Amd’s Rdna4 GPU Architecture

I refer to the RDNA4 instruction set manual ([1]), page 90, Table 41. WMMA Instructions.

They support FP8/BF8 with F32 accumulate and also IU4 with I32 accumulate. The max matrix size is 16x16. For comparison, NVIDIA Blackwell GB200 supports matrices up to 256x32 for FP8 and 256x96 for NVFP4.

This matters for overall throughput, as feeding a bigger matrix unit is actually cheaper in terms of memory bandwidth, as the number of FLOPs grows O(n^2) when increasing the size of a systolic array, while the number of inputs/outputs as O(n).

1. https://www.amd.com/content/dam/amd/en/documents/radeon-tech...

2. https://semianalysis.com/2025/06/23/nvidia-tensor-core-evolu...

To wager a guess, would that optimization also help push the envelope when one application needs all the power it can get while another monitor is just sitting idle ?

Another angle I'm wondering about is longevity of the card. Not sure if AMD would positively care in the first place, but as a user if the card didn't have to grind much on the idle parts and thus last a year or two longer, it would be pretty valuable.

Some people appreciate leaving the pc open for light tasks even at night, and wasting too much power doing nothing is... well, wasteful. Imagine a home server that has the GPU for AI or multimedia stuff.

The same architecture will also be used in mobile, so depending on where this comes from (architecturally) it could mean more power savings there, too.

Besides, lower power also means lower cooling/noise on idle, and shorter cooldown times after a burst of work.

And since AMD is slowly going to the (ever next-time) unified architecture, any gains there will also mean less idle power draw in other environments, like servers.

Nothing groundbreaking, sure, but I won't say no to all of that.

Another pane of AMD GPU R&D is the _userland_ _hardware_ [ring] buffers for near direct hardware userland programming.

They started to experiment on that in mesa and linux ("user queues", as "user hardware queues").

I don't know how they will work around the scarse VM IDs, but here, we are talking near 0 driver. Obviously, they will have to simplify/cleanup a lot 3D pipeline programming and be very sure of its robustness, basically to have it ready for "default" rendering/usage right away.

Userland will get from the kernel stuff along those lines: command/event hardware ring buffers, data dma buffers, read/write pointers & doorbells memory page for those ring buffers, and an event file descriptor for an event ring buffer. Basically, what the kernel currently has.

I wonder if it will provide some significant simplification than the current way which is giving indirect command buffers to the kernel and deal with 'sync objects'/barriers.

In terms of heat output the difference between an idling gaming PC from 10 years ago (~30-40 W) and one today (100+ W) is very noticeable in a room. Besides, even gaming PCs are likely idle or nearly idle a significant amount of time, and that's just power wasted. There are also commercial users of desktop GPUs, and there they are idle an even bigger percentage of the time.

I am not saying that this was the reason I bought it, but I recently purchased a Radeon 9070 and I was surprised how little power this card uses in idle. I was seeing figures between 4W~10W on Windows (sadly slightly more on Linux).

In general this generation of Radeon GPUs seems highly efficient. Radeon 9070 is a beast of a GPU.

I definitely value lower power usage when idle. My desktop PC uses ~150W when idle. Sometimes I leave it on overnight simply downloading a remote file or some other extremely light operation.

It doesn't make sense that it would draw this much power. A laptop can do the same thing with ~10W.

This sort of improvement might not increase sales with one generation, but it'll make a difference if they keep focusing on this year after year. It also makes their design easier to translate into mobile platforms.

Rumors have been floating around about some kind of PS6 portable or next gen steam deck with RDNA4 where power consumption matters.

There's also simply laptop longevity that would be nice.

Might be helpful to get some perspective on this. Most cards idle in the 5-10 watt range, but there have been outliers, like the Intel A770 before the drivers were mostly fixed, which ran at 45 watts.[1] I believe there have been even more stupid incidents with cards combined with early drivers, where, if you just happened to have a multi-monitor setup running at two different resolutions, it would go up 70+ watts. Obviously, these are outlier situations probably caused by some optimisations being disabled while the drivers were developed, and never re-enabled before release. 20 watts and below should be easily achievable with pretty much any card, but it's easy to forget that this is still some work that should be done, and shouldn't be left to chance and happenstance.

[1] https://thinkcomputers.org/here-is-the-solution-for-high-idl...

Even if you save 0.01kw a year, if you multiply that by the number of computers, you’ll save around 100Mw of power. Even small improvements have macro level implications.

Power efficient chips will result in more overall performance for the same amount of total power drawn. Its all about performance/watt

If the tech was developed, then might as well deploy it to both laptops (likely it's original intent), as well as Desktop.

the architecture is shared between desktop and mobile. this sounds 100% like something that they did to give some dual display laptop or handheld 3 hours extra battery life by fixing something dumb.

MI300A's are very fun. very fortunate to have one to mess around on at work.