Any chance of nexgen consoles having asymmetrical cores and a task scheduler?

HJuggernaut

Member
Asymmetrical CPU cores and GPU Compute Units combined with an intelligent task scheduler, offer significant benefits for video game consoles by optimizing performance, power efficiency, and responsiveness. In this architecture, high-performance cores handle demanding tasks like physics calculations and AI, while smaller efficiency cores manage background operations such as audio processing or system services without wasting power. Similarly, GPU CUs can be designed with varying capabilities—some optimized for complex shading and others for simpler rendering or post-processing. A smart scheduler dynamically assigns workloads based on the nature and priority of each task, ensuring that resources are used efficiently. This enables smoother frame rates, reduced latency, and improved thermal performance, all while extending hardware lifespan and supporting more immersive and complex game worlds without requiring developers to micromanage hardware resources.

yes I used chatgpt but its my idea.
 
I asked ChatGPT to make it simpler:

Game consoles can use big fast cores for hard work and small ones for easy stuff to save power.

The graphics chip does the same with different parts for different jobs. A smart system decides who does what, so games run smoother, cooler, and better.
 
Yes. Very likely we'll see a mix of Zen6 and Zen6c (if they don't just use the c-variant entirely). AMD will also introduce a new Zen-LP core, which seems to be similar to Intel's LPI concept so they may end up having that too.
 
Kristen Bell Reaction GIF
 
This complexity demands highly sophisticated scheduling and developer tools, making it harder to program for and potentially leading to optimization headaches if not managed perfectly. It can also introduce bottlenecks if tasks aren't distributed optimally negating some efficiency gains.

I used Gemini.
 
Asymmetrical CPU cores and GPU Compute Units combined with an intelligent task scheduler, offer significant benefits for video game consoles by optimizing performance, power efficiency, and responsiveness. In this architecture, high-performance cores handle demanding tasks like physics calculations and AI, while smaller efficiency cores manage background operations such as audio processing or system services without wasting power. Similarly, GPU CUs can be designed with varying capabilities—some optimized for complex shading and others for simpler rendering or post-processing. A smart scheduler dynamically assigns workloads based on the nature and priority of each task, ensuring that resources are used efficiently. This enables smoother frame rates, reduced latency, and improved thermal performance, all while extending hardware lifespan and supporting more immersive and complex game worlds without requiring developers to micromanage hardware resources.

yes I used chatgpt but its my idea.
Rotum GIF
 
Man, that's a lazy OP even with ChatGPT. You can get it to do deep research, present proper arguments, get it to format some pro/con, have it make an opening statement , body, conclusion and so on.

Hell, get it to output to easily digestible slide deck! I mean put a modicum of effort!
 
if ai doesn't simplify, or better optimize, for real the code base every single hw evolution or "paradigm shift" is going to bring massive development time and costs. we already have enough power to make something wonderful, the problem is time needed and cost associated. This is also the problem with the kind of "safe" game design we see today, when there's a lot of money to be lost no one want to risk, back in the days a flop would be bad for the studio but was manageable, now a single flop can kill a proven studio. We need better hardware to manage the things that really free resources (like raytracing, that even if costly on the hw does really simplify game development, other than looking better than baked lights).
 
Think your own thoughts and write your own ideas... while you still can: https://www.media.mit.edu/publications/your-brain-on-chatgpt/

On topic, I read a while back about the efficiency gains you can get by incuding a scheduler in the CPU. I'm not sure how much of an effect it would have on games, or how legit the claims of a startup are, but they were talking up to 100x performance *in sime scenarios.* If it's that good, you can bet it will be rolled into hardware in the near future, especially with Moore's Law slowing down.
 
