I think the other thing getting lost here is this…
Microsoft can’t put out that tweet without AMD okaying it.
Just saying.
15 Likes
I’m starting to think maybe Sony is going for a PS5 / PS5 Pro deal in a few years, PS5 will get the early lead on games because of the GDK Xbox has put out late so it will take some time to adjust, then by the time PS5 Pro comes out it’ll be RDNA2 or w/e and all the learnings from XSX will help with the PS5 Pro.
I think we might be on a HW war lol Sony will want to fix the PS5 ASAP if it truly is missing pretty big features, then MS wouldn’t want them to be ahead so will also release a new console. I seriously would not be shocked if we’re gonna have consoles every 3-4 years if they go for this tit for tat strategy.
3 Likes
This makes sense. Some early games might look better on PS5, but chances are XSX’s advantage will become more and more noticeable as time goes on, assuming third party developers will bother customizing the game that is. But if they already have the various graphics settings from the PC version, figuring out which console runs which is no biggie.
2 Likes
I suspect the caching arrangements for Apu and discrete gpu will be different. Also Sony and MS are never going to stump up for that much video cache when they are fitting a cpu and custom blocks (such as acceleration for decompression) on the same chip.
1 Like
I was just trying to imagine last night, whilst being unable to sleep, what the reaction would be had Sony tweeted this out following the AMD event yesterday. I can honestly say it would have been social media full of the usual nonsense, other forums with hundred pages of concern and breaking down exactly what tech Xbox had used and why it was inferior and why they’d cheaped out…and I can’t honestly say hand on heart it would have been anything other than that.
We do need to wait for a bit to see what the longer term implications are for games but at this point if someone ran empirical benchmarking and showed a big gap between them the benchmarking would be called ‘inaccurate’.
6 Likes
It would have initiated the biggest Gaming media ‘circle Jerk’ in history.
6 Likes
Absolutely this, I’ve seen some come at me and others for pointing to the fact that the Series consoles are the only full-fat RDNA 2. I provide the statements from AMD and MS, and an responded to with vague screen grabs taken from who knows where with no citations. It’s sad really because we all know the PS5 will still be a great machine but the statements and DF analyses are all pretty clear on the state of the things; Microsoft simply hedged their bets on waiting for the architecture to be finished, whilst developing the suite with AMD, therefore creating will be an even more powerful console than what the spec sheet implies.
5 Likes
Oh dear time to run for the shelters, Metldown happening in 3…2 … 1
For those of you without a shelter?
HIT THE DECK!
I joke of course, please don’t feel too hurt Sony fans! I’m a big fan of their phones :).
It just is funny how some are already going :“Well Sony has their own implementations of this! Cerny said so”, I mean really what must happen for this man to become a normal man again, he really is talented but Sony fanboys have him up there in Sainthood as if he never can do ANY marketing speak.
2 Likes
Everyone, as it was mentioned before, let’s cool it on the console war rhetoric. There is absolutely no reason to be talking about people in the industry or even comparisons to the PS5 because it serves no purpose. Please keep the discussion on the Xbox hardware and what could mean for games running on it.
4 Likes
The problem is that this actually makes the Series S a lot more capable than given credit for. Just a few days ago people were again on the band wagon of hating that “small little engine that can” due to some raytracing stuff not being implemented (While other devs do implement it). Saying it would hold their Sony console back. But now it seems to be more capable than given credit for. When you know that “FLOPS” are really mostly about graphic calculations at a given resolution (I develop CAD systems for a living for big companies, and have programmed a few 3D engines), you know that it won’t hold back anything at the resolutions it’s intended to run at.
So while we shouldn’t be fanboys, I do believe you cannot get around the fact that people will compare it to Sony’s PlayStation 5 console. And that the feature set will make the little system a lot more capable than first assumed. Certain features will simply help with optimization, features the PS 5 simply won’t have. So while the PS5 is a very capable console, there is no doubt that any image it can produce at 4k? The Series S will be able to reproduce at 1080p! If the developer chooses to keep in mind the feature set that it offers.
Again “FLOPS” in GPUS are really a bad metric to look at.
5 Likes
FACTS. I thought I was the only one. People on twitter share with me some paragraph. IDK where it came from, who is saying it, or even the context lol
3 Likes
Is it bad that I actually want a Series S for my office? I’ve got two X’s preordered for my wife and I, but I’m genuinely curious about its capabilities (plus my office TV is 1080p). I think that the GDK’s scalability is going to be pretty awesome to see, especially given that the Series S is still a beast of a CPU with great memory bandwidth.
I"m not the judge of that, I can’t even have an X due to the download requirements (I live in a city were the ISP’s refuse to put in any new cables and are actually taking out copper cables. YAY privatization of government services! :P).
I do think the Series S is a wonderful little machine, and for smaller screens (Like in Little Timmy’s bedroom) more than capable. It is also why this crap at “Resetera and grannies” annoys me to no end. No Grandmother is getting “LIttle Timmie” a Series X for Christmas, not at 500 Freedom Dollies! I mean she usually has more than one grandchild, at several families. Each time I brought that up, you’re ridiculed as “It is a serious problem! Little Timmie will be angry” . Well Little Timmie is a spoiled brat then! :P.
