How much money are you willing to pay for a next generation console?

  • Up to $199

    Votes: 33 1.5%
  • Up to $299

    Votes: 48 2.2%
  • Up to $399

    Votes: 318 14.4%
  • Up to $499

    Votes: 1,060 48.0%
  • Up to $599

    Votes: 449 20.3%
  • Up to $699

    Votes: 100 4.5%
  • I will pay anything!

    Votes: 202 9.1%

  • Total voters
    2,210
Status
Not open for further replies.

Outrun

User requested ban
Banned
Oct 30, 2017
5,804
Just a quick question:

What is the performance delta between the two consoles that we would deem as negligible?

7%?
 

modiz

Member
Oct 8, 2018
18,655
the ones that obey the laws of physics which seem to change after every gonzolo leak. Next Consoles wont go over 1.5 ghz! Few months later, laws of physics are bent by the will of Cerny and 1.8 ghz becomes the law of the land.

Two months later, laws of physics are set aside as 2.0 ghz becomes the magical standard.



you are not helping lol
I don't think I have been that inconsistent.
I originally thought PS5 would be 11~13TF, then Navi got revealed with the IPC increase so I lowered expectation of the raw tf number so it would perform according to my expectation.
I don't see why equivalency doesn't exist anymore.
 

sncvsrtoip

Banned
Apr 18, 2019
2,773
Before navi premiere I assumed performance level between vega 56-64 (10.5-12.5tf vega). After premiere it seems that we get 9.2tf but with performace 1.1-1.15x above my expectation so I'm happy. Hope we will get 16GB ram for games + some ram for os/cache and not only 16GB total.
 

FSavage

Member
Oct 30, 2017
562
Oberon, because It is information that was leaked from AMD and has PS4 and PS4 Pro GPU clocks listed as previous generations and the 2000MHz for the new generation. Then Oberon has Ariel GPU, which also connects Gonzalo and flute.

I've been following this set of rumors since the first tweet that mentioned Ariel = PS5?. While it seems the most probable rumor, I don't think it's yet time to declare it "absolutely correct".

Too many posters in here convincing themselves of certain specs when nothing is certain yet. Ya'll gotta chill.

Edit- ya'll look silly getting hostile over other poster's predictions and thoughts, gotta realize that you're all cherry picking rumors...
 

Outrun

User requested ban
Banned
Oct 30, 2017
5,804
I've been following this set of rumors since the first tweet that mentioned Ariel = PS5?. While it seems the most probable rumor, I don't think it's yet time to declare it "absolutely correct".

Too many posters in here convincing themselves of certain specs when nothing is certain yet. Ya'll gotta chill.

Aint that the truth. Next week, shit will change once more.
 

modiz

Member
Oct 8, 2018
18,655
I've been following this set of rumors since the first tweet that mentioned Ariel = PS5?. While it seems the most probable rumor, I don't think it's yet time to declare it "absolutely correct".

Too many posters in here convincing themselves of certain specs when nothing is certain yet. Ya'll gotta chill.
Oberon is definitely PS5 because what else would have PS4 and PS4 pro's gpu clocks listed as previous generation?!
I will repeat myself, this APU has a GPU with 3 clock speed configurations, PS4's 800MHz, PS4 pro's 911MHz, and 2000MHz, this is an almost assured proof that this is the next generation PlayStation
 
Oct 27, 2017
7,317
Somewhere South
Early PS5 devkits were likely packing a Vega 64 or a Radeon VII - i.e. the were performance target devkits - and that's most likely where we get rumors like Warren's or Benji's from. So, 12 to 13TF, higher than the 10.7TF from Stadia.

We now know that, performance-wise, the RX 5700XT pretty much aligns with that.
 

Astronut325

Member
Oct 27, 2017
5,948
Los Angeles, CA
  • Tom Warren Rumor:
  • Scarlett die size estimates in this very thread point to 56 CUs. Reiner said PS5 is stronger which means the PS5 also has at least 56 CUs, clocked higher to get more tflops. Which could also explain why Colin said the PS5 is DEFINITIVELY more powerful.
  • 4Chan leak in April mentioned 1849 mhz clockspeed which gets us 13.2 tflops. It also lines up with the gonzolo rumor which showed 18 (or 1.8 ghz). its entirely possible that gonzolo was 1.849 this entire time and 13 tflops instead of 12.9 like we all though.
  • RuthenticCookie who called the Sony E3 cancellation a week before it happened also said the PS5 is a monster. That Sony is going to lose $100. This is before months the RDNA reveal. I doubt anyone was
  • Jason Schrier in this very thread said both MS and Sony were aiming to top Google's 10.7 tflops number. Again, this is BEFORE we knew about RDNA. He specifically said number.

