Next-gen PS5 and next Xbox speculation launch thread - MY ANACONDA DON'T WANT NONE
- Thread starter Phoenix Splash
- Start date
- Status
Capacity and supply I guess. The suppliers have been focused on other memory types, and have been charging a premium for HBM. You also need an interposer, which throws a little more on top. The design work is more expensive also, although this isn't so much of a deal over a multi-million-unit product I guess.
I suppose if higher volume customers came along and more suppliers switched more capacity to HBM, the price might come down. I believe it's expected more capacity will come on line for HBM. But I haven't seen any forecasts about pricing.
Through silicon via. It's basically an interconnect that goes all the way from the top to the bottom of a die.
It's fun to speculate about these setups, but I think things are far from accepted wisdom right now. Patents, reddit posts etc. etc. all in mind.
16GB of GDDR6 still seems as likely as anything right now IMO, if not more likely. It's expensive by the standards of memory last gen, but it's still the cheapest of the scenarios offered here also (I think).
Man you want the PS5 to be stronger and cheaper than the XBOX so bad.
And you're a childish buffoon if that's what you read from my post.
https://www.youtube.com/watch?time_continue=1185&v=5jD0mELZPD8 well I hope a few indie games do path tracing with w/e hardware is on both consoles. this looks so so good.
True, but from what others have posted it seems that HBM2 would have more potential for cost reduction throughout the life cycle of the console, no? If they are going to wind up selling a $500 machine and are taking a loss on the initial price then maybe that could be part of the higher price?
Now I'm just talking out loud here, but if GDDR6 has much less room for cost reduction over time, does that mean it might be closer to being on its way out to be replaced by something newer? (ie. HBM) Over the 10 year life cycle of a console, if GDDR6 is no longer really used because HBM has become much more commonplace then wouldn't that possibly keep the price of the console high because they have to source parts based on an outdated memory type that's not used in many other devices?
Also guessing there's absolutely 0 chance of HBM3 in next gen machines? :)
no, I cant see hmb replacing gddr6 in most gpus, so if anything as more gpus start using gddr6, it should push the price down.
Ahh, so there is room for a decent amount of price reduction still on GDDR6?
I doubt very much that Microsoft would take any loss even if they sold the Xone SAD at $199.
And this story of not worth it because Sony has already won does not make much sense. If that is the case, they would not even have spent money on the X project, or even SAD. Sony in the last gen. would not have made the PS3 Slim and Super Slim as well.
how much I dont know, but it will be the defult for at lest sevral gens of gpus, so.
Just FYI...6TF on Jaguar based silicon and 4TF on Zen based silicon are two different things. All TFLOPS aren't created equal, so your argument isn't as clear cut as you make it. But we shall see the games in a year. :)
again its 6tfps with rt hardware. people should watch the df video just posted today on minecraft path tracing. btw that doesnt even use the rt hardware, imagine ms releasing a version that worked with the rt core.
Well, I think 'everyone' is saying that because that's how the reddit rumour initially helped to rationalised the choice. It's really hard to verify the likelihood of that... there is talk of HBM supply opening up, and of course if a mass device or devices created demand it would help, but I have no idea where hbm memory prices will go, or how well anyone outside of contract negotiations for these kinds of things can gauge it.
IF it's the case that they see a significantly better cost curve on HBM vs alternatives it would be a good reason to use it even if short term it was more expensive. But whether they see that or not... I dunno.
TeraFLOPS are exclusively a measure of GPU compute power, at least in the context we talk about it here (because CPU TeraFLOPS are trivial and don't account for the majority of what CPUs do). A 4TF GPU with Zen could more easily handle games with higher (60) frame rate games than a Jaguar console, but a 6TF GPU with a Jaguar CPU would be capable of higher graphical detail, despite its slower CPU.
In modern systems, geometry draw is primarily limited by the CPU (which is why no one bothers publishing triangle per second numbers like back in the PS2 era) and most improvements in GPUs come down to higher bandwidth and more RAM for higher resolutions or higher resolution textures or more powerful general compute power for shader, lighting, physics compute and a number of other extraneous things. Ironically, this means that the more graphically simpler your game is, the more your CPU is what holds back the average framerate. There's also been some advancements to try and decouple GPU performance from CPU performance more, by having the GPU itself handle more of the geometry draw operations, but the results of that haven't been terribly conclusive in PC performance.
that isnt fully true. as fp16 becomes more common, gpus that support it (vega rtx, navi) get better results at the same tfp. in this case a 6 tfp navi chip, whill beat a 6tfp polaris chip. add rt harware on top of that, and yes there will be a world of difrence dispite both chips saying 6 tfps.
The real question is will 6tflops navi be comparable to 6tflops GeForce or not?
we will find out in a few months, though its less that a rx 580 is worse than a gtx 1060, there are a bunch of games in which the 1060 loses, its fore it doesnt scale well. its why it takes such a power hungry card to beat a 1080 in vega 64. hopefuly navi is better.
Yeah, too many factors.
anexanhume explained a lot of things well on how it could be credible but... we need leaks from more credible sources to see how they'd go with it.
As he states Digitimes reporting that Playstation 5 will be packaged and tested by ASE and SPIL does contradict what is posted in that post too.
I do think that both next gen systems will use fan out packaging though, and I won't be surprised if it's Microsoft that goes with InFO_MS instead. Also I can't believe I spent a good time trying to find out if SPIL and ASE has fan out packaging tech.
A comparison was made to the point that Lockheart at 1080p would look crap compared to the PS4 Pro when it came to cross generation games because the latter would simply play them at 1440p.
Why should that opinion be taken seriously if Navi is supposed to be a more efficient version of GCN? It makes no sense, especially because we do not know what the comparison between the two will be. Also it is such a tedious argument that does not take into account what both consoles try and accomplish.
call me crazy, but I will 100% take 1080p with raytracing, than 1440p without.
See my earlier point: "MS are just profiting as much as they can".
I never said it wasn't worth it - I said the strategy of introducing the SAD as a loss-making console this late in the gen makes no sense.
Raytracing, better load times, better AI going forward, better physics once developers move from cross gen stuff. It is an opinion totally lacking in understanding of where Microsoft wants to go with gaming i.e. have a cheaper console that can play next generation games, a lower entry point to attract people that have traditionally waited for a price drop to get a console.
And I'm sure they're hoping it'll help them win back gamers who switched from 360 to PS4. Instead of waiting for the PS5 to drop in price people will be enticed to buy into next gen earlier with Lockhart.
Last edited:
The thing is, Sony has gotten touchy with underperforming divisions for so many years that re-constructuring these areas from the ground has let them to reset expectations and visions. A poster above brought up a good point with the 4k Bluray argument, that Sony left out 4k Playback from being integrated on the Pro, just so they could have a high-end console not take a loss and even out manufacturing costs. This looks good to investors, and red negative numbers (however big or small) in the hardware sector, does not.
Another thing that people should note, is that Hardware has been a sort of a stigma for Sony for so many years. That now that they are finally in a full profiting state with Playstation, I see them getting complacent with their new profiting hardware status.
For example, they've dialed down in price cuts substantially this gen compared to their prior ones even during holidays (No deals or price cuts last December), Pro didn't go a digit over $399, leaving out costly components to diminish manufacturing costs, etc.
Of course, I am open to the idea of selling powerful hardware at a loss. As revenue for Network Services for PSN are looking so mighty right now they could just as easily do the heavy lifting for offsetting costs / losses until parts come down in price.
And, the leaks so far have shown very competent hardware components that 5 months ago, people would have called you crazy for basing predictions within such a high spectrum. But more leaks keep popping out consistently in the same vicinity of high-end hardware, so perhaps times are changing.
In the end, there is course of action history from Sony being reductive with manufacturing costs, as there is course of action history with Sony selling hardware at a loss (PS3, PS4 to a lesser extent).
Just to play it super safe, I am leaning towards the same approach they took late this gen with being picky with parts just to hit the the magic $399 that gave them a thunderous applause at E3 and helped position Playstation as the traditional console market leader again.
That strategy was a monumental success for them this generation.
I know I hope that 10th generation consoles (PS6, XBox equivalent) around 2027 will be using HBM4 :D
Check out this projection from 2004 (from an article posted in July 2005) by Nvidia's then Vice President of Technical Marketing, Tony Tamasi.
https://www.tomshardware.co.uk/xdr2-to-quintuple-memory-data-transfer-speeds-by-2007,news-16613.html
Even the last versions of HBM2 (Aquabolt and Flashbolt) cannot reach 3TB/s AFAIK. HBM3 might, but as of 2016, it was estimated 64GB HBM3 would offer 2.4 TB/s bandwidth. So then, HBM3+ (6TB/s ?) and HBM4 are predicted (not announced) to happen between 2022 and 2024, offering more and more throughput for the data center's insatible appatite for bandwidth.
(old slide from 4+ years ago)
The problem that MS will face with the lower end Lockhart box is multifold, and depends too much on the competition.
