Baseless Next Generation Rumors with no Technical Merits [post E3 2019, pre GDC 2020] [XBSX, PS5]

Status
Not open for further replies.
I never said they didn't but as far as I know nothing of the short was really announced for the new controller? The only noteworthy update was a share button like the DS4 has, but N2O has been insisting that what was announced for the DS5, the XSX already has, which is not true.
I think their messaging regarding the controller is the same as the rest of the Series X: Show but don't tell all that much.

Would not surprise me if they snuck in some new features.
 
Has anyone done a comparison between the blocks tested/missing between Arden Vs Oberon?

"quite a few blocks... missing" definitely suggests that it wasn't a complete native test of the chip(s).
I'll check, but all chips (Oberon/Arden/Renoir/MI100) had native sheets which consist of specific blocks, but not all. Since native Arden test had shader block we could get CU count. Since Oberon/Ariel had texture fillrate and CB blocks we could get native CU count. (But this one also contained referrences to full chip being 18WGPs and "CU_count = 36 for native")
 
XB1 already has, not XSX.
You can keep thinking how innovative it will be though, not the first time some people falling about this kind of feature end up no one use it.

It's not the same freaking thing and XB1 does not have it. It's more than rumbling. For something like Horizon Zero Dawn for example, having resistance in the trigger changes gameplay itself as you have to be much more precise and it makes it harder.
 
Last edited by a moderator:
The force-feedback triggers on DS5 sound rad. I remember how much force-feedback wheels added to the immersion of driving games.
 
i remember my jog con with ridge racer type 4

SSONPSACCNAMCON01_l.jpg
 
I think their messaging regarding the controller is the same as the rest of the Series X: Show but don't tell all that much.

Would not surprise me if they snuck in some new features.

I would not be surprised either, but for now we don't know.

It's just while MS message about XSX has been about power, Sony has been about the experience, of which the focus on the controller was an example.
 
Heres the part about the new controller from the gamespot article when XSX was relevealed:

https://www.gamespot.com/articles/xbox-series-x-exclusive-details-meet-microsofts-ne/1100-6472190/

The New Xbox Series X Controller
Not everything has been given a dramatic makeover, however. The new Xbox controller shipping with Series X consoles looks a lot like the old one. The biggest difference, from a distance, is the addition of a share button in the middle, implemented to help you quickly share your gaming moments with friends. I was also told that the transition from the face of the controller to the top and over to the back is more rounded than before, which is likely more of an aesthetic feature than an ergonomic one--we'll have to wait and see.

Getting down to the details, Spencer had the following to say about the controller's new features: "There were certain things that we've learned through doing the Elite controller and just listening to fans. One of them is on the d-pad we have a new hybrid d-pad that we've been working on that we think is important, so you'll get a sense of that in the new Xbox wireless controller. We do have a share button. We've heard the feedback. We're not the first ones to do a share button, so we're not going to say that we invented that, but we've heard feedback that sharing is such a part of a gaming experience now for many of our players that I wanted a dedicated hardware button to share, so you'll see that. We'll still have all the rumble triggers and haptic feedback that you've had in the console before."

Spencer also noted that while the current Xbox One controller fits the majority of hands, according to Microsoft's research, the new design is ever-so-slightly smaller in subtle ways to increase that coverage from 95% to 98%. And in terms of cross-compatibility, not only will Xbox One controllers work on Series X, but its new controller will also function on Xbox One consoles and on PCs.

Ergonomics and inputs aside, under the hood, Microsoft is also working to improve the already excellently low input latency of Xbox controllers, which rely on a proprietary radio, rather than Bluetooth--though some Xbox One controllers have Bluetooth support for added compatibility across non-Xbox devices. Ronald explained that one tool developers have, "dynamic latency input," allows inputs to directly sync up with a game's rendering path. And in Spencer's opinion, reducing the disconnect between button presses and actions on screen is one step down the path to greater immersion. "So when we talk about things like refresh rate," he added, "and we talk about input latency, this is all about the most immersive experience game designers can create, where the visuals are stunning, my ability to get into the experience [is] very timely, it's as great as it's ever been with the I/O speeds and the load times we're going to see, and the input and the ability for just my control and activation of my character or of the game itself becomes a subconscious thing and not something that I think about."
 
In fairness, no one has used it yet to determine what difference there is.
https://www.wired.com/story/exclusive-playstation-5/

The controller (which history suggests will one day be called the DualShock 5, though Cerny just says "it doesn't have a name yet") does have some features Cerny's more interested in acknowledging. One is "adaptive triggers" that can offer varying levels of resistance to make shooting a bow and arrow feel like the real thing—the tension increasing as you pull the arrow back—or make a machine gun feel far different from a shotgun. It also boasts haptic feedback far more capable than the rumble motor console gamers are used to, with highly programmable voice-coil actuators located in the left and right grips of the controller.

