• Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
  • We have made minor adjustments to how the search bar works on ResetEra. You can read about the changes here.

When do you think the PS5 reveal will take place?

  • January

    Votes: 6 0.3%
  • February

    Votes: 1,172 65.7%
  • March

    Votes: 273 15.3%
  • April

    Votes: 81 4.5%
  • May

    Votes: 116 6.5%
  • June

    Votes: 48 2.7%
  • Later

    Votes: 89 5.0%

  • Total voters
    1,785
  • Poll closed .
Status
Not open for further replies.

KeRaSh

I left my heart on Atropos
Member
Oct 26, 2017
10,335
Yeah they have destination PlayStation event for retailers and third party that makes going to E3 basically much less important from that aspect and why Sony wants the ESA to remove that aspect and make it around fans instead and possibly one of the main reasons they won't go anymore.

Someone somewhere from Twitter in the last thread also pointed out that this year is the first in many that Sony themselves won't manage it and instead some other party will and belive it's because Sony is already occupied with other things in February to plan if I remember correctly.
That would make sense. I'm sure Sony would like to fully announce the PS5 before Destination Playstation so they don't have to fear leaks from retailers.
 

Chaostar

Member
Oct 27, 2017
107
New thread, this is needed:

😂

I'm sorry, I had to screen grab this post because of the ad.

Literally pissing ourselves with excitement.

4tUjWHQ.jpg
 

Kenzodielocke

Member
Oct 25, 2017
12,871
Retailers will want to know before May, holy shit people really considering that there wont be an event in a month or two?
 

Skeff

Member
Oct 29, 2017
1,628
According to DukeBlueBall's calculation, a single SE is over 23mm^2, so we are talking about a pretty big die. If we look at the 251mm^2 5700 XT, we've got extra 32-bit memory controller (8mm^2), extra shader engine (23mm^2) and extra 10 WGPs (6 disabled because it has 3 SEs, which is 45mm^2) which results in 327mm^2 just for the GPU before any RT modifications. That's pretty big, even without most of the L3 in the CPU, if we add some more die area for RT, it will be close to 400mm^2.

Yep, Huge, but i think series x and ps5 are both getting huge APU's if were getting 12TF.
 
Oct 27, 2017
7,166
Somewhere South
I said this before and I'll repeat for the new OT: AMD has never done a 3 SE GPU and I feel it's very unlikely it can be done without issues to work scheduling. Navi architecture kinda mitigates most of the need to go with a higher number, too, since it changed cache structure and the backend so they scale with shader arrays, instead.
 

Thera

Banned
Feb 28, 2019
12,876
France
You can click on the tweet and read his bio, it took less time than writing a post here...
I have a gut feeling sony is going to play this real tight to the chest and wont say anything until summer.
Greg Miller thinks May.
Godfall CEO said "Spring" when asked "when will we see them / have news next", that's why I haven't hope for next month. But I am ready to be surprised.
 

Gamer17

Banned
Oct 30, 2017
9,399
You can click on the tweet and read his bio, it took less time than writing a post here...


Godfall CEO said "Spring" when asked "when will we see them / have news next", that's why I haven't hope for next month. But I am ready to be surprised.
Maybe god fall will be at second event about price and games etc... and not PS meeting reveal ?
 

DrKeo

Banned
Mar 3, 2019
2,600
Israel
I said this before and I'll repeat for the new OT: AMD has never done a 3 SE GPU and I feel it's very unlikely it can be done without issues to work scheduling. Navi architecture kinda mitigates most of the need to go with a higher number, too, since it changed cache structure and the backend so they scale with shader arrays, instead.
I don't know if 3 SEs will be a thing or not, but we don't really know what RDNA needs or can do because we have so little GPU samples to draw examples from. If RDNA is very flexible, maybe the SE count is more flexible than before too. We did get the 5500 with a single SE, so we know the SE count doesn't have to be even. But I do agree that it's most likely 2 SEs, especially now that we know that a single SE can handle 12W WGPs, so 2 SE can probably handle at least 24 WGPs which is pretty close to the XSX's 30 WGPs.
 

modiz

Member
Oct 8, 2018
17,907
You can click on the tweet and read his bio, it took less time than writing a post here...


Godfall CEO said "Spring" when asked "when will we see them / have news next", that's why I haven't hope for next month. But I am ready to be surprised.
I get this... but then i am thinking:
"With all the silence on leaks that sony are doing around their PS5 plans, they just let a third party dev give a ballpark of when to expect the event when even they arent willing to?"
 

Expy

Member
Oct 26, 2017
9,876
You can click on the tweet and read his bio, it took less time than writing a post here...


Godfall CEO said "Spring" when asked "when will we see them / have news next", that's why I haven't hope for next month. But I am ready to be surprised.
Why would they show GodFall at the reveal?
 

Skeff

Member
Oct 29, 2017
1,628
52CU with 3 SEs will actually be bigger than a 56CU XSX iGPU because of the 3 SEs and the extra WGP that will need to be disabled.

I expect both to be 60CU chips, so it would only be the extra 23mm2 between the chips? But yes ps5 would be slightly bigger in this scenario.
 

Cyborg

Banned
Oct 30, 2017
1,955
Here is your daily pastebin:(salt and all)
  1. Verona SDK (Prototype 2): All Sony internal studios have these kits.
  2. Prospero SDK (Prototype 1): All major 3rd party studios have these kits.
  3. PS5 SDK (Final): June-July timeframe for all studios.

  4. Verona and Prospero SDKs are similar in design, but spec'd differently for auditing and security reasons. The hardware in Verona are much closer to the finalized SDK and consumer PS5 system.

