Well guessing often is wrong.I don't mean insiders. I mean people who have no problem telling all of twitter and gaf/era that they prefer xbox.
There were a couple write ups showing what can fit in a ~400mm2 die. The limit on 7nm was 56/60 cu accounting for lowered l3 cache.
7nm+ is supposed to bring a 20% density increase.
https://www.resetera.com/threads/ne...arks-for-chance-at-glory.165711/post-28427929
Baseless Next Generation Rumors with no Technical Merits [post E3 2019, pre GDC 2020] [XBSX, PS5]
- Thread starter BRiT
- Start date
- Status
No one ever described it as a decimal form architecture. Mostly because that's not a real thing.
It could mean that certain RT blocks are optional like PowerVR’s Coherency Engine. The CE block is not a fundamental part of Imagination’s RT solution and licensees can choose not to license that particular solution. It may be the case that both MS and Sony excluded certain blocks because they value more general compute than making space for hardware limited to RT tasks.
upnorthsox
Veteran
1.81 which also matches the clock speed.
MS: "Can we say it's based on rdna2?"
AMD: "Sure"
Sony: "Can we say it's based on rdna2?"
AMD: "Sure"
Whatever they altered, optimized, or even removed doesn't change what AMD calls the architecture. Both will be customized in some way with secret sauce, maybe even removing stuff that are less relevant on consoles. And adding stuff like LZ and Jpg decompressors on the amiga blitter, ID buffer, etc...
The 1.9 reference seems to be like calling a Zen2 cpu with less cache, or some vector instructions missing, "it's kind of Zen1.9 so to speak, or Zen2 lite". We can understand what is meant without trying to argue whether or not Zen1.9 is an actual registered trademark protected by a legal denomination of origin.
AMD: "Sure"
Sony: "Can we say it's based on rdna2?"
AMD: "Sure"
Whatever they altered, optimized, or even removed doesn't change what AMD calls the architecture. Both will be customized in some way with secret sauce, maybe even removing stuff that are less relevant on consoles. And adding stuff like LZ and Jpg decompressors on the amiga blitter, ID buffer, etc...
The 1.9 reference seems to be like calling a Zen2 cpu with less cache, or some vector instructions missing, "it's kind of Zen1.9 so to speak, or Zen2 lite". We can understand what is meant without trying to argue whether or not Zen1.9 is an actual registered trademark protected by a legal denomination of origin.
It's about the base architecture.
If you started with RDNA 1 as the base and pulled features from RDNA 2, then it would be RDNA 1+
If the base was RDNA 2 and you removed/cut things down (e.g. cache size) then it would be RDNA 2 derived.
Depending on what the differences between RDNA 1 & 2 are will determine how big a difference it makes actually is.
But it is legit to call it 1.5 or 2 etc.
If you started with RDNA 1 as the base and pulled features from RDNA 2, then it would be RDNA 1+
If the base was RDNA 2 and you removed/cut things down (e.g. cache size) then it would be RDNA 2 derived.
Depending on what the differences between RDNA 1 & 2 are will determine how big a difference it makes actually is.
But it is legit to call it 1.5 or 2 etc.
That's RDNA 'the last decimal of pi'What if RDNA1 with added features and RDNA2 with removed features are the exact same?
D
Deleted member 13524
Guest
Then he's very clearly not suggesting AMD calls it 1.9.
"Both systems are rdna 2, but not. They are essentially rdna 1.9 so to speak, with features bolted on, meaning that you can call it the next generation of rdna, you could even call it rdna 2 if you wanted. I wouldn’t myself but that seems to be what they are doing so ok."
How is this any different from me saying "I myself wouldn't call KBL-G a Vega, but that seems to be what they're doing so ok"?
RDNA√(-2)What if RDNA1 with added features and RDNA2 with removed features are the exact same?
Cause how does one at his position even qualify that statement. Without knowing what MS or Sony has done at the architecture level in comparison to a base RDNA 2 that no one has this level of information on.
and without any shred of evidence is there any reason for us to believe him.
Yeah it gets fuzzy since we're into hypotheticals.
It's not like GCN vs RDNA. If we assume RDNA is an evolving architecture with a big list of independent improvements, the designation for which series of upgrades they put in the rdna2 folder or the rdna3 folder might be arbitrary, and based on their own gpu plans regardless of semi-custom contracts.
If there's a specific feature which is not ready on time for ps5/xsx but is ready on time for their own first gpu they officially call rdna2, is it really different from removing that feature to save die space? If the semi-custom design gets all except one of the rdna2 improvement list, the distinction doesn't matter.
I assume both semi-custom designs and the upcoming rdna2-designated retail gpu were all developped in parallel, and implementing the same new features from that rdna2 list of improvement. The work being done once by AMD and applying it to all three implementations?
Could you end up at the same place, sure.
But the starting point is different.
I've never heard anyone claim 4pro is Vega minus some bits. It's southern islands (think) + Vega bits.
Until we know what the difference between RDNA 1 and 2 is we don't know how much difference it actually is.
Either way you can't blame MS unless their lying about what the base design is from saying RDNA 2.
But the starting point is different.
I've never heard anyone claim 4pro is Vega minus some bits. It's southern islands (think) + Vega bits.
Until we know what the difference between RDNA 1 and 2 is we don't know how much difference it actually is.
Either way you can't blame MS unless their lying about what the base design is from saying RDNA 2.
Personally I don't think trying to interpret what he means by 'RDNA 1.9' even worth the keyboard wear-and-tear...
D
Deleted member 13524
Guest
How do I, or any anonymous account in a forum, qualify for saying I think KBL-G shouldn't be called Vega?
I don't. Neither does he.
It's the internet, people give opinions over matters they haven't presented evidence they're qualified to all the the time..
Yes, but why not RDNA 1.7?
j/k sorry
Odium is apparently verified as a..."QA tester". I mean are we just ignoring that? That's a game tester=low level lackey in my book. But apparently some people just go "oh, he's got the verified stamp, nobody dare question 100% legit!". What's next, verifying gamestop clerks?
Again, of course. Low level lackeys are the only ones going to post repeatedly on Neogaf/resetera and break NDA's right and left.
Imagine how seriously sued, or more likely fired I guess, any "real" insider would be in danger of being. Including having to show identifying info to the mods to get verified.
- Status
Similar threads
- Locked
