AMD reveal Ryzen 7000 X3D processors, desktop 65W CPUs and new mobile chips

I wish they also included the max TDP, like Intel in the previous announcement, and that there were 35W TDP AMD desktop AMD processors. It was also about time they recognized the design error on the 7900 XTX, and start applying RMAs.

I wish they also included the max TDP, like Intel in the previous announcement

But TDP numbers are nearly useless, all manufacturers use their own incomprehensible algorithms to conjure up their magical numbers. Gamers Nexus has addressed this many times over the last couple of years if you want to dive into this particular swamp.

On mobile I don't recommend using the latest amd CPU at all if you use Linux. Had this terrible experience with a tuxedo laptop with ryzen 6900HX where the keyboard did not work when using kernel newer than 5.19. So using distros like Garuda or pop_os was impossible.

I think same thing is happening with the GPUs now, we are always beta testers. Maybe after one year of launch it is time to buy.

I wish they also included the max TDP, like Intel in the previous announcement

But TDP numbers are nearly useless, all manufacturers use their own incomprehensible algorithms to conjure up their magical numbers. Gamers Nexus has addressed this many times over the last couple of years if you want to dive into this particular swamp.

Thanks, I didn't know. Will have a look!

It was also about time they recognized the design error on the 7900 XTX, and start applying RMAs.

They have and are doing !

https://wccftech.com/amd-confirms-radeon-rx-7900-xtx-throttling-issue-related-to-thermal-solution-used-in-reference-design/

I wonder if they still plan to make the long rumoured AM4 5600 X3D. Would be sure nice for a cost efficient upgrade on AM4 systems.

Having a 5800X3D for about a month I will never go back to a non 3d sku for gaming. Compared to my 5950X the 5800x3d is a substantial improvement in the games I play (Hunt Showdown, Sea of Thieves, Valheim) and especially those running through proton. Sea of Thieves specifically gains over a hundred fps in specific areas maxed at 1440P, without a GPU change.

I feel no pressure to upgrade from my 5800X3d/6900XT for 1440P/144, but as always, I'm tempted.

One huge problem is that on the 7900x3d and the 7950x3d the extra L3 3d-cache is only connected to one of the CCD:s so there will now be a scheduler problem that is worse than on Alder Lake in that it have to decide which thread to give extra L3 to and which thread to give extra potential cpu boost frequency to, which of course is not something the scheduler can do (a thread could very much switch between the two needs as well).

there will now be a scheduler problem that is worse than on Alder Lake

Well, a potential problem at least. If it even is a problem at launch I expect it to be addressed quickly, much like Alder Lake was.

there will now be a scheduler problem that is worse than on Alder Lake

Well, a potential problem at least. If it even is a problem at launch I expect it to be addressed quickly, much like Alder Lake was.

How? I cannot think of how a scheduler can know which thread needs larger L3 vs higher clock frequency.

I wonder if they still plan to make the long rumoured AM4 5600 X3D. Would be sure nice for a cost efficient upgrade on AM4 systems.

It seems as though they've moved on this from idea. 5800X3D was the last AM4 Desktop chip that we'll get. AM5 is all they care about going forward.

I feel no pressure to upgrade from my 5800X3d/6900XT for 1440P/144, but as always, I'm tempted.

with the generational stagnation in performance, your probably good to run that system for another 5 -7 years before you absolutely reach a point where games need too many of the settings turning down to justify.

for context (although its not exactly how things work like for like) your card has 35TFLOPS and the Steamdeck, xboxone (which is still a huge target platform) have less than 3 TFLOPS (1.8 for the steamdeck) and they run games just good enough at 720p/1080p. Even an old 1660Ti has 5.5 Tlops and can play most every game on medium/high at 60fps. Sure if you want ultra settings pegged at a constant 144hz at 1440p then you probably will be looking at a yearly or bi-yearly update.. seems excessive but it's your money and energy wise cards are not getting lower TDP's. a 4090 is supposed to be 85TFLOPS (not sure if that's accurate but possibly) that said its a 450w card!

edit* just checking and the xbox series x and the PS5 are under 12TFLOPS (again not an exact marker for performance). But im not sure if sony or MS are going to put a 450w GPU plus competing CPU in a small form factor console. So with your hardware you are definitely covered for a loooong time.

Perhaps lower raytracing performance will be the issue with your hardware though.


Last edited by Lofty on 6 January 2023 at 5:08 pm UTC

How? I cannot think of how a scheduler can know which thread needs larger L3 vs higher clock frequency.

