drkohler said:
ok. Explain how you got from point 1 to point 2. Give sources to point 2 as well. I really want to know that. |
You missed my point completely I see.
| Lafiel said: only if at the end they have more chips that pass the quality check per wafer - "super high clock rate" surely has a significant influence on how many of the chips per wafer are substandard for use in scorpio |
Clock Rates aren't really directly attributed to yields.
Chips are designed to operate at a certain frequency at any given voltage.
Some chips can hit a lower voltage, others need higher, so what do you do? You be conservative so you can have as many chips as possible meet a certain clockrate.
Of course you also have different types of fabrication processes which allow for transisters to be ultra low power and a moderate clock, the opposite is true for that too of course.
The caveat though to using a chip with a higher clockrate is going to be voltage, not yields... And voltage has a direct relationship with power consumption and the heat generated by the chip as well.
But that can also be mitigated with a better cooling system.
The other issue is electromigration, but being only console levels of performance, won't be much of a drama anyway.
| DonFerrari said: It's kinda of pointless to arguee on rumours putting all good points in favor of one side and the other having only negatives. You aren't talking about a big company and a nobody (so that you could apply that only one of them know what they are doing). We already have the processing capacity and architeture of Scorpio and the rumoured for Neo. If both uses the same architeture and Scorpio is an overclocked version and Neo an undercloked one of the same chip then either one is smaller than the other, or will have more failures to produce and consume more power. So unless you think MS system will run on unicorn rainbow and Sony on pigs fart them you can't seriously give MS smaller chip, higher processing, less consumption, higher die output success and cheaper to produce. It doesn't make sense in any way. |
I'm just giving plausable possible scenario's, nothing that's based on legitimate information on these upcoming devices anyhow.
I have no doubt that the Neo will actually be a slower device with a lower price point than Scorpio, unless Microsoft is willing to eat the costs to shift more units.
| DonFerrari said: And worse yet you want MS chip to be totally different than Sony and still use improvements on the fabrication process... are you that pushing for a wish situation that MS put something magical when they come from RROD and a oversized console? |
That's not what I am getting at all.
| DonFerrari said: no, not all improvements apply to everyone. The chips may be similar, but there will be differences in setup and configuration, so not all improvements will carry over. And considering bulk production and cost of setup if Sony have 1 year lead to production and keep leading after Scorpio release Sony will always have a cheaper production on similar architeture. I respect your knowledge on computation, but your scenario in this one is so one sided that would only work in the bizarro world and on "it could happen impossible situation". |
If the chips are designed by the same company (AMD) uses the same licensed technology (x86, HyperTransport, GCN etc'.), all fabbed at the same place (Global Foundries) then when it comes to the actual fabrication of the chips, then improvements will carry over.
Let's say Global foundries does a tweak to it's process which allows for better power characteristics or better patterning due to learning how the process works with the Neo, then that will apply to Microsoft's SoC as well. They are all using the same tools.
Of course that works in the reverse as well. If Global Foundries somehow manages to optimise it's fabrication whilst fabbing Scorpio, chances are it will also benefit Neo.
Of course Sony and Microsoft petition AMD for their own various touches and changes and such to the designs, which will have an impact... But overall, whoever has the smaller chip will have the cheaper manufacturing, to a point. - In some cases building a big chip on an old process can be cheaper than having it on a new process as an idle fab is wasted money and retooling a fab for a new process can cost billions, hence why Intel historically kept chipsets/eDRAM at an older fab, but CPU's on the latest to prolong the old fabs life, but that's not relevent when both chips are using the same process.
Also remember, Sony and Microsoft by law are not allowed to deal with the production of these chips.
--::{PC Gaming Master Race}::--