Haptics aren't new, that is.

They mention more, in both articles, I think.
 
The only visible changes are the share button, the d-pad and the design near the top of the controller/bumpers are:


3615036-xbox-new-controller.jpg
 
I listen the podcast it is before the TGA the 5th december. From 25 minutes Jason Schreirtalk about the fact Lockhart is not cancelled finally and they heard complain about the fact to release on two SKUs. And at 38 minutes he told PS5 and Xbox are very powerful and it compares it positively form what he heard from dev on the GPU side talking about Anaconda and PS5 as more powerful than RTX 2080. Interesting podcast.

https://podcasts.google.com/?feed=a...CAUQieUEegQIARAE&ep=6&at=1575634615373http://
So I listened to this part and this is what Jason said...

I think...I mean, I'm pretty sure, don't hold me to this up there everybody...rumormongers out there... but I'm pretty sure that the specs these high end PS5 and Anaconda are going for is more powerful then capabilities of graphics cards like RTX 2080, the stuff you can get now on the market today and they are trying to future proof it. I have heard alot of poisitive buzz about power level these guys are working with

I consider Jason to be probably best source in the industry, and if you look at this quite, I mean, its not like he has specs in his hand he is working with. He thinks they are future proof, he thinks that they are better then RTX 2080, but there is really no exact substance. He even says "Don't hold me to this one up there everybody..."

I am saying this because, difference between updated Navi for next year at around 10TF, with RT/VRS and more BW then 5700 or RX 2080, could very well end up being in that ballpark.

This is why I am "agnostic" to this issue. I think document leak we got is legit, its a snapshot of regression test during late Spring/early summer and I cannot but take it at face value. For some people 9.2TF Navi is slow. For me, even Navi 10 (with less BW then what Oberon shows) is a card only few % behind Nvidia 2070 Super.

Again, anyone can go back to Beyond3d (or any other forum) and find posts when first Oberon leaked. You will see that alot of people, who are now saying 36CU at 2.0GHz make no sense and 9-10TF is low end, sure saw alot of sense back then, because XT performance results were very impressive. That card, even with less BW then Oberon and lacking RT/VRS, decimated Vega 64 and pretty much matched R7. Only, only reason why people do not believe its real is because XSX seems to be pointing at 12TF. Thats only reason. Remove XSX, tell them improved 5700XT with more BW, RT and VRS is card that will find its way into PS5 and people would absolutely rejoice. Since XSX seems to be 12TF, we are now trying to make sense of all the vague leaks that say otherwise.

I am still waiting on complete picture, I still think PS5 is 36CU chip (so, GPU portion should be bigger then one that got into PS4/Pro), but I am willing to stay open minded and wait for confirmation or better leak. Unfortunately, current situation is such that no one actually provided ANY number bar Github. In fact, Github shown Sparkman APU during the summer when every single insider said it was cancelled back in early 2019 (so, someone like Matt and Klee), which was as we can see by data, almost certainly wrong.

Oh, and I want to add. From benchmarks I saw, overclocking Navi core is not going to yield big % gains, but memory bandwidth is. It is well worth to mention that TF per TF, Navi and Touring are ~1%. Alot of people forget it, but it is true, Navi is as good as Nvidia per TF. This is why 5700 and XT are actually closer then their TF count indicates, because per TF 5700 has more BW available.

My opinion on what is Ariel and what is Oberon is...Ariel was GFX1000. Its referred as GFX1000. Its a first Navi chip. It was referred as "Navi 10 Lite" because it was iGPU derived from Navi 10. Oberon is actual APU going to PS5, with additional features refreshed/RDNA2 chips will be getting in 2020. Ariel was used as chip to make sure BC is working, but didn't have RT/VRS and anything else PS5 will have perhaps.

This GPU, in PC space, is 251mm2. Its 10% bigger then equivalent that went into PS4Pro and 20% bigger then the one that went into PS4. This is without counting RT/VRS or any additional hardware logic going in it above what 40CU Navi has in PC space. This is also one of the things that leads me to believe Sony is going for such chip, because even "puny" 40CU Navi with RT/VRS is well above what they put in PS4 and Pro.
 
Last edited:
So I listened to this part and this is what Jason said...



I consider Jason to be probably best source in the industry, and if you look at this quite, I mean, its not like he has specs in his hand he is working with. He thinks they are future proof, he thinks that they are better then RTX 2080, but there is really no exact substance. He even says "Don't hold me to this one up there everybody..."