That's my thought too, but I think this so called "big little" issue exists for a while (started with ARM?) and may be there was some work for Linux on that front before in regards to asymmetric cache?

I also wonder what will happen if the scheduler will be unaware of any of that and scheduling will be random. I.e. what will perform better, 7950X or 7950X3D? Some thorough benchmarks comparing them will be for sure needed.


Last edited by Shmerl on 9 January 2023 at 7:07 am UTC

How? I cannot think of how a scheduler can know which thread needs larger L3 vs higher clock frequency.

That's my thought too, but I think this so called "big little" issue exists for a while (started with ARM?) and may be there was some work for Linux on that front before in regards to asymmetric cache?

I also wonder what will happen if the scheduler will be unaware of any of that and scheduling will be random. I.e. what will perform better, 7950X or 7950X3D? Some thorough benchmarks comparing them will be for sure needed.

ARM big.LITTLE is a easy problem in this complex problem space. There the scheduler have the "simple" choice of "should this thread run on a high performing CPU or on a low performing CPU", it does this by doing some metrics (which still after all these years are far from perfect). Alder Lake have very similar design but there they added metric collection in the hardware (Intel Thread Director) buf AFAIK this is still far from perfect even in Windows 11 where compile jobs sometimes gets scheduled to the e-cores and thus takes 55min to complete instead of 17.

The problem space here is that both cores are high performance, just in different ways. I mean trying to determine if your thread/application would benefit from a higher clock or a larger cache is something that takes endless long benchmarks in numerous runs for application developers today (to determine which cpu to recommend to the enterprise to run the system on).

My guess is that MS (at this point in time AMD have only talked to Microsoft AFAIK) will simply (if they will do anything at all) try to detect if the application is a game or not and if so run it on the larger cache cores while running everything else on the higher boost cores.

To really benefit here the app/game developers would have to bench this individually and set the thread affinity but the number of combinations in combination with this probably going to be a niche cpu I have a hard time seeing this been done.

I think the most telling of all is that neither AMD nor Intel have any plans what so ever to implement either of these strategies on the server market.

The problem space here is that both cores are high performance, just in different ways. I mean trying to determine if your thread/application would benefit from a higher clock or a larger cache is something that takes endless long benchmarks in numerous runs for application developers today (to determine which cpu to recommend to the enterprise to run the system on).

I wonder if AI can help with scheduler for that. It feels like prediction problem based on some moving sample input.

AMD are now adding AI chips to some of their APUs, so may be this can even be hardware accelerated in the future.


Last edited by Shmerl on 9 January 2023 at 8:12 pm UTC

The problem space here is that both cores are high performance, just in different ways. I mean trying to determine if your thread/application would benefit from a higher clock or a larger cache is something that takes endless long benchmarks in numerous runs for application developers today (to determine which cpu to recommend to the enterprise to run the system on).

I wonder if AI can help with scheduler for that. It feels like prediction problem based on some moving sample input.

AMD are now adding AI chips to some of their APUs, so may be this can even be hardware accelerated in the future.

Possibly

Edit: The 7900 XTX is lovely, other than the fact I can't see my cursor under X11. Still don't need a CPU upgrade though.

Edit: The 7900 XTX is lovely, other than the fact I can't see my cursor under X11. Still don't need a CPU upgrade though.

Good to know! Are games working well? And does the cursor work in KDE Wayland session?


Last edited by Shmerl on 24 January 2023 at 9:55 pm UTC

In every DX11/DX12/Vulkan game I've tried thus far, yes*. Before commenting back I fired up the oldest game I currently have installed for shits and giggles, GTA IV.

!GTA4_borked

I installed plasma for you but no, the cursor is not visible there either (enabling the software cursor works). I'm not the only one so I'll double check a bug hasn't been filed already.

*Even the Witcher 3 enhanced seemed to run fine under DX12 but appeared to have a floor texture bug. More investigation needed.

Edit: I really notice the GPU performance uplift in Valheim (vulkan). Not the typical open world but in areas with lots of terraforming and buildings.

Edit 2: Until mutter supports VRR I've switched to a Plasma wayland session as my default. Freesync is working as expected. I should also note this isn't exactly stock Fedora. It's Fedora 37 with Rawhide's (so Fedora 38 currently) no-debug kernel (currently 6.2.0-0.rc4.31.fc38.x86_64).


Last edited by drlamb on 25 January 2023 at 8:51 pm UTC