By using this site, you agree to our Privacy Policy and our Terms of Use. Close
potato_hamster said:

Yes. Low powered Ryzen APUs exist. No one has argued otherwise. Do they have the right TDP and price to be used in a portable PS4? We don't know that. You believe that's possible, but from what I can see it hasn't been demonstrated.

And what kind of capacity would you like to have it demonstrated?
I can get some underclocking/undervolting graphs if that is your cup of tea.

And I most certainly believe it to be possible as that is my educated point of view. But I am more than happy to meet the requirement of evidence, I typically don't make a claim unless I am certain I can back it up. ;)

potato_hamster said:

Yes, obviously there are exceptions to the rule, but if they bear no relevance to today's technology and design principles, they're irrelevant aren't they? It's kinda like a biology teacher teaching their class that humans have 23 chromosomes, and someone interjecting and saying "well actually people with Down Syndrome have an extra chromosome, so that's false!". No, it's still true. While it's just not a hard rule to describe the DNA of every human out there, it still applies to over 99% of humans.

Irrelevant? No. Considering it was one of the very founding principles of the entire Core architecture that Intel tends to iterate on a yearly basis... And which dominates the high performance computing spaces.
Rambling on about Chromosomes and so on is what is truly irrelevant.

potato_hamster said:

My apologies, emulate was not the right word. I meant more "replicate". All I'm asking is an APU that can be packaged in a sub $400 handheld with 1080p screen, and 3+ hour battery. You keep asserting this can be done using custom version's of AMD's existing APUs, so I'm assuming there must be some company out there that is using one of these APUs in a sub $400 handheld with a decent screen resolution and battery life that gets similar performance to a PS4. If there isn't, are you just theorizing that this hypothetically could be done?

Replicate? That is also not required.
The only requirement is that they have a sufficient baseline of performance whilst being completely ISA compatible.
The rest falls onto the software stack.

There is a reason why the PC has been able to retain backwards compatible for decades, that's not just an assertion either.

potato_hamster said:

Is "pretty inexpensive" inexpensive enough to fit in a hypothetical BOM of $400 PS4 Portable?

If it can fit into my $100 AUD Gamepad/Tablet. Yes.

potato_hamster said:

Fair enough. With all that being said, do you think something like an AMD APU would be more expensive or less expensive on a 7nm process vs a 14nm process for example? If not, where do you think we're currently at in terms of the most cost effective fabrication process that would get us the TDP a mobile PS4 would need to have? If you can't answer something like that with confidence, that's fine.

It's hard for me or anyone to quantify any legitimate numbers as that kind of information isn't typically released into the public domain.
With that in mind... I would say that 14nm would be more cost effective in the short term for a mid-sized SoC, but 7nm most certainly has the long term pricing locked up.

Right now though, 12nm-10nm seems to have good enough power characteristics (As demonstrated by what Ryzen offers) to make a PS4 portable entirely feasible... With the correct binning of course.

potato_hamster said:

No of course it doesn't have to be 7nm. It just needs to be the right cost and have the right TDP to make sense in a PS4 portable. I think a lot of this conversation we're having is completely unnecessary because I wasn't arguing against the points that you were making, I was arguing against the points our friend Kev was making. He's the one that believes it's necessary for Sony to use an APU made using a 7nm fabrication process in order to get the TDP necessary for a PS4 portable. He's the one that thinks it would automatically be cheaper to fabricate chips using a 7nm fabrication process. He's the one that thinks Sony can produce a PS4 portable using this 7nm APU for $399 or less as soon as six months from now, and be able to play future PS5 games at a lower resolution/frame rate. I'm pointing out the holes in his nonsense, and you've more or less stepped in and argued on his behalf for some strange reason.

The chips need to be designed first. Then they need to go through multiple steppings/spins.
Then the layouts need to be optimized...

Even leveraging AMD's already established library's, it's still not going to happen within 6 months.

I will argue points I fundamentally disagree with, if it actually interests me.

--::{PC Gaming Master Race}::--