  • Then there are the more less reliable leaks, but have proven to be more correct than 99% of the people in this thread. They confirmed hardware RT and SSD months before the official confirmation when almost everyone here was laughing at that suggestion.
  • The project episilon leak is riddled with errors and most likely fake, but i am surprised by how much this pastebin got right months before we knew anything. https://pastebin.com/PY9vaTsR
    • - No Sony at E3.​
    • - GTA6 set in Miami and NYC. (this was corroborated by another gta6 rumor a few months later)​
    • - 8k upscaling (curious how sony mentioned this)​
    • - retina tracking and wireless PSVR (someone here found a sony patent for this)​




So we have 4 industry vets, 2 saying higher than 10.7 tflops, 2 saying PS5 is more powerful, plus a bunch of leaks than line up with the gonzalo leaks. But almost everyone seems to ignore all of this and had no trouble believing era insiders who created threads right after the PS5 reveal claiming the Xbox Anaconda was more powerful based on no information on PS5 tflops counts. They even ignored Brad Sams who changed his tune from MS aiming to be more powerful to simply matching Sony on performance. This was back in March.

I just find it funny that when Gonzalo was first leaked to have 1.0 ghz clock, everyone assumed a 64 CU part. 8 tflops. Then when 1.8 ghz was revealed, everyone went down to 36 CUs. Once again, 8 tflops. When Navi was revealed and it looked like 40 CU might be on the cards. And now that it has a 2.0 ghz clock, we are down to 36 CUs? We seem to downgrade the PS5 after every piece of good news. Its hilarious.

I can't thank you enough for this. Excellent summary and helps me catch up on this thread.

Edit:
This gives me flashbacks to the speculation threads at the old place prior to the start of the current gen. EVERYONE thought the PS4 would be 10-20% more powerful than the Wii U. I was just not buying it.
 

DukeBlueBall

Banned
Oct 27, 2017
9,059
Seattle, WA
  • Tom Warren Rumor:
  • Scarlett die size estimates in this very thread point to 56 CUs. Reiner said PS5 is stronger which means the PS5 also has at least 56 CUs, clocked higher to get more tflops. Which could also explain why Colin said the PS5 is DEFINITIVELY more powerful.
  • 4Chan leak in April mentioned 1849 mhz clockspeed which gets us 13.2 tflops. It also lines up with the gonzolo rumor which showed 18 (or 1.8 ghz). its entirely possible that gonzolo was 1.849 this entire time and 13 tflops instead of 12.9 like we all though.
  • RuthenticCookie who called the Sony E3 cancellation a week before it happened also said the PS5 is a monster. That Sony is going to lose $100. This is before months the RDNA reveal. I doubt anyone was
  • Jason Schrier in this very thread said both MS and Sony were aiming to top Google's 10.7 tflops number. Again, this is BEFORE we knew about RDNA. He specifically said number.

  • Then there are the more less reliable leaks, but have proven to be more correct than 99% of the people in this thread. They confirmed hardware RT and SSD months before the official confirmation when almost everyone here was laughing at that suggestion.
  • The project episilon leak is riddled with errors and most likely fake, but i am surprised by how much this pastebin got right months before we knew anything. https://pastebin.com/PY9vaTsR
    • - No Sony at E3.​
    • - GTA6 set in Miami and NYC. (this was corroborated by another gta6 rumor a few months later)​
    • - 8k upscaling (curious how sony mentioned this)​
    • - retina tracking and wireless PSVR (someone here found a sony patent for this)​




So we have 4 industry vets, 2 saying higher than 10.7 tflops, 2 saying PS5 is more powerful, plus a bunch of leaks than line up with the gonzalo leaks. But almost everyone seems to ignore all of this and had no trouble believing era insiders who created threads right after the PS5 reveal claiming the Xbox Anaconda was more powerful based on no information on PS5 tflops counts. They even ignored Brad Sams who changed his tune from MS aiming to be more powerful to simply matching Sony on performance. This was back in March.

