need some time to pull my thoughts together on this one.
Not exactly. Oberon B0 was a physical device. We have the IP addresses of 3 units. But it does point to that Oberon A0 was a c.model. And possibly Ariel A0, maybe Ariel B0.
Now that is a lot to think about.
That was my impression, given how it was labeled in the file. I don't know for certain.Oh I see, I always thought those IP adresses were tied to machines where testing was done, which would not prevent a test on some model being done on them and the IP still showing on the data.
That was my impression, given how it was labeled in the file. I don't know for certain.
Another implication is that all those test results that were found, including the >20k firescore were all model results.
It would also indicate Oberon B0 in the V dev-kit.
Months ago when we first heard about it. It's been damaged goods for a while IMO.
need some time to pull my thoughts together on this one.
Not exactly. Oberon B0 was a physical device. We have the IP addresses of 3 units. But it does point to that Oberon A0 was a c.model. And possibly Ariel A0, maybe Ariel B0.
There are other c.models in GitHub and we know from AMD FAD that they were using these for rapid prototyping.
Now that is a lot to think about.
Edit: I am not at my desk so I can't back up or debunk these claims.
Yes.Hmmm...just speculating from the surface, but would it be possible for Oberon B0 to be a model revision of a virtual device?
It just seems odd for a revision of a virtual device to be a final chip in a console. I would think they would rename it. Maybe they did.
Hmmm...just speculating from the surface, but would it be possible for Oberon B0 to be a model revision of a virtual device?
Possible but unlikely as we found IP addresses for three physical units named Oberon 1, 2, 3...
Why do IP addresses need to be associated to physical devices and not virtual machines?Possible but unlikely as we found IP addresses for three physical units named Oberon 1, 2, 3...
And what was in the dev kit? And tested in fire strike? No sure about anything at this point,except that PS5 is going to be amazing.
I've heard several times now that development has been a bit rough :-/
Hopefully it all comes together, I love Halo and I like the direction the game seems to be going in.
Sort of like an Uncharted 4 situation. Continuing creative differences that lead to other issues.
But that's just what I've heard, and like I said, that's not unusual for Halo and it's generally worked out in the end.
And if it makes you feel any better, I've heard quite the opposite for a few other projects.
My first thought as well.Why do IP addresses need to be associated to physical devices and not virtual machines?
Not required, but none of the c.models had the same treatment.Why do IP addresses need to be associated to physical devices and not virtual machines?
We understand it is speculation. I am not at my desk and am going by memory at the moment. A few of the inconsistencies he mentioned we had found in the past as well, and did not seriously consider as it did not Fit well with the thinking at the time.Instead of speculating what the guy Randy is posting on twitter I never understand why nobody has the idea to challenge/ask those guys on twitter directly.
I asked (again) and guess what, he is speculating. It is not a claim, it is his theory. And such kind of theories are likely to be not accurate, just like the last one just in the recent days.
To III-V BlacknGoldBlood wowbaggerBR
Instead of speculating what the guy Randy is posting on twitter I never understand why nobody has the idea to challenge/ask those guys on twitter directly.
I asked (again) and guess what, he is speculating. It is not a claim, it is his theory. And such kind of theories are likely to be not accurate, just like the last one just in the recent days.
To III-V BlacknGoldBlood wowbaggerBR
Edit:
Randy @Carthemanbra
Replying to @WrexTheNerd
This is just speculation, obviously, ...
You don't have to use those kind of words.people just want any information or scoop they can get. Even if its just from some goober on twitter.
the last one ... really not necessary imho ...
There you go: "a random person online" :)i dont think its offensive? But anyway its just a light hearted term for a silly person. Im just saying i think people will take any "news" they can get with how quiet everyone is being. Even if its just from a random person online.
Doesn't that also pretty much extend to everyone on this very forum?
I'd like to hear some people more familiar with hardware design/testing/manufacturing speak to what the purpose/usefulness of this would be opposed to how they did it in the past or how other companies are currently doing things.
Also: have there been other things virtualized and then tested like this before? anything else in Github from those cases?
"Virtualized stepping" is most likely running synthesized logic on FPGA. This lets them to test out the soundness of their logic design without the huge cost of actually manufacturing silicon.