A $299 box that plays next gen games is an appealing idea in a vacuum. A $299 box, presumably restricted to 1080p is a less appealing idea if the PS5 is $399 (vs. say $499, where the gap is notable). It's an extremely risky strategy, and I'm interested to see how the pricing of the PS5 and both Xbox boxes go, because it's very easy to price yourself out of the competition.
I'd wager that MS would benefit the most from the PS5 being $499 if Lockhart is $299, as that would give the Lockhart SKU far more appeal.
A $299 box that plays next gen games is an appealing idea in a vacuum. A $299 box, presumably restricted to 1080p is a less appealing idea if the PS5 is $399 (vs. say $499, where the gap is notable). It's an extremely risky strategy, and I'm interested to see how the pricing of the PS5 and both Xbox boxes go, because it's very easy to price yourself out of the competition.
I'd wager that MS would benefit the most from the PS5 being $499 if Lockhart is $299, as that would give the Lockhart SKU far more appeal.
That is what Microsoft seems to be gambling on. In theory, it sounds good, and it should work, but in the world of man things are not always straightforward.
What was the appeal to the Gameboy handhelds when Sony was putting out far better hardware in the PSP?
Why is the Switch doing better in Japan? Why is it selling at a faster rate compared to more capable consoles? Resolution is not everything.
Maybe. Plenty of cheaper systems have failed. They have to make it seem next gen enough too
Price isn't even the issue. The issue (imho) is developers would have to build their games with the lower spec machine in mind, putting them at a technical disadvantage again. In essence, Lockhart would be their next-gen system and Anaconda would just play superficially better versions of Lockhart games (like what the X is to the S today).
Resolution, power are not everything.
It wont be for quite some time because majority of consumers are still on 1080p sets, and there is not a lot of content out on 4K worldwide that you do not have to pay a premium for.
You develop for the higher SKU and scale down. It has already happened this generation. And as DrKeo said in the previous reincarnation of this thread, there is nothing to worry about so long as there is enough buffer in reserve i.e. instead of having the weaker SKU being 1/4 as powerful, have it as 1/3 or less. Everything would work, no issues.
Well they haven't ever really gone crazy with pricing at the start of a gen either have they? I mean the one was $500 and then $400 when Kinect went away
360 was $400 right? Wasn't sold at as much of a loss as ps3
OG Xbox I have no idea what the loss or gain was on that at launch but I think it was equal with ps2 at worst
I really don't see how much powerful anaconda can be to really show a difference from PS5 in games..... In PC space, many times, the difference between ultra and high are negligible and not worth it for the performance hit. I don't think that multiplats will vary too much between the two consoles. Microsoft must show the power through their exclusives. Can't wait to see Halo Infinite.
Let's say, for example, that Anaconda has an SSD and Lockhart has an HDD. How would a developer be able to make a game that takes advantage of the SSD and have is scale down to a system that technically can't come anywhere close? They would have to develop with the limitations of an HDD in mind, no?
if the next MS console has dedicated hardware for RT i sure as hell hope playstation does as well. i want that shit common place
On pc, people with hdd just load a bit slower, in Anthem I would load in 15 seconds into the mission before my friend would, in gameplay it didn't really affect
The OG Xbox was sold at a loss, as was the 360 - around a $125 loss for each if memory serves me correctly, but could be wrong.
The big advantage now is that there are so many additional income streams that allow console manufacturers to claw that money back. Online play subscriptions, streaming subscriptions, game library subscriptions etc. on top of the usual software streams etc. My guess is that both Sony and MS will take a big loss on hardware at launch.
Follow-up of a discussion we had and heads up to those who argued any kind of RT cores are confirmed because of the first tweet:
Everything points to both having SSD, only differing on RAM amount, GPU power and maybe storage amount.
Jaguar is a CPU and has nothing to do with the discussion you're referencing.
Also, you're wrong. A floating point operation is a finite mathematical operation performed by a computer; thus a FLOP by one computer is identical to a FLOP by another.
What I think you're trying to reference is the architectural features that make the difference in real world performance between processors, e.g. Nvidia and AMD GPUs.
Considering, however, that Navi is confirmed to be GCN your idea that any architectural improvements can make up a difference of 50% peak theoretical performance (i.e. 4TF vs 6 TF) is wishful thinking at best. I think you probably need to rein in your expectations a tad.
As much as I see how this appears. I'm half suspecting this follow-up tweet is purely damage control after the tweeter in question got a slap round the wrist for revealing stuff that hasn't been publicly announced yet.
I mean, the guy is a Tech. Artist at Naughty Dog for heaven's sake. If anyone knows what PS5 is going to be it's Naughty Dog.
- Status