Playstation 5 [PS5] [Release November 12 2020]

PS5 is so big that there is a chance it will have an [almost] passive cooling mode for UI & base PS4 BC clocks, and very slow fan speed for everything else. Since the console is created around fixed max power output, they will never get surprised by the heat that APU can create.
 
Software disabled circuits can still leak and therefore consume a lot of power, so no, not the same.
If AMD's on-die power network isn't configured to isolate individual CUs, it might not matter when the CUs are disabled. If there is individual CU/WGP power gating, which I'm not sure is done by AMD, then the disabled CUs would be close enough to zero or as close as they can be. Absent a physical breaking of the lines feeding the CUs at the factory, even the dead CUs would still be powered to some extent. I've seen fuses discussed used for configuration changes, but not for wires used for power delivery. Physically breaking the connections with something like a laser seems impractical as well.

would't that make it easy to topple over if it's top heavy?
If it's centrifugal, the heatsink assembly could be closer to the midline. The fan and shroud should be outweighed by the metal in the cooler and power supply. Other components might cluster nearer the middle or bottom as well, like the optical drive and metal caging around the electronics. If the fan is at the top, that portion of the box's volume may have a decent amount of room just for air flow or plastic shrouding.
 
If the fan is at the top, that portion of the box's volume may have a decent amount of room just for air flow or plastic shrouding.
If you think about the design, the CPU cooler(whatever it is) will likely be in the middle(centrifugal) or towards the bottom. I think it's in the middle so that there's a balance to the system. The way the stand is built on the DE, leads me to believe this is the design. You might say but what about the disk version. Well the Bluray weighs differently to the DE version hence the shape of the stand, which is also built around the shape of the Bluray to balance it.
 
would't that make it easy to topple over if it's top heavy?
Hmm I guess having the power supply, heat sink and optionally the disc drive located at the lower half should keep the centre of mass low? And the fan should be made of plastic. Just my speculation!
 
If AMD's on-die power network isn't configured to isolate individual CUs, it might not matter when the CUs are disabled. If there is individual CU/WGP power gating, which I'm not sure is done by AMD, then the disabled CUs would be close enough to zero or as close as they can be. Absent a physical breaking of the lines feeding the CUs at the factory, even the dead CUs would still be powered to some extent. I've seen fuses discussed used for configuration changes, but not for wires used for power delivery. Physically breaking the connections with something like a laser seems impractical as well.
Ah, didn't know power gating reduces the overall power draw to near zero. For GPU power saving I guess it's more practical to reduce frequency and voltage? Is there any GPU known to power gate cores for power saving?
I think it may be practical for them to test each die on the silicon "assembly" line, then laser cut it.
 
Hmm I guess having the power supply, heat sink and optionally the disc drive located at the lower half should keep the centre of mass low? And the fan should be made of plastic. Just my speculation!

It can also be used the other orientation, so it can't be too fat bottomed or top heavy.
 
It can also be used the other orientation, so it can't be too fat bottomed or top heavy.
Indeed!

Perhaps something similar to One X's?
Pop05Ap.jpg


But with a very different proportion (i.e. the fan and heatsink is much larger/ taller).
 
Hmm I guess having the power supply, heat sink and optionally the disc drive located at the lower half should keep the centre of mass low? And the fan should be made of plastic. Just my speculation!
Yea you’re probably right; I just assumed heat sink and fan assembly are combined when they can be separated
 
Software disabled circuits can still leak and therefore consume a lot of power, so no, not the same.

I don't speak about power consumption. I speak about people saying 36 CUs is linked to backward compatibility. They could have gone for 40 CUs and disable four if is needed fo backward compatibility. They did not go this way because for the cost and power they target 36 CUs is ideal.

And it seems 2.23 Ghz is not so high when we see AMD has a 40 CUs at 2.5 Ghz.;)
 
With how huge it is, it'll be hilarious if the quietness will still be a lottery like PS4 and PS4 pro due to Sony using two fan suppliers; loud delta fan and quiet nidec fan.

Btw I just turned on my Xbox 360 again after left unused for eons. Whoa, this thing is quiet! (but the DVD drive is loud as heck)
 
I don't speak about power consumption. I speak about people saying 36 CUs is linked to backward compatibility. They could have gone for 40 CUs and disable four if is needed fo backward compatibility. They did not go this way because for the cost and power they target 36 CUs is ideal.

And it seems 2.23 Ghz is not so high when we see AMD has a 40 CUs at 2.5 Ghz.;)
And I didn't speak about what Sony wanted, I spoke about how disabling CUs in driver is different from disabling them in the factory.

Also Mr Cerny said 2.23 GHz was the limit before the logics go haywire, but how did AMD reach 2.5 GHz? Did they reserve the best for themselves or was PS5's GPU not truly RDNA 2? :???:
 
And I didn't speak about what Sony wanted, I spoke about how disabling CUs in driver is different from disabling them in the factory.

Also Mr Cerny said 2.23 GHz was the limit before the logics go haywire, but how did AMD reach 2.5 GHz? Did they reserve the best for themselves or was PS5's GPU not truly RDNA 2? :???:

Or maybe more logic this is the limit for the power envelope they target.

https://forum.beyond3d.com/posts/2160211/

Here RDNA 2 leak from a ROCM in Mac os
 
Last edited:
And I didn't speak about what Sony wanted, I spoke about how disabling CUs in driver is different from disabling them in the factory.

Also Mr Cerny said 2.23 GHz was the limit before the logics go haywire, but how did AMD reach 2.5 GHz? Did they reserve the best for themselves or was PS5's GPU not truly RDNA 2? :???:
What about XSX then? Because 1.825 ghz seems like a typical RDNA1 clock. Anyways, I want to see the receipts for that supposed 2.5ghz clock.
 
Back
Top