  5. Sony Interactive Entertainment and Sony Pictures Entertainment division are working closely together after last year's acquisition of Nurulize. Deeper investments into merging both films and video games across both divisions. Bringing a higher production value and film related tech to next generation PlayStation games.
Did we ever hear the name Verona? Is that Shakespeare related also ?

Didn't ZhugeX say that final Devkits would be around E3? Seems in line with 3rd point
 

VN1X

Banned
Oct 28, 2017
2,027
Oh god imagine if the PS5 reveal is in bloody May.

That's like a life time away.
 
Last edited:

DrKeo

Banned
Mar 3, 2019
2,600
Israel
I expect both to be 60CU chips, so it would only be the extra 23mm2 between the chips? But yes ps5 would be slightly bigger in this scenario.
If they have 3 SEs while MS has 2 SEs and both are 60 CUs, then the XSX will be 56 CUs while PS5 is 54 CUs and has a bigger iGPU. On the one hand, MS has 4% more CUs, on the other Sony has 50% more SEs, and we don't know which one has higher clock speed. In that scenario, the PS5's pixel fillrate will destroy the XSX's.

But the more reasonable scenario is that both have 2 SEs, we already know from the Github leak that XSX has 60 CUs (56 active) and we have no idea what Sony has if the Github leak is outdated.
 

Deleted member 8136

User requested account closure
Banned
Oct 26, 2017
1,004
How about some SSD speculation? We had a rumor about the XSX using a Phison controller. If we assume it's a PCIe 4.0 controller, and I think it's safe to assume that, then we only have three options. The mid-range option is the E16, and we already have some SSDs on the market which uses it. One example is the Corsair MP600.

Here is the Corsair MP600 VS the launch Xbox One HDD - the Samsung Spinpoint Momentus ST500LM012:
OiCm4tT.jpg

samsung-momentus-st500lm012-500go-2-5-pouces-5400tpm-05.jpg


As you can see, x44 times the sequential read speed (which matches Phil's 40 times faster statement) which is a good example for resuming a hybernating game out of a non-compressed memory dump (around 4 seconds to resume if we assume 13GB-17GB memory available for games) or initial loading of a game if the developer duplicates data and/or takes the time to arrange the initial loading data on the drive in a sequential order.

As for streaming randomly accessed data, like in Star Citizen (which Dictator has pointed out), we get over 100 times faster data read which is an indicator of performance during the streaming of a large open world.

This looks very likely, good find!
 

MrKlaw

Member
Oct 25, 2017
33,254
Timing:

Before end March - Sony end of financial year report. Likely to include a look ahead to next year (April 1 onwards) and some discussion of PS5. Maybe less relevant to get ahead of this as the name at least has been mentioned along with release window of 2020.

Before Mid-March - Pre GDC to allow for streams if needed or even just to get ahead of technical discussion on engines etc that will be relevant for next gen.

Before end of Feb - Pre Destination Playstation which is normally end of Feb where Retailers will want to know more about content for rest of year

(Godfall Spring comment doesn't impact that - they've recently shown so could easily miss a February reveal and come back for a later presentation in June)


Scope: This would be for providing enough initial information to get traction/news and avoid leaks/spoilers from those two events. looking at what PS meeting did I think thats a good idea of the scope you'd be looking at.



For final price/game announcements you wait until MS has shown its hand at E3 so you go later in June
 
Last edited:

Gamer17

Banned
Oct 30, 2017
9,399
If they have 3 SEs while MS has 2 SEs and both are 60 CUs, then the XSX will be 56 CUs while PS5 is 54 CUs and has a bigger iGPU. On the one hand, MS has 4% more CUs, on the other Sony has 50% more SEs, and we don't know which one has higher clock speed. In that scenario, the PS5's pixel fillrate will destroy the XSX's.

But the more reasonable scenario is that both have 2 SEs, we already know from the Github leak that XSX has 60 CUs (56 active) and we have no idea what Sony has if the Github leak is outdated.
When will they say how many SE they have .i have never seen it mentioned in spec sheets . Is it considered in calculation of TF(i doubt )
 

modiz

Member
Oct 8, 2018
17,907
Word of advice
Stop basing PSM date over sony hall vacancy, that was only relevant under the january 22nd rumor that specifically stated sony hall, otherwise, there are many other places where it could be held.
 

Deleted member 23046

Account closed at user request
Banned
Oct 28, 2017
6,876
Word of advice
Stop basing PSM date over sony hall vacancy, that was only relevant under the january 22nd rumor that specifically stated sony hall, otherwise, there are many other places where it could be held.
22nd january converted from the roman GCN calendar or natively coming from the RDNA gregorian calendar ?
 

Deleted member 12635

User requested account closure
Banned
Oct 27, 2017
6,198
Germany
You can't get to 288bit. Each GDDR6 controllers are 64 bits.

The only way 18GB makes sense:

A) 12gbit chips on 384 bit bus.
B) 12gbit chips on 192 bit bus clamshell
C) Mix of different density chips on 320+ bit bus,
D) Mix of different density chips on 256 bit bus clamshell.
Are you sure? I though GDDR6 uses two 16bit channels per controller:
GDDR6-Memory-Interface-Subsystem-Example.png

DukeBlueBall

DrKeo is correct. The GDDR6 PHY on the 5700 XT and 5500 are all 32-bit. If you look on the die shots of those chiös you see 4 pairs of it on the 5700 and 4 single ones on the 5500. If have no idea how you come to the conclusion the bus-width has to be divisible by 64.

First 5700, Second 5500. Green highlighted areas are the PHYs
ZN6jMih.png
Ft2isy7.png
 
Last edited:
Status
Not open for further replies.