Playstation 5 [PS5] [Release November 12 2020]

don't know, but Carmack had developped new codecs to stream high quality videos with less bandwidth.

Codecs for streaming video and codecs for streaming real time content like frames to a VR headset differ greatly. I don’t think most video compression codecs would work well for VR as you need to buffer multiple frames to do the typical intra-frame compression and motion estimation which would introduce a bunch of frame lag.
 
I would be quite an upgrade if Sony could update their SSD to comply with zoned namespaces. It would possibly increase the available space by reducing overprovisioning and also boost lifespan ( which is possibly a concern with a soldered ssd)
From what I understand it just requires an ssd firmware update and Linux kernel 5.9 may support it with a driver update.
https://www.anandtech.com/show/15959/nvme-zoned-namespaces-explained

Doesn't this assume that they are accessing the device as normal block storage device and haven't implemented some other paradigm on there custom controller?.
 
How so?, this could be the exact reason that add on devices need more bandwidth to make up for the efficiencies lost by the other implementations.
I would have thought that having a unified way to access data (block storage or named spaces) , would be the easier path , but i am no software engineer. I mean it would reduce the number of software layers i guess?
 
Well now with the new SSD, they will be able to do a lot of new stuff. I can't imagine what they will come up with. Whatever it is, it's likely to be awesome.
 
Gofreak on RE posted this liquid metal patent from Sony, complete with method to insulate nearby components and keep the LM contained to the chip and heatsink interface.

https://patentscope.wipo.int/search...20162417&tab=FULLTEXT&_cid=P20-KDUB59-01124-1

Edit: apparently Steve from GN has known for a while.


Sounds good, but if it ends up sounding like a jet engine after a year your up the creek with higher performance options.

At least when everyone uses thoes naff sticky blocks for thermal transfer you knew you have a very easy win to improve the stock design.

Sounds like a kick ass teardown when it finally happens.
 
Gofreak on RE posted this liquid metal patent from Sony, complete with method to insulate nearby components and keep the LM contained to the chip and heatsink interface.

https://patentscope.wipo.int/search...20162417&tab=FULLTEXT&_cid=P20-KDUB59-01124-1

Edit: apparently Steve from GN has known for a while.

If this is used, I'm curious if the full reasoning for the decision would be disclosed.
The thermal conductivity is higher versus a more conventional thermal compound, although there are other AMD products with high thermal output at 7nm that seem like they get by with standard TIM.
Liquid metal could reduce the needed dimensions of the cooler, potentially give some additional tolerance in terms of interface thickness, better long-term thermal performance versus grease, or may allow for the power management algorithm to assume a more optimistic baseline for heat transfer out of hot spots.
The downside does appear to be the additional complexity in isolating the compound from electrical components, and different measures against oxidation or reacting with metal components like the heatsink.
It seems like there was a fair amount of variability in compound application and heatsink mounting with the PS4, although if Sony took this jump perhaps there were additional pressures like hot spots being a more acute concern than before.
 
The downside does appear to be the additional complexity in isolating the compound from electrical components, and different measures against oxidation or reacting with metal components like the heatsink.
Agreed, this doesn't make a whole lot of sense for PS5 based on the conventional alternatives, which are known quantities.
 
Agreed, this doesn't make a whole lot of sense for PS5 based on the conventional alternatives, which are known quantities.
The patent is broad enough to cover a lot of bases in terms of the material or how it's applied.
In terms of how AMD GPUs and Sony's consoles have had spotty results in terms of mounting pressure and grease application, an interface material that could give good enough performance even if their application skills do not improve might be tempting.

One scenario is a metal compound solid at room temperature that melts during operation, which might allow applying the material as a pad like the phase change pads on some coolers, or like the graphite thermal pad used on the Radeon VII.
It would combine a more readily positioned and applied layer with at least some of the performance benefits of the metal.
Unfortunately, the references I've seen for this sort of application seem to have come from the manufacturers of such materials, and I didn't see reference to an application with the chip size and production rate like a console APU.
Given the extra effort, is a fancy TIM a better solution than improving some uninspired cooler assembly processes? If there are other reasons that are more pressing, like hot spots or thermal density, it might be considered an neat expenditure of creative engineering for a design corner that seems like a mistake to find oneself in.

Gallium would be an example of a pure element that would have the desired melting point, although chemical and cost considerations intrude. It'd be interesting to see what sort of warranty sticker Sony would need to absolve itself of responsibility with a TIM that could potentially disintegrate parts of the tools or setup of a user trying to service their console.
 
Agreed, this doesn't make a whole lot of sense for PS5 based on the conventional alternatives, which are known quantities.
The factors at play are known quantities. They will have material scientists and mechanical/reliability engineers that will understand all the risk factors.

The actual challenge is scaling to manufacturing millions of units and dealing with those tolerances, as well as proving the lifetime reliability of the system with sufficient sample sizes and test cycles. They will likely need a stringent article inspection schedule, and perhaps periodic production sampling to ensure their vendors and manufacturers are holding the tolerances to drawing specifications.

The PC space is a veritable Wild West comparatively. Heatsink and heat compound manufacturers don’t coordinate to guarantee material composition, purity or consistency, provide compatibility lists, have a way to guarantee consistent clamping force, have no control over die or heatsink surface area, can’t control orientation, and the list goes on.
 
Last edited:
Back
Top