I just find it funny that when Gonzalo was first leaked to have 1.0 ghz clock, everyone assumed a 64 CU part. 8 tflops. Then when 1.8 ghz was revealed, everyone went down to 36 CUs. Once again, 8 tflops. When Navi was revealed and it looked like 40 CU might be on the cards. And now that it has a 2.0 ghz clock, we are down to 36 CUs? We seem to downgrade the PS5 after every piece of good news. Its hilarious.


tumblr_o16n2kBlpX1ta3qyvo1_1280.jpg
 

OnPorpoise

Avenger
Oct 25, 2017
1,339
I always think back to the "2.5 teraflop" Sweeney quote and the ensuing fiasco it caused. He wasn't really wrong, but it didn't exactly hold true in the console space.

It's better to allow for some wiggle-room with tweets/quotes/etc from reliable sources, no reason to call them out as liars, and no reason to treat them like they were carved in stone from on high either.
 
Oct 27, 2017
4,018
Florida
If Xbox gives me the estimated 380mm2 7nm(+?) die without going too low on frequency I'm going to be a happy man overall. Can't ask for too much more than that in a console in 2020. Let's see how they use the silicon.
 
Dec 31, 2017
1,430
I actually believe Sony will have a more powerful console, but I also expect a repeat of the 360/PS3 Gen where their console will be more powerful but also more expensive, andthis time around I have a feeling MS is aiming for a lower price tag like the 360 gen. No fancy new controller that adds useless features, no Kinect, just pure gaming console at a reasonable launch price.
 
Oct 26, 2017
6,151
United Kingdom
I really doubt anyone is going to have a bigger die than the XB1X this gen. 7nm is just way too expensive for that.

You go too big and you get stuck with a huge die for the rest of the gen, because future die shrinks just don't justify the huge hike in wafer price to net you any real meaningful cost reduction.
 

Chamber

Member
Oct 25, 2017
5,311
I just find it funny that when Gonzalo was first leaked to have 1.0 ghz clock, everyone assumed a 64 CU part. 8 tflops. Then when 1.8 ghz was revealed, everyone went down to 36 CUs. Once again, 8 tflops. When Navi was revealed and it looked like 40 CU might be on the cards. And now that it has a 2.0 ghz clock, we are down to 36 CUs? We seem to downgrade the PS5 after every piece of good news. Its hilarious.
Accurate, Gonzalo could be 3Ghz tomorrow and it'd still somehow wrap back around to 8-9TF with the armchair engineers talking about thermal budgets and backwards compatibility.

Gonzalo/Oberon/Ariel is almost certainly PS5 but all we have on that is a clock speed that keeps changing and a vague benchmark result.
 
Jan 17, 2019
964
where does it say that Navi 10 can only have 56 CUs?

Does Navi 10 have RT? PS5 has RT.

Why can it have RT and not have more CUs? a 2.0 Ghz clock almost assures that its on 7nm+. 7nm+ gives 15-20% die size reduction. that should give allow them to add 56 CUs without going over 320mm2. smaller than the pro die size.



Yes. 14.2 tflops is exactly what the pastebin leak back in october mentioned. 14.3 tflops was on that hilarious project epsilon leak back in April 2018. Ruthentic cookie also hinted at 14 tflops but i cant find that post.

Maybeeeee this will help :

 

Kleegamefan

User requested ban
Banned
Dec 16, 2017
980
I'm not gonna say anything specific, but back in June, I asked a buddy of mine who is currently working on a next Gen game what kind of final performance to expect...

PlayStation 5 is not 9.2TF.

Both it and Scarlett are greater than that...

You can choose to believe me or not, I don't care.....I'm not in the gaming industry any more so....
 
Feb 23, 2019
1,426
I'm not gonna say anything specific, but back in June, I asked a buddy of mine who is currently working on a next Gen what kind of performance to expect...

PlayStation 5 is not 9.2TF.

Both it and Scarlett are greater than that...

You can choose to believe me or not, I don't care.....I'm not in the gaming industry any more so....

GCN or RDNA TFlops? Big distinction
 

Deleted member 40133

User requested account closure
Banned
Feb 19, 2018
6,095
I'm not gonna say anything specific, but back in June, I asked a buddy of mine who is currently working on a next Gen what kind of performance to expect...

PlayStation 5 is not 9.2TF.

Both it and Scarlett are greater than that...

You can choose to believe me or not, I don't care.....I'm not in the gaming industry any more so....