This is the general flow of silicon design. It doesn't matter whether the test is conducted on actual silicon or FPGA, we knew the high-level design is done, as we saw in the sheets that theoretical values and test patterns are all set. If the virtual stepping of Oberon A0 is indicative of anything, I'd say AMD is committed to the specifications!
States are now shutting down restaurant businesses and bars to curtail the spread of COVID-19. Retail can't be far behind. Thinking any sort of announcement is coming from Sony any time soon under these conditions is wildly optimistic.
Yes, this let them find bugs and design flaws early, which, regarding the size of their projects, is far less costly than making a new silicon to test for every revision. But it has its limitation, testing on real silicon is still absolutely necessary, as the manufacturing process will determine the electrical performance characteristics of the silicon, which in turn affects final performance (e.g. how conductive are the physical transistors → how high can frequency go).Nice! Thanks for the reply and the info.
Is this something a lot of companies do? Seems like it would save time and money, but I don't know how accurate it would be.
The same in The Netherlands... Restaurants (at least the physical ones where people can sit down), nurseries, schools and all kinds of gathering places like gyms, sport clubs en stuff. Even libraries are shutting down. So yeah fun times..... and idiots are stocking up on stuff without it being needed and only bring problems with it. There is enough food and products..... for weeks and weeks on end.States are now shutting down restaurant businesses and bars to curtail the spread of COVID-19. Retail can't be far behind. Thinking any sort of announcement is coming from Sony any time soon under these conditions is wildly optimistic.
Yes, this let them find bugs and design flaws early, which, regarding the size of their projects, is far less costly than making a new silicon to test for every revision. But it has its limitation, testing on real silicon is still absolutely necessary, as the manufacturing process will determine the electrical performance characteristics of the silicon, which in turn affects final performance (e.g. how conductive are the physical transistors → how high can frequency go).
Not required, but none of the c.models had the same treatment.
We understand it is speculation. I am not at my desk and am going by memory at the moment. A few of the inconsistencies he mentioned we had found in the past as well, and did not seriously consider as it did not Fit well with the thinking at the time.
A large part of this theory is tied to recent finding, that OberonA0 was associated with CPU-ID 8F40. However, I believe that Ariel_B0 is associated with 8F10 in GitHub, which works with our current understanding.
Now I am not in agreement with him yet, will need to look over my notes tomorrow. At this point I am thinking OberonA0 may have been c.model only.
It will never die... Even when Sony comes out and says it's a 14tf system ( :) ) people will still call it a <9.2tf system.... Look at the previous OT... It always comes down to <9.2tf hahah even before the "mighty github"..... Its just the way of life :D...
Only GitHub revisitParrigatooooo (@isaparrot) | Twitter
The latest Tweets from Parrigatooooo (@isaparrot). Opinions are my own, Parrots and feathered creatures welcome. Seattle, WAtwitter.com
Don't know his source (github leak or another files), but also interesting.
Hardware designing is really similar to software. Instead of laying out components on a blueprint by hand, you code in Hardware Description Language like Verilog or VHDL, then you "compile" (synthesize) your HDL code into some form of gate level description, which can then be burnt onto a FPGA for testing or put into other CAD softwares for, e.g. floor planning, all before a mask is made and sent to foundries like TSMC. Pure software simulation also exists, though they are obviously slower than FPGAs.I see.
So they can virtualize some stuff and check the results, and then make a physical iteration for testing, then virtualize some changes to that, and sort of get it closer and closer to the final product before finalizing it and signing off on it?
I think I'm becoming so starved for PS5 news that I'm starting to become more interested in a lot of the tech stuff within the hardware industry now.
Changes nothing as far as I can tell given B0 is a real chip afaik
Indeed. That is why I thought the wording "goober" strange.Doesn't that also pretty much extend to everyone on this very forum?
I'm still betting on a Feb reveal
I think we should all get our avatars back. It's not like we ever gave Transistor a year prediction.
I am. Just that my twitter handle doesn't match my name here.Instead of speculating what the guy Randy is posting on twitter I never understand why nobody has the idea to challenge/ask those guys on twitter directly.
I asked (again) and guess what, he is speculating. It is not a claim, it is his theory. And such kind of theories are likely to be not accurate, just like the last one just in the recent days.
To III-V BlacknGoldBlood wowbaggerBR
Edit:
Randy @Carthemanbra
Replying to @WrexTheNerd
This is just speculation, obviously, ...