Has anyone kept a log of Pachter predictions and how often they pan out?
I mean he has been this industry "insider" for years now with a thread on Neogaf every time he puts a prediction out there. I would love to see a breakdown of how often he is correct. If it is over 50% I would be surprised.
This would be fine and it's the most likely way this industry will continue. I agree.
Lets assume PS5 is coming Fall 2020 using 7nm FinFET - semi-custom Zen2 and Navi APU. PS5 Pro in 2023 or 2024 - on one of the proposed 5nm or 3nm nodes.
https://arstechnica.com/gadgets/2017/06/ibm-5nm-chip/?comments=1
PS5 Pro could also use AMD's modular APU design with cpu and gpu chiplets and future stacked memory. The GPU architecture can be anything. The same as what PS5 uses, or forward from Navi.
Whatever performance the base PS5 has, PS5 Pro would have 2x to 4x the GPU performance, the same CPU architecture but clocked quite a bit faster. It would function exactly like the PS4 Pro.
He's an industry insider? Like, access to inside information? I thought he was an analyst instead.
2019 could be the end yeah but I'd prefer this gen stays living till 2020.He said 2019 before, yeah. I think 2019 and 2020 are both equally likely.
Based on previous evidence- which is all we have- a late 2019 or early 2020 release is a pretty safe bet.
Last time MS got significantly behind they rushed into the next gen and beat Sony to the punch. Whatever the dates are, I doubt Sony will let that happen again. wouldn't be shocked at fall 2019 for both based on that. No idea how technology will impact things, though.
Sony aren't worried about Microsoft.
early 2020 was my guess too but I don't remember any actual evidence for it.
Someone suggested the idea of launching in September as PS's 25th anniversary which was interesting. I expect both Sony and Microsoft to have something out before the end of 2020.
May I ask why people are optimistic that AMD is able to come up with an overwhelming (or at the very least not underwhelming) console CPU next time? Seriously asking, I really have no idea if AMD offers more bangs for the buck nowadays.
He's an industry insider? Like, access to inside information? I thought he was an analyst instead.
Pro doesnt need to be a baseline. Itll still get most PS5 games from third parties because theyll still be using PC development and will cut back to support pro (and PS4) for at least a couple of years after PS5 launch.
Basically cross gen but so much simpler for developers that itll be even more common than PS3-PS4
Would be nice to see - but I'm not holding much hope yet. Considering you get 2+ speedup by default - for a small portion of an already small market, the motivation for specialized support is next to none.c0de said:Really? Well, that's interesting. Perhaps GTS will be a good candidate to implement it.
PS3 codebase deviation was rather significant, I guess the expectation is that won't be the case with next gen (akin to PC where codebases change slower and over much longer periods of time than has been the case traditionally on consoles).TheThreadsThatBindUs said:What specifically makes PS4 Pro cross-platform development inherently any easier than PS3 cross-platform development?
PS3 codebase deviation was rather significant, I guess the expectation is that won't be the case with next gen (akin to PC where codebases change slower and over much longer periods of time than has been the case traditionally on consoles).
Last time MS got significantly behind they rushed into the next gen and beat Sony to the punch. Whatever the dates are, I doubt Sony will let that happen again. wouldn't be shocked at fall 2019 for both based on that. No idea how technology will impact things, though.
So pachter thinks Sony will do it like Apple and their systems will be cross generational, if so it makes sense to draw the line somewhere and end up the support of the lowest model so they can take more advantage of ps5 power and won't be held back by the lowest common denominator (base ps4).
was gonna post something like this when i saw the quote. he changed his old 2019 prediction. can't in anyway see Sony putting out a PS5 before 2020. the pro allowed them to mitigate the new 7+ year cycle. we thought it was an anomaly last gen but it's probably what is required now to get the desired CPU change. 4-5 years simply isn't enough time anymore.
But the PS4 Pro will hold back the PS5...
The Pro still has the same 2013 notebook-level CPU. If the PS5 launches with a Zen CPU (which of course it will) the PS5 CPU will be light years ahead of even the Pro in CPU performance, meaning that even setting the Pro as a base min. spec. will be significantly hobbling next-gen games development for the next 7+ years...
...no thanks.
I really don't think that PS4 Pro will get its own version of next gen games if there is no base PS4 version. Base PS4 together with Pro will get the usual launch year crossgen ports and then they are abandoned for next gen only titles.
I really don't think that PS4 Pro will get its own version of next gen games if there is no base PS4 version. Base PS4 together with Pro will get the usual launch year crossgen ports and then they are abandoned for next gen only titles.
That and longer development times means nobody is ready to move forward a generation in 4-5 years.
But the PS4 Pro will hold back the PS5...
The Pro still has the same 2013 notebook-level CPU. If the PS5 launches with a Zen CPU (which of course it will) the PS5 CPU will be light years ahead of even the Pro in CPU performance, meaning that even setting the Pro as a base min. spec. will be significantly hobbling next-gen games development for the next 7+ years...
...no thanks.
Perhaps Patcher meant that PS4 Pro will be the only version available after PS5 launch, which I highly doubt since it's not for the mass market.
Furthermore it has the least/smallest userbase between the two so why would developers even bother with that model, it's a business suicide.
if the PS4 Pro can be reduced to ~$199 by the launch of PS5 there would be no reason to sell the standard PS4 anymore. People wouldn't be losing out on anything but an even cheaper price. People still holding on to base model PS4s wouldn't be left out because the pro games would still run on it regardless. them sharing a CPU means almost no matter what backwards compatibility is ensured.
So TLOU2, Death Stranding and probably Sucker Punch new IP to be PS4's swan song games?
how? the PS3 didn't hold back the PS4 and the PS2 certainly didn't hold back the PS3. The change to a more PC like architecture allows devs much more freedom in scaling. You will see instances of games not only looking better but having different tangible features because of CPU disparity.
You're misunderstanding what a "PC-like architecture in a console" actually means. It certainly doesn't allow developers more freedom in scaling software. The developers own game engine and production pipeline achieves that.
The biggest benefit of the traditional console generation is the hard cut-off and significant raising of the minimum target specification that a new-generation of hardware provides.
Whilst the PS4 Pro provides greater performance than the base PS4 model, the CPU is essentially the same. It's also significantly weaker than a next-gen Zen-based CPU would be. So while, with the transition from PS3 (with a fairly weak CPU) to PS4 (again a fairly weak CPU), game experiences didn't seem to differ too much in terms of simulation complexity (i.e. the 3D simulation driven by the CPU) and the biggest improvements were in graphics, a next-gen Zen CPU in a console promises to offer desktop PC-level CPU performance; thus an improvement much more significant than the last generational transition.
If developers are mandated to continue to support the PS4 Pro as a minimum spec. then game simulation complexity simply cannot move forwards, because the Jaguar CPU simply cannot compare to a Zen. Allowing developers to focus on Zen development (later in the gen when PS5/NxtBox become the exclusively targeted platforms), allows developers to create games they simply couldn't run on a PS4 and by extension the Pro... not because of the GPU performance, but because the CPU simply wouldn't be able to cope at all.
When it comes to games, whilst game engines today are probably more scale-able than they have ever been, games themselves are not infinitely scaleable. If the developer envisions a game designed around simulating fluid mechanics, for example, that a Jaguar CPU simply couldn't run at real-time framerates, then in a situation where he/she is forced to target the Pro as a baseline min. spec. he/she would either have to forgo the design for the game entirely or essentially gimp it on that platform to the point in which it becomes a completely different game (i.e. a non-starter for developer who values their reputation and the quality of the products they produce).
Last time MS got significantly behind they rushed into the next gen and beat Sony to the punch. Whatever the dates are, I doubt Sony will let that happen again. wouldn't be shocked at fall 2019 for both based on that. No idea how technology will impact things, though.
i think you're confusing my point. Zen v Jaguar obviously has a large disparity, but my point is that it would not hold the PS5 back. If the developer wants to do something the PS4 cannot do, that wouldn't be a cross-gen game. It would simply be a next gen exclusive. Sony would never mandate that every PS5 game must run on PS4. That would be foolish. But every dev is not creating super magical CPU heavy games that cannot run on that older hardware. Sure the PS4's CPU was old by PC standards when it released but it was still a difference from the PS3 and here we are today with amazing games that aren't held back by the PS3. A more similar structure eases the developers work I'm assuming here but holding it back is a pretty strong assumption in itself
The Driveclub thread that was posted a few days ago had me thinking about how much better games can look with perfect weather & lighting effects & now I'm really hoping that PS5 take a page from PS1 / PS2 /PS3 book & have a coprocessor for the effects.
Why do you need a co-processor for weather effects? Why not spend that portion of the silicon budget on more GPU ALU; keeping everything tightly integrated on a single processing unit (with its associated memory subsystem)?
https://www.youtube.com/watch?v=BYEgitCAcys
I looked it up and they are indeed recruiting someone for some next-gen game for current and next-gen platforms.
One for the oppertunities: http://avalanchestudios.com/careers/7837/
So shit is in full swing behind the scenes, color me surprised?
Because it could be a specialized processor with it's own cache or whatever & just do the job more efficiently.
Don't get stuck on whether PS4 was easy to adopt or not. The costs at that point were still disproportionately high - divergent code-bases and wildly incompatible hw targets (not unlike past gens really).TheThreadsThatBindUs said:By that point, every developer had spent the last 7 years maturing their engine codebase and tools for PS3, so developing a game simultaneously for both PS3 and PS4 was no issue at the time
Can you elaborate on that? What exactly happened with the CPU decision-making process?That's not really what happened.
Pro doesnt need to be a baseline. Itll still get most PS5 games from third parties because theyll still be using PC development and will cut back to support pro (and PS4) for at least a couple of years after PS5 launch.
Basically cross gen but so much simpler for developers that itll be even more common than PS3-PS4
(i) I'm not sure if computing weather effects would really benefit from dedicated fixed function processing units in silicon; as there isn't any set or fixed way of doing these effects, but rather multiple, which means devs are better off having less specialized, more general computing cores to do the heavy liftingmeaning GPU is likely more than up to the challenge.
(ii) Adding in specialized fixed function or general processing cores for off-loading tasks, makes sense when the workloads being offloaded are largely independent with limited data dependency across the rendering pipeline. For something like weather effects, by moving tasks like that off the GPU, you incur penalties that would make the overall rendering less efficient, as well as incurring additional cost and complexity in your APU design.
E.g. look at the problems the SHAPE audio unit and MOVE engines created with the OG XB1. All these additional processing blocks did was consume die-area which could have been dedicated to more GPU ALU and resulted in a console that would perform much closer to the PS4 at launch.
In a fixed size, console APU, die-area is your most precious commodity. Specialized cores need their own cache and memory takes up a significant amount of die-space that you really want to be devoting to maximizing the area taken up by your execution units.
I definitely agree with this. I fully expect PS5 to be the start of a new "gen" and have exclusive first party games, but the normal third party offerings will be cross-gen for quite some time. I expect another PS2 situation where at the very least, we'll see third party sports games on PS4 for 2-3 years into the life of PS5.
Can you elaborate on that? What exactly happened with the CPU decision-making process?