Any indication of advantages one may have over the other? Bandwidth, ram etc
 

Pheonix

Banned
Dec 14, 2018
5,990
St Kitts
Ok wait... people here are expecting a 56CU (52CU active) GPU????

It hasn't occurred to some said people that the 40CU 5700XT pulls in an average of 235W while powering less RAM and components than a console will need to power? Oh and also lacking a CPU.

I mean I like throwing numbers around like the best of em... but how does anyone expect any of this to be possible?

If we take a 52CU Navi GPU and couple that with 8GB it would need at least 270W to drive that GPU alone past 1800Mhz. Now let's throw on a CPU and more ram and an SSD, and that's putting power consumption at over 300W. Now lets even assume they are using 7nm+. Thats only gonna give 15% power efficiency. That's still putting the system power draw at over 270W.

How are they gonna shave off 80-120W of power draw to fit inside console ranges?
 

orochi91

Member
Oct 26, 2017
10,275
Canada
I'm not gonna say anything specific, but back in June, I asked a buddy of mine who is currently working on a next Gen game what kind of final performance to expect...

PlayStation 5 is not 9.2TF.

Both it and Scarlett are greater than that...

You can choose to believe me or not, I don't care.....I'm not in the gaming industry any more so....

Ayyyyy

This literally changes everything.

Bravo, lol
 

StrykerIsland

Member
Oct 25, 2017
2,308
Ok wait... people here are expecting a 56CU (52CU active) GPU????

It hasn't occurred to some said people that the 40CU 5700XT pulls in an average of 235W while powering less RAM and components than a console will need to power? Oh and also lacking a CPU.

I mean I like throwing numbers around like the best of em... but how does anyone expect any of this to be possible?

If we take a 52CU Navi GPU and couple that with 8GB it would need at least 270W to drive that GPU alone past 1800Mhz. Now let's throw on a CPU and more ram and an SSD, and that's putting power consumption at over 300W. Now lets even assume they are using 7nm+. Thats only gonna give 15% power efficiency. That's still putting the system power draw at over 270W.

How are they gonna shave off 80-120W of power draw to fit inside console ranges?
I'm confused as to why power draw is such a concern in this thread. What's to say these boxes won't be bigger than ever before, with a new type of cooling system that they've never used before? I say, go big or go home. I want a power-hungry beast that sucks up the juice like no console before it. The thirst is real.
 

RoboPlato

Member
Oct 25, 2017
7,161
Accurate, Gonzalo could be 3Ghz tomorrow and it'd still somehow wrap back around to 8-9TF with the armchair engineers talking about thermal budgets and backwards compatibility.

Gonzalo/Oberon/Ariel is almost certainly PS5 but all we have on that is a clock speed that keeps changing and a vague benchmark result.
Honestly the great thing about that is that 9.2tflops is most likely the MINIMUM we can expect for PS5 if it actually is Oberon. Given that even that would be higher in game performance than my previous predictions (Vega 56-64) and we're getting raytracing and an SSD (neither of which I thought were possible), I'm going to be thoroughly pleased with whatever the final specs are and I think most people in here will too.
 

AegonSnake

Banned
Oct 25, 2017
9,566
Thank you!

Of course its RDNA. Why would devs talk about Vega tflops is beyond me. Tom Warren is posting devkit rumors a day before E3. Two months after the gonzolo qualifying sample leaked and Cerny said Devkits are going out. Why would Cerny release devkits with Vega GPUs especially when gonzalo was so close to being ready?
 

Pheonix

Banned
Dec 14, 2018
5,990
St Kitts
Just a quick question:

What is the performance delta between the two consoles that we would deem as negligible?

7%?
Anything less than 20% will be negligible. That's like saying a 3-6fps drop which can be massaged out with specific coding.

As it stands, things like the SSD type and implementation is more likely to have a more noticeable impact between the two. And that's already really pushing it cause even that is likely going to be negligent.
 
Oct 25, 2017
18,223

Bitch Pudding

Member
Oct 25, 2017
4,202
I'm not gonna say anything specific, but back in June, I asked a buddy of mine who is currently working on a next Gen game what kind of final performance to expect...

PlayStation 5 is not 9.2TF.

Both it and Scarlett are greater than that...

You can choose to believe me or not, I don't care.....I'm not in the gaming industry any more so....

Ah, totally fine for me. But referring to a friend instead of your uncle sure was a missed opportunity ;).
 

