something again.
ymdhisRDNA4m is a marketing name, going by the internal codename this would be more like RDNA3.7 (GFX11.7, RDNA3.5 is GFX11.5, RDNA3 is GFX11, and RDNA4 is GFX12).

Also this is still 1 years from release, and 3 years from being ported to AM5.


RDNA4 did not perform to expectations so they didn't bother developing it further; AMD put his R&D into RDNA5 instead. Same as RDNA1, that one didn't have high end nor igpu versions either, iGPUs went from Vega to RDNA2.

Radeons and APUs are both considered third rate backup products by AMD compared to the Zen gravy train, so they don't update iGPUs unless the stars align. It's really very annoying.
In what way did rDNA 4 not perform? From efficiency to RT performance per core rdna4 was a massive step up in every way.
Posted on Reply
#10
GenericUsername2001
I can see why AMD would do this, rather than use RDNA 4 for lower end mobile products. RDNA 4 is a big leap in terms of transistor count over RDNA 3; to compare two 32 CU GPUs the RX 9060 XT has 29 billion transistors vs the RX 7600's 13 billion. A lot of those extra transistors went to ray tracing hardware + a side of supporting higher clock speeds.

Mobile iGPUs aren't going to be fast enough for heavy ray tracing games even if they included the RT hardware, and obviously clocks are limited on mobile platforms for power use reasons. On the flipside though, better hardware upscaling is appealing for low end GPUs. So adding just FP8 & a few other things to RDNA 3.5 was probably easier than trying to cut back RDNA4.
Posted on Reply
#11
ymdhis
TheinsanegamerNIn what way did rDNA 4 not perform? From efficiency to RT performance per core rdna4 was a massive step up in every way.
They wanted to do flagship cards using chiplets a la RX7800 and RX7900, but it didn't perform to expectations (I forgot the exact problem, it was either power draw or latency or simply not being fast enough). So in that way, it did not perform as intended - all they could get out of the tech was a mid range monolithic chip, while nVidia was showing off the 5090.

There was also the buggered launch that got delayed months for whatever reason with cards already sitting in retailer warehouses.
Posted on Reply
#12
qlum
TheinsanegamerNWow. That took, what, a week from panther lake launching for AMD to backtrack on the whole "rDNA 3.5 until 2029" thing? :roll:
Given that amd loves to keep using other cpu-cores on the low-end it would not surprise me that tha ryzen ai confusing name using zen4 will continue to exist for a while yet with "new" models.
Posted on Reply
#13
dgianstefani
TPU Proofreader
TheinsanegamerNWhat stopped them was that intel couldn’t yet compete with rDNA 3.5. Now that panther lake and Xe3 are here AMD has to get off the couch and do something again.
Lunar Lake with 140V beats Z2 and is a significantly better handheld chip, so, not really.

It's just cheaping out, as usual. RDNA 4 should have been rushed to laptops, but there's no presence dGPU or iGPU.

The reality is though that Intel dominates the laptop market, and if Panther Lake is anything to go by, I wouldn't be surprised to see them dominate the handheld market too. Stuff like Strix Halo doesn't move the needle, might as well be a technology demonstrator, because it's not out there in volume and doesn't really have a market.
Posted on Reply
Add your own comment
Nov 19th, 2026 01:09 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts