bitbydeath
Member
Either wishful thinking or they know something we don’t due to the level of data they produced. Leaning more towards the first though.
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
- Status
Either wishful thinking or they know something we don’t due to the level of data they produced. Leaning more towards the first though.
Negotiator
Member
mckmas8808
Mckmaster uses MasterCard to buy Slave drives
Well...
And now it's posted publically.
ethomaz
Banned
You know when a "insider" says "do not post publically" it means "why did you not posted yet?"
Last edited:
DemonCleaner
Member
that are overclocked results. i wrote a whole lot on that topic few pages ago,
stock firestrike combined results:
Chronos24
Member
Nice my 1080ti still hanging near the topthat are overclocked results. i wrote a whole lot on that topic few pages ago,
stock firestrike combined results:
ArabianPrynce
Member
Lol didn’t know you have inside connections. Does he the know the specs of next gen???Well...
Aceofspades
Banned
that are overclocked results. i wrote a whole lot on that topic few pages ago,
stock firestrike combined results:
This makes the Gonzalo leak score very impressive tbh.
Bryank75
Banned
My local GameStop in a small town in west of Ireland has 10 preorders for PS5 already, all over a hundred euro. Kinda pissed that I’m not first on the list.... ha
Reassuring numbers above in any case, with the memory speed and hardware based raytracing, we should see some really cool things!
Reassuring numbers above in any case, with the memory speed and hardware based raytracing, we should see some really cool things!
ArabianPrynce
Member
people on era are hyping up the 20k. they say its close to 2070 and close to 2080
ArabianPrynce
Member
wait you mean the gonzaloThis is PS4Pro...
Seems all Sony console chips start with DG1XXX, while MS are DG3/4/5.
Nope, PS4Pro retail SOC. I think there was revision DG14XX.
What we are missing is DG15XX which might not exist, or its actually PS4 Superslim that was rumored. We'll see.
ResilientBanana
Member
This was not the reason they didn't go Zen. It's been stated that it was a different architecture and that compatibility with Xbox One games would be an issue.Well...
It's starting to become clear, Microsoft and Sony have the same APU with different customizations. There will be minimal difference, each with their own storage technologies.
Last edited:
LordOfChaos
Member
ArabianPrynce
Member
What significant points did you take away from the video?
ethomaz
Banned
Well the guy said 5k is PS4 and 20k is PS5 but he didn’t say what benchmark.
5k to 20k is pretty small jump imo.
ResilientBanana
Member
4 x a lot is a lot more.
SonGoku
Member
Maybe apisak faked the numbers based on previous known numbering scheme
It does? i see the RTX 2070 above in that bench and Radeon 7 is faster than the 2070
Bigger chips clocked lower are more power efficient
LordOfChaos
Member
The benchmarks against the 2070 were against the Founders Edition for one, and they say they're trying to take a more honest approach to performance expectations than past AMD. They said they wanted to bring ray tracing when it's ready for everyone, and the current focus was competing with a 599 card and bringing that down to 450.
Some architectural details on Ryzen too mostly on Winodws scheduling, taking the clock picking latency down from 30ms to 1-2, cross CCX performance fixes in compilers, etc, I assume consoles are already well aware of all that
Negotiator
Member
Ha, good one!
John no longer works at MS, so I thought it's fine to post it now (could have done it long time ago).
I don't think so. Haven't talked to him in a while.
Agreed (if we're talking about a 4x GPU jump).
The 1080p -> 4k transition alone is going to render the jump meaningless.
1.84 x 4 = 7.36 TF (GCN flops) = 5.88 TF (Navi flops)
It definitely doesn't deserve the "monster" moniker IMHO.
Uhm, we're talking about MS, the company that brought 360 BC.
Scarlett will be Zen/x86-based and it will run XB1/Jaguar compiled binaries just fine. Is there any doubt about that?
The "different uarch" excuse would be more believable for Sony, but even in that case, Cerny said that the PS5 will offer PS4 BC (via native execution, not emulation).
Don't you find their statements a bit contradictory? If Zen uarch was an obstacle for BC/native execution, they wouldn't choose Zen 2 for next-gen consoles.
Keep in mind that there was also a rumor (on this very forum) about a more expensive PS4 Pro SKU with a Zen CPU, but it would have cost $499 and it would have been delayed to March 2017. Instead, they chose to hit a $399 target and late 2016 release, so Jaguar it was.
SonGoku
Member
Many assumptions made on his postSonGoku
Next-gen PS5 and next Xbox speculation launch thread |OT5| - It's in RDNA OT
If my word was worth anything, I can confirm that nextgen GPUs ARE Navi 10 aka RDNA 1.0. There is a lot of talk about RT, and how that's an RDN 2.0 feature, but it's very very common to see console APUs be more of a .5 variant of a chip, being based on specific gen but incorporating some...www.resetera.com
Speculation=/confirmation
This point is valid if RDNA2 is indeed 2021 but... (see below)
Can you ask him how he came to that conclusion?
In all likelihood AMD is prioritizing their semicustom business, finishing nextgen consoles RDNA2 chips, so they could already have working engineering samples
To release a Navi20 lineup they have to prepare multiple chips (80CU-72CU-64CU) and their in between disabled versions compared to consoles single version.
Its in AMDs best interest to make sure consoles are powered by their latest arch, that way they benefit from a generation of games tailored to their GPU architecture
Sony/MS wouldn't have chosen 2020 as a release date if they knew 7nm+ and RDNA2 wasn't going to be ready on time. 7nm DUV for all intents and purposes its not a node beneficial for big chips (APUS) they know this and that is why they are not launching earlier than 2020
That's my take on it.
Last edited:
ArabianPrynce
Member
Makes a ton of fucking sense, hence my potential rage and dissapointment if sony/ms announce the 7nm duv chips lmao. thats if they do it.
transformer
Member
Quite a bit of analysis in this thread relating the new consoles to the Technology available now, or very soon. Keep in mind the new consoles come out ~1 year from now. Xbox one x gained the power advantage solely on a one year gap in technology, as state by Phil Spencer. He said they waited a year to hit 4k and would have been where the PS4 pro was if they released a year earlier. So if you think 8-9 TF is realistic based on this year’s technology, there will be advances made for next year.
CyberPanda
Banned
New AMD RT patent if anyone cares:
www.freepatentsonline.com
TEXTURE PROCESSOR BASED RAY TRACING ACCELERATION METHOD AND SYSTEM
Complete Patent Searching Database and Patent Data Analytics Services.
www.freepatentsonline.com
ethomaz
Banned
Well PS4 is not a lot anymore
4x PS4 is a bit better than what Xbox One X is.
Ridiculous for a next-gen.
SonGoku
Member
Its not a linear scale on that bench, RTX2080 (my expectation) is "only" 22k that's just slightly over 4x compared to the PS4 but we all know the RTX2080 is much better than 4x the PS4Well PS4 is not a lot anymore
4x PS4 is a bit better than what Xbox One X is.
Ridiculous for a next-gen.
Last edited:
SpinningBirdKick
Banned
Well PS4 is not a lot anymore
4x PS4 is a bit better than what Xbox One X is.
Ridiculous for a next-gen.
Lockhart - Zombie edition.
Everyone forgot to tell Lockhart that it's dead.
SonGoku
Member
Would be hilarious if Lockart was gonzalo and now that ms ditched lockart AMD offered the chip to the SuborZ team
Last edited:
LordOfChaos
Member
Well PS4 is not a lot anymore
4x PS4 is a bit better than what Xbox One X is.
Ridiculous for a next-gen.