However the Series S gets to a point where a Family might just buy one for the whole family as a console for Christmas or a birthday. People forget, Gaming is an enthusiast hobby, and most people don’t spend that much on it. So then the little machine becomes a very interesting proposition. Especially when you’re in a store, and you see both on display and the image quality doesn’t seem to differ that much (outside of resolution which on a smaller screen will be a lot harder to see).
Realism is lost to many gamers on forums (no offense intended to you personally), budgets for families exist and for most a cheaper alternative like the S will be the better option. For goodness sake, we got a Master System in 1993! SEGA was still selling them back then as a “budget console!”.
But yeah… it’s a cute little console, with a feature set that will get it a lot closer to the PS5 than people will think is possible if they just look at the Floating point numbers, especially if the resolution doesn’t go above 1080p :P.
PS: It could also be a nice streaming machine from your Xbox Series X to your S within your house. Although I’d wait, as Microsoft seems to be rumored to actually wish to launch a “streaming box” as well for like under 100 Freedom Points/Europoly monies.
1 Like
While it’s clear why the narrative is being pushed, there’s precedent for that. A hardware implementation not necessarily ensures a tangible better performance than some of the workarounds you can do with hardware.
For example we have seen people using virtual texture and virtual geometry ever since 360, which lacked any of the hardware support the implementation in rdna2 has. It was still able to achieve similar results and increase performance over the old pipeline.
Also for example, something like mesh shaders. It’s getting rid of some old fixed function pipeline and handling the geometry entirely in compute. But since it’s compute you can basically mimick the same approach in any compute enabled architecture including gcn.
In this particular case, there are some limitations. One of the features to support mesh shader per hardware is for example that now a compute shader can directly feed the rasterizer. So you get the whole flexibility and a first citizen aspect when doing graphical workloads on compute. That direct path is not available on old architectures but can be circumvented by having an intermediate buffer generated by the compute shader and a subsequent small vertex shader that will just read this buffer and feed the rasterizer. Is it more complicated and slightly less performant due the extra copy? Yes. Is it enough to make a tangible performance difference? Honestly I think not, at least with the exception of some very specific edge cases.
And game development is like this all the time. Ps4, Xbone and even X lacks the hardware for Checkerboard rendering but with a “hack” in the AA hardware they are able to mimick that, with similar results and performance gains than what was seen on the Pro.
In this point Alex is 100% yes those are features that ps5 is currently lacking, but the real world implications versus a workaround on a architecture that lacks those features are not known yet, and might be indeed small.
Man; Reading the other place, I’m already getting tired of the defense force and the mind gymnastics to make sure that it’s completely on par! Nobody is bloody denying that, this is biggers news for the Series S than the Series X! People there are so stuck in their bubble of being a “Gaming Enthousiast” that they completely forget what this means at lower resolutions and that developers will have means to get a lot more out of the transistor count in that system than given credit for. VRS alone will really help at lower resolutions, let alone Direct ML and other tricks to smooth out pixels. Of course it won’t be as sharp as a Series X or a PS5, but you’ll get a lot closer than people think.
It seems to really go over their head that this console is “not for them” in their privileged bubble, thinking it’ll “hold their toy back”. When really it’ll get next gen in a lot more households and “drag their toy forward into next gen features, kicking and screaming if it has to!” as people switch earlier from an Xbox One/PS 4 to this cheaper machine.
I really don’t think devs are gonna leave performance on the table and completely ignore VRS, mesh shaders, SFS, and machine learning later on in the gen even for third party games. All those hardware features being used in conjunction will have a compound effect in terms of performance gains, I think.
2 Likes
Of course not, it is part of the API, and most graphics engines will obfuscate it behind tool sets. So if they are working in say “Unreal Engine”, the engine will help them use the feature set of the Series S to gain performance. So it’ll be a lot easier than people think. It won’t be a “press of a button” of course, and testing will be required, but it isn’t something that should take manyears to implement. We are looking at a man month at most.
Most of the work will be done by the engines developers are working with.
2 Likes
I agree that flops aren’t everything but the comparisons are still useless. We don’t know all of the customizations in the PS5 and likely never will because it likely isn’t as feature filled as what we have in the Xbox consoles. How much games or performance benefits from these features may also not be anything earth shattering. This is just like the scenario with the Pro and it supporting RPM and an ID buffer. Those features were touted as major features but it never mattered in the end. I’m sure these features will help in the end but nothing that is worth the celebratory nature of this thread. That is especially true if developers don’t fully utilize the features since they are exclusive to Xbox. It’s for this reason that I don’t think anyone should be happy that these features are missing on the PS5. With more systems supporting these features, that would lead to higher adoption among developers and better support.
Really it comes down to us not knowing what this will mean for both the Xbox and PlayStation consoles, so any comparisons would just be baseless assumptions and nothing more. People here should be able to discuss the features themselves in isolation and make the effort to try to learn and discuss how they might benefit developers IF they are properly utilized.
I don’t think so either, especially since there’s feature parity with the PC implementation and DX12U makes the process even less of a hurdle. Microsoft partnering so