I am saying this because, difference between updated Navi for next year at around 10TF, with RT/VRS and more BW then 5700 or RX 2080, could very well end up being in that ballpark.

This is why I am "agnostic" to this issue. I think document leak we got is legit, its a snapshot of regression test during late Spring/early summer and I cannot but take it at face value. For some people 9.2TF Navi is slow. For me, even Navi 10 (with less BW then what Oberon shows) is a card only few % behind Nvidia 2070 Super.

Again, anyone can go back to Beyond3d (or any other forum) and find posts when first Oberon leaked. You will see that alot of people, who are now saying 36CU at 2.0GHz make no sense and 9-10TF is low end, sure saw alot of sense back then, because XT performance results were very impressive. That card, even with less BW then Oberon and lacking RT/VRS, decimated Vega 64 and pretty much matched R7. Only, only reason why people do not believe its real is because XSX seems to be pointing at 12TF. Thats only reason. Remove XSX, tell them improved 5700XT with more BW, RT and VRS is card that will find its way into PS5 and people would absolutely rejoice. Since XSX seems to be 12TF, we are now trying to make sense of all the vague leaks that say otherwise.

I am still waiting on complete picture, I still think PS5 is 36CU chip (so, GPU portion should be bigger then one that got into PS4/Pro), but I am willing to stay open minded and wait for confirmation or better leak. Unfortunately, current situation is such that no one actually provided ANY number bar Github. In fact, Github shown Sparkman APU during the summer when every single insider said it was cancelled back in early 2019 (so, someone like Matt and Klee), which was as we can see by data, almost certainly wrong.

I agree with pretty much all of this.

The only red flag around Ariel/Oberon for me though is those clock speeds. If you look at the clock speeds from the HD 7000 series in 2012 up to the RX 5000 series now there has been a steady increase in the clocks of AMD's GPU's but 1.8-2.0GHz still seems really high for a console....especially 2Ghz.
 
Last edited:
So I listened to this part and this is what Jason said...



I consider Jason to be probably best source in the industry, and if you look at this quite, I mean, its not like he has specs in his hand he is working with. He thinks they are future proof, he thinks that they are better then RTX 2080, but there is really no exact substance. He even says "Don't hold me to this one up there everybody..."

I am saying this because, difference between updated Navi for next year at around 10TF, with RT/VRS and more BW then 5700 or RX 2080, could very well end up being in that ballpark.

This is why I am "agnostic" to this issue. I think document leak we got is legit, its a snapshot of regression test during late Spring/early summer and I cannot but take it at face value. For some people 9.2TF Navi is slow. For me, even Navi 10 (with less BW then what Oberon shows) is a card only few % behind Nvidia 2070 Super.

Again, anyone can go back to Beyond3d (or any other forum) and find posts when first Oberon leaked. You will see that alot of people, who are now saying 36CU at 2.0GHz make no sense and 9-10TF is low end, sure saw alot of sense back then, because XT performance results were very impressive. That card, even with less BW then Oberon and lacking RT/VRS, decimated Vega 64 and pretty much matched R7. Only, only reason why people do not believe its real is because XSX seems to be pointing at 12TF. Thats only reason. Remove XSX, tell them improved 5700XT with more BW, RT and VRS is card that will find its way into PS5 and people would absolutely rejoice. Since XSX seems to be 12TF, we are now trying to make sense of all the vague leaks that say otherwise.

I am still waiting on complete picture, I still think PS5 is 36CU chip (so, GPU portion should be bigger then one that got into PS4/Pro), but I am willing to stay open minded and wait for confirmation or better leak. Unfortunately, current situation is such that no one actually provided ANY number bar Github. In fact, Github shown Sparkman APU during the summer when every single insider said it was cancelled back in early 2019 (so, someone like Matt and Klee), which was as we can see by data, almost certainly wrong.

Oh, and I want to add. From benchmarks I saw, overclocking Navi core is not going to yield big % gains, but memory bandwidth is. It is well worth to mention that TF per TF, Navi and Touring are ~1%. Alot of people forget it, but it is true, Navi is as good as Nvidia per TF. This is why 5700 and XT are actually closer then their TF count indicates, because per TF 5700 has more BW available.

My opinion on what is Ariel and what is Oberon is...Ariel was GFX1000. Its referred as GFX1000. Its a first Navi chip. It was referred as "Navi 10 Lite" because it was iGPU derived from Navi 10. Oberon is actual APU going to PS5, with additional features refreshed/RDNA2 chips will be getting in 2020. Ariel was used as chip to make sure BC is working, but didn't have RT/VRS and anything else PS5 will have perhaps.

