how did you get that amount? All I've found is there are 2 simd-32s in one CU.3584 ops/clk/simd32
there are 64 simd-32s in a CU.
3584 / 64 = 56
This hint?
how did you get that amount? All I've found is there are 2 simd-32s in one CU.3584 ops/clk/simd32
there are 64 simd-32s in a CU.
3584 / 64 = 56
This hint?
If the X is 12 tf, there is no chance it is going to be under £500 in the UK, particularly with Brexit. It could even be more than that depending on how Lockheart is priced. Mostly likely is probably £300 and £500 I reckon with tales from my ass. Even at that, won't be surprised if they are selling at a decent loss. Otherwise, they might go all in for subscription models.
At those specs, no way PS5 can go for above £400.
You are 100% sure so you are amd engeenieer working on both socs for nextgenconsole ? :d
I agree except the last sentence. No way could/would Sony come in at/above £400 as the only other time they did briefly was with PS3 before getting a dose of smelling salts and pushing a £299 revision out the door 6 months after launch. Like you say this time we're in uncertain economic times too.
It will be interesting to see if Microsoft do bring Lockhart at launch. I'm thinking maybe not because if we say Lockhart is £249-£299, PS5 is £349 and XSX is £500+ then many could simply deduce PS5 is the Goldilocks model to get? Microsoft might be better to just launch XSX and take a bigger hit on price to get to around a £100 of PS5?
This is all probably moot because there are almost certainly going be twists and turns to come...Probably starting today at some point!
The Series X isn't going to be 12TF RDNA.
Eventually folks will come to accept that. Better now than the crushing dissapointment of later.
It's funny, over a year ago, 9-10TF RDNA was seen as the high end goal here. And now this thread has skipped past it and expects AMD to create a console level GPU that has 2TF more than their fastest Navi card and put it in a $500 console, all in less than a year.
Talk about absurdly whack expectations. For a tech focused thread, that's something...
The same way you guys are 100% sure they will be 12 tflops :D and says it is confirmed :DYou are 100% sure so you are amd engeenieer working on both socs for nextgenconsole ? :d
Blame Tom Warren, Windows Central and Digital Foundry then as they have all stated that they believe it's a 12TF RDNA console.
Richard from DF made a video that show 9TF is enough to be x2 the One X.I suppose on the other hand, just over 10 tf would align with the specs being similar and 2 x Xbone X in actual performance. Also might allow MS to deliver the console at a price close to PS5.
Meaning?You've considered all possibilities except one : the integrity, timeline and provenance of the GitHub document.
Df didn't they asked Ms and didn't have an answer.Blame Tom Warren, Windows Central and Digital Foundry then as they have all stated that they believe it's a 12TF RDNA console.
Blame Tom Warren, Windows Central and Digital Foundry then as they have all stated that they believe it's a 12TF RDNA console.
£350 for PS5 would be great value, but I would be surprised if they could get it with those specs without taking a loss. I really don't think MS can achieve £249 with Lockhart unless they take a big loss. At 12tf I reckon £500 is the cheapest they can go with X, unless they take a big loss.
I'm not this guy but for sure many leaks and words hints 12tfThe same way you guys are 100% sure they will be 12 tflops :D and says it is confirmed :D
If he did, we won't be having this outbreakJust a question, did Phil Spencer say / confirm the 12 tf number anywhere?
Just a question, did Phil Spencer say / confirm the 12 tf number anywhere?
I have posted this multiple times but belief is not a confirmation (DF has not confirmed it for example). And hey, it's not like insiders haven't been wrong before.
Wouldn't be the first time, wouldn't be the last. I trust Phil's specific word usage more than I do reports where it isn't even abundantly clear that they know what the difference between RDNA and GCN is. Phil said twice the performance of the XB1X GPU. That's about 10TF RDNA.
But mostly, jumping from a 9.75 TF PC GPU to a 12TF console GPU in the span of a year, all aimed at a $500 console sounds like the thing fan drama are made of.
You've considered all possibilities except one : the integrity, timeline and provenance of the GitHub document.
The Series X isn't going to be 12TF RDNA.
Eventually folks will come to accept that. Better now than the crushing dissapointment of later.
It's funny, over a year ago, 9-10TF RDNA was seen as the high end goal here. And now this thread has skipped past it and expects AMD to create a console level GPU that has 2TF more than their fastest Navi card and put it in a $500 console, all in less than a year.
Talk about absurdly whack expectations. For a tech focused thread, that's something...
And is no one concerned about this?!Files are from April - June 2019, haven't found any timestamps for tests in the files.
9-10 RDNA for both is good enough. I would rather pay 399 than 499.
Is it possible that windowscentral took the twice as X1X and came to the conclusion of 12 Tf?No. He even dodge answering that question with asked by DF for clarification. MS only confirmed twice Xbox X GPU
For the third time: Arden has 56 CUs (3584 Shaders) in the internal AMD testing documents --> 1,7 GHz x 3584 x 2 = 12,1 TFlops
Oberon has 36 CUs and 2 GHz
For me a leak must correspond to two things :
Mentions doesn't mean anything about document provenance, edition and timeline. I could find these on Linkendin.The documents are legit, there are personal informations relating to AMD employees there.
Interesting...
Mentions doesn't mean anything about document provenance, timeline ans integritiy, I could find these on Linkendin.
Well, i wont post email-adresses, IPs and user infos, but this stuff all checks out.
Those arent offical documents for anyone other than internal employees to see.
Whatever as I said, all these mentions could be true, as information on it, that doesn't say anything about the document itself.Well, i wont post email-adresses, IPs and user infos, but this stuff all checks out.
Fair enough. So I suppose 12tf is still up in the air. This may simmer about until E3...
Whatever as I said, all these mentions could be true, as information on it, that doesn't say anything about the document itself.
I am really curious about the cooling system for 2GHz GPU in the retail console box!Probably, but i doubt they will reach 2 GHz like Oberon, so this means that Oberon will have better pixel/geometry rates than Arden.
Duke, i called it brother. 56 cu dream came true. What a time to be alive.3584 ops/clk/simd32
there are 64 simd-32s in a CU.
3584 / 64 = 56
Df didn't they asked Ms and didn't have an answer.
They showed 9.7 tflops doubles a Xbox one x
They showed 10.2 tflops is 8x Xbox one
They showed 11 tflops is 8x Xbox one s
And they always said 12 tflops is a target.
There is no guarantee a console will hit a target.
The simple fact that no 3rd party devs cared to use the extra power of xox over the ps4p should tell you what will happen if ps5 is 9tf and xsx is12tf. Games will be designed around 9tf limits. So if ps5 is cheaper you will be wasting money on the xsx power.
I am not stating anything about it, I am just calling to follow elementary rules of prudence when looking at it, because during almost 24 hours I read people seriously talk about "drivers", believing that a complete array of APU prototypes will end in Catalayst.Its not a "document", these are results in csv/excel with Python Scripts.
This is nothing official or for anybody other than internal employees to see.
True. Good thing MS has now more than a dozen of first party studios.The simple fact that no 3rd party devs cared to use the extra power of xox over the ps4p should tell you what will happen if ps5 is 9tf and xsx is12tf. Games will be designed around 9tf limits. So if ps5 is cheaper you will be wasting money on the xsx power.