Chamber

Member
Oct 25, 2017
5,311
Honestly the great thing about that is that 9.2tflops is most likely the MINIMUM we can expect for PS5 if it actually is Oberon. Given that even that would be higher in game performance than my previous predictions (Vega 56-64) and we're getting raytracing and an SSD (neither of which I thought were possible), I'm going to be thoroughly pleased with whatever the final specs are and I think most people in here will too.
Oh, everyone is going to be pleased no doubt. Vague speculation about teraflops and RAM is fun but nobody is going to care once we see the games which are guaranteed to look amazing.
 

Andromeda

Member
Oct 27, 2017
4,997
Early PS5 devkits were likely packing a Vega 64 or a Radeon VII - i.e. the were performance target devkits - and that's most likely where we get rumors like Warren's or Benji's from. So, 12 to 13TF, higher than the 10.7TF from Stadia.

We now know that, performance-wise, the RX 5700XT pretty much aligns with that.
I agree.
I'm not gonna say anything specific, but back in June, I asked a buddy of mine who is currently working on a next Gen game what kind of final performance to expect...

PlayStation 5 is not 9.2TF.

Both it and Scarlett are greater than that...

You can choose to believe me or not, I don't care.....I'm not in the gaming industry any more so....
What ? But how does it work for PS5 as everything point to PS5 having Navi 10 with 40CUs ? this is really puzzling. They really can't overclock it above 2ghz, can they ?

Thanks for this BTW ! You spoke AND did know something !
 

Kleegamefan

User requested ban
Banned
Dec 16, 2017
980
Oh, everyone is going to be pleased no doubt. Vague speculation about teraflops and RAM is fun but nobody is going to care once we see the games which are guaranteed to look amazing.


Yeah. The performance figures(once Sony and Microsoft divulge them) won't blow you away but the games will, IMO.

Your body isn't ready.....mine wasn't!
 

AegonSnake

Banned
Oct 25, 2017
9,566
Ok wait... people here are expecting a 56CU (52CU active) GPU????

It hasn't occurred to some said people that the 40CU 5700XT pulls in an average of 235W while powering less RAM and components than a console will need to power? Oh and also lacking a CPU.

I mean I like throwing numbers around like the best of em... but how does anyone expect any of this to be possible?

If we take a 52CU Navi GPU and couple that with 8GB it would need at least 270W to drive that GPU alone past 1800Mhz. Now let's throw on a CPU and more ram and an SSD, and that's putting power consumption at over 300W. Now lets even assume they are using 7nm+. Thats only gonna give 15% power efficiency. That's still putting the system power draw at over 270W.

How are they gonna shave off 80-120W of power draw to fit inside console ranges?
  • 5700XT at 2.0 ghz consumes <160w.
  • 5700XT at 1.6ghz consumes <100w.
  • Add 40% more CUs at 1.6Ghz and you are looking at <140w for a 11.4 tflop 56 CU GPU on 7nm.
  • 7nm+. 15-20% smaller die and/or more power efficient. Thats 112-126w for a 56CU GPU at 1.6 Ghz on 7nm+.
  • As for Sony, youtubers were able to get CPU wattage down to 10-20w for the new Ryzen 3700. PS5 and most likely Scarlet will have a cutdown version of that.
  • HBM2+DDR4. 200-225w console.
 

Andromeda

Member
Oct 27, 2017
4,997

Deleted member 40133

User requested account closure
Banned
Feb 19, 2018
6,095
Could you at least tell us that the target specs are very close to each other's ? I think that would put everyone at ease as there won't be a weak baseline (currently XB1) for nextgen games !


Navi 10 has 40 CUs.
Why can't they go over 40 though? It is semi custom

Edit: more cu is more plausible than clocked over 2, is all I'm saying
 

Bashteee

Member
Oct 27, 2017
1,194
I agree.


What ? But how does it work for PS5 as everything point to PS5 being Navi 10: 40CUs ? this is really puzzling. They really can't overclock it above 2ghz, can they ?

Thanks for this BTW ! You spoke AND did know something !


Just a thought, but what if Gonzalo is the rumored 2019 chip they had so far in dev, that they are now using it in dev kits until they can ship the 2020 chip in dev kits? Does that make any sense?

Why can't they go above 40 CU?

They can, but Navi 10 LITE seems to be based on the Navi we see today in the shape of 5700 and 5700XT with 36 and 40 CUs.
 

Kleegamefan

User requested ban
Banned
Dec 16, 2017
980
Status
Not open for further replies.