This GPU, in PC space, is 251mm2. Its 10% bigger then equivalent that went into PS4Pro and 20% bigger then the one that went into PS4. This is without counting RT/VRS or any additional hardware logic going in it above what 40CU Navi has in PC space. This is also one of the things that leads me to believe Sony is going for such chip, because even "puny" 40CU Navi with RT/VRS is well above what they put in PS4 and Pro.

You think someone not tech savvy will pull this out of thin air. He just repat what he was told same when he speaks about how to tackle Lockhart.

He just did not give more precise information because last time he said Xbox Anaconda and PS5 have more TFlops than Stadia someone made an article.

He was pissed off and he said this is why I am not posting more on era.

The 5700XT is clearly less powerful than a RTX 2080.

He said from what I heard.

Do I say Github is not legit? Matt the era mod told( verified by other mod to work in the videogame industry) told this is outdated data. Same said by two other people working in the industry with verified tag on era.
 
What is this Project Infinity he mentions? Never heard about that before.
Can't say I know..

They do, but they wouldnt break their NDA and leak it to those posting on forums.
Because that hasn't happened in the past, with pretty much all consoles so far?
It's one thing to try to discredit this or that rumor.
Assuming every single rumor out there must be fake because all developers this particular generation (unlike all the others) are super afraid of being caught for using anonymous forum accounts to say "they're close in power" or "around 12TF" and "less than 1TF apart" seems a bit unrealistic to me.



It wasn't nit-picky, but talking generalisations. He posted a few examples of things people would look for. Heck, he even mentioned resolution would be different in terms of the pixel-counters, and said the pixel-counters would find it very difficulty - not that they were the same resolution.
For the pixel counters to have difficulty noticing the differences (the same pixel counters who notice differences in the XBoneS vs. PS4 with 28.6% difference), then we're not talking about a 30% difference in shading throughput.



With his descriptions, going by the rumours many are taking as set in stone, PS5 has less TFs, loads the game in 7 seconds instead of 10, and renders the game at the same 60 fps in a way that is nigh indistinguishable from the XBSX because it's 'better balanced', such that it's probably 1880p with superb upscaling vs 4k native.

1 - "PS5 has less TFs" -> he actually says the opposite:

(note: the post is from yesterday, so the year is 2020)

2 - "better balanced" -> I see no mention of balance in his posts. Is this purely your speculation or is there another post from him that I'm not taking into account?

3 -"probably 1800p" -> I see no mention of upscaling in his posts. Like the post above, is this your speculation?



That's not in any way explicit though. The point of the post was that end users aren't going to see a big difference, and all the fussing about specs was silly. Even if your console of choice has less TFs or a slower SSD, it's not going to make a meaningful difference to your experience of the games machine. That's the only message intended here. People trying to read into it that the machines are the same performance or not are reading between the lines things never intended.
He says end users aren't going to see any difference, which is different from users not seeing a big difference. Which means we're not talking about 9.2 vs. 12 TFLOPs.


There is no RT in Oberon regression test because tests are split by blocks. RT is (at least for Arden) in GLX block, and Oberon had no GLX block tested (it only tested Ariel GPU specs), hence no RT related entries. In fact, both Arden and Oberon had quite a few blocks in regression tests missing that other had, and its most certainly not missing in hardware (as certain blocks are default for Navi).

Disproving the leak like this just shows some people never bothered to get the repo and look at it themselves...
Yes, you've made us aware many times that you're convinced that the Gospel tells the whole story. To the point of believing the SoC isn't even using any CUs for redundancy and there's no more than 36 physical CUs (which is unprecedented for all GCN consoles so far).

Right, so when some 'jounalist' claims they are 'close', they perhaps aren't even reffering to TF numbers.
Yes, specifically when this same 'journalist' actually referred to TF numbers himself.
Jason Schreier said:


He's not talking about "general power", nor "effective performance" nor any other subjective term. He's saying they're both aiming higher than 10.7 TFLOPs and there's really no way around that, regardless of the architecture.
 
I have the Github repo and I kept a copy on one google drive account and it was deleted by an AMD demand. I never said it is not legit but from one trustable people said it is outdated. Same things when @Dictator told on era Richard Leadbetter heard by someone trustable in the industry a PS5 version have Jaguar and he spoke about it inside one Digitalfoundry video, this looks like something 1i heard about Sony exploring the possibility of a PS4 Pro 2. It was just one of the setup they explored and it was abandoned later.

screen12.jpg
 
Last edited:
Status
Not open for further replies.
Back
Top