Last edited:
i1l26W3_d.webp


On the subject the HD generation ditched out of order instructions because it was not cost effective. The better you make a chip the more is going to take in yields. Although some people are chanting victory because MS is going to smash prices with 900€$ hardware that actually should cost 1.000€$ this is a market of people rising an eyebrow every time something crosses the 299 price tag. Ready with their keyboards already on fire. If, if, the advantages outclass the space in silicon, like with AI cores for machine learning, it's implemented. If not:
https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4aaee399-f14a-4f83-9ef5-eeacdf229d0c_619x238.png

Out with it!
 
Last edited:
The point of the asymmetrical design is to have some bigger CUs do bigger tasks and smaller ones do smaller tasks and therefor being a better optimized system. The task scheduler would use AI to distribute the workloads ofcourse human input would the vast majority of task delegation but some of the other stuff leaving it to the machine would save programing time.
 
Asymmetrical CPU cores and GPU Compute Units combined with an intelligent task scheduler
- Consoles have been using asymmetrical CPU cores for decades
- Unified GPU compute units weren't even a thing until XB360 (and PS3 was still asymmetric)
- Console OS has had task schedulers for about - 20-ish years
- Most games also do their own micro task scheduling(you can debate its intelligence - but the purpose is to optimize resource utilisation) - it's literally how games have been built for... around 45? years now.
 
Last edited:
The point of the asymmetrical design is to have some bigger CUs do bigger tasks and smaller ones do smaller tasks and therefor being a better optimized system. The task scheduler would use AI to distribute the workloads ofcourse human input would the vast majority of task delegation but some of the other stuff leaving it to the machine would save programing time.
No? The point of asymmetrical design is to have energy saving cores do the small stuff in order to save battery in idle or light task for computers and smartphones. Like the ones Apple uses. That is completely useless on a regular console and a waste of silicon that will better be used on another performant core in case workloads hit the CPU limit. Coincidentally workloads hitting the CPU limit is what we've been dealing with since the PS3 passed away.
 
Asymmetrical CPU cores and GPU Compute Units combined with an intelligent task scheduler, offer significant benefits for video game consoles by optimizing performance, power efficiency, and responsiveness. In this architecture, high-performance cores handle demanding tasks like physics calculations and AI, while smaller efficiency cores manage background operations such as audio processing or system services without wasting power. Similarly, GPU CUs can be designed with varying capabilities—some optimized for complex shading and others for simpler rendering or post-processing. A smart scheduler dynamically assigns workloads based on the nature and priority of each task, ensuring that resources are used efficiently. This enables smoother frame rates, reduced latency, and improved thermal performance, all while extending hardware lifespan and supporting more immersive and complex game worlds without requiring developers to micromanage hardware resources.

yes I used chatgpt but its my idea.
The current Xbox Series X / S, PlayStation 5 / 5 Pro, and Steam Deck use compact Zen 2 models i.e. they are not the full-fat desktop Zen 2 cores.

AMD's compact Zen cores are budget chip area cores with lower clock speed potential and reduced L3 cache.
 
i1l26W3_d.webp


On the subject the HD generation ditched out of order instructions because it was not cost effective. The better you make a chip the more is going to take in yields. Although some people are chanting victory because MS is going to smash prices with 900€$ hardware that actually should cost 1.000€$ this is a market of people rising an eyebrow every time something crosses the 299 price tag. Ready with their keyboards already on fire. If, if, the advantages outclass the space in silicon, like with AI cores for machine learning, it's implemented. If not:
https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4aaee399-f14a-4f83-9ef5-eeacdf229d0c_619x238.png

Out with it!
AMD BC250 has a PS5 APU with defects. PS5 Zen 2 is the compact model.
 
No? The point of asymmetrical design is to have energy saving cores do the small stuff in order to save battery in idle or light task for computers and smartphones. Like the ones Apple uses. That is completely useless on a regular console and a waste of silicon that will better be used on another performant core in case workloads hit the CPU limit. Coincidentally workloads hitting the CPU limit is what we've been dealing with since the PS3 passed away.
You need bigger CU for bigger workloads like advanced shaders and RT and smaller CUs for post processing effects and geometry processing, the fact that it saves power is an added bonus.
 
Top Bottom