That's only based on this alleged benchmark that magically has Windows running on PS5 hardware. Even if we say that's credible and don't retread on the conversation about how strange that is and how easy it is to fake benchmark driver strings, that's running software that isn't completely tailored for it, possibly in some sort of wrapper that's getting Windows draw calls translated to FreeBSD/GNMX, the raw hardware potential is higher than 4x.
If it does nothing but match the 5700 that's more like north of six times faster than the PS4 GPU.
Fake
Gold Member
More like a plot twist.
bitbydeath
Member
SonGoku
Member
Yeah and the only source linking PS5 with Ariel is/was komachiAriel and Arden are very close together for names in the alphabet.
He's been known for putting speculation on his tweets
vpance
Member
https://www.neogaf.com/threads/reddit-anaconda-dev-kit-dante-leak.1475108/post-253934045
So was there any truth behind the Dante dev kit leak? Could MS be going large and slow, and Sony small and fast?
Scarlett 385mm2 60-70CU 1.475 GHz
So was there any truth behind the Dante dev kit leak? Could MS be going large and slow, and Sony small and fast?
Scarlett 385mm2 60-70CU 1.475 GHz
Ovech-King
Member
@ 672GB/s ? TRIGGERED !!!! Here I go again... this is exactly the bandwith of a 384 bit bus with 14Gbps chips (as seen in Scarlett reveal) but with 24GB GDDR6 though.
Still favors my theory that PS5 will be as much as 30% more powerful just some leaks said because ''apparently'' 18Gbps chips on a 384 bit bus same amount of ram would be 864 GB/s . Screw it , i'm starting a new hashtag; #30%morepowerful
vpance
Member
I still think PS5 will be more powerful too, but maybe only by 20% max. If MS dropped the ball and doesn't have EUV then it's very likely to be 20-30%.
SonGoku
Member
Die size estimates for 7nm EUV
72CU chip 379mm2
80CU chip 394mm2
Let's play it safe and use the 72CU chip (64 enabled)
64CUs @1475Mhz = 12TF
That's RTX2080 levels and then some!
Of course this is just wild speculation on my part
demigod
Member
Ariel and Arden are very close together for names in the alphabet.
You might be onto something. We had Neo and Morpheous. Arden Cho singing Ariel’s song.
ArabianPrynce
Member
There was one aspect i forgot to mention. There is a possability that sony will stick to rdna 1 because of ps4/gcn backwards compability. What do you think.
Imtjnotu
Member
So ms won't and lose backwards compatibility? Nah
ArabianPrynce
Member
Wait i am confused??? I am saying that sony may no use rdna 2 because of BC for gcn.
dead_lighter
Member
Talkin about specs isn’t my kink, both consoles will be powerful enough.
I’m probably one of the few hoping for a backward compatibility with kinect for the xbox.
This gear really differentiates a console, proposes games that can’t be put in a streaming game service, and it’s really fun with young kids (kinect party by double fine, sesame street, just dance, etc.). My XBox360 still is useful today and that’s kind of a shame for the next gen.
I’m probably one of the few hoping for a backward compatibility with kinect for the xbox.
This gear really differentiates a console, proposes games that can’t be put in a streaming game service, and it’s really fun with young kids (kinect party by double fine, sesame street, just dance, etc.). My XBox360 still is useful today and that’s kind of a shame for the next gen.
SonGoku
Member
its bs, rdna1 already has pretty significant changes compared to gcn. If we aplied that logic Sony would use vega for bc
Whatever change rdna2 brings won't be as big as the change in cpu
Last edited:
bitbydeath
Member
More power is always better than less for BC.
PS4 wasn’t powerful enough to be BC with PS3.
ethomaz
Banned
That makes no sense.
Both RDNA and what will come after is ISA compatible with GCN.
BTW GCN will have new version (7.0?) for prosumer cards.
Last edited:
Negotiator
Member
And Jaguar at 2.7 GHz (courtesy of 7nm EUV)...its bs, rdna1 already has pretty significant changes compared to gcn. If we aplied that logic Sony would use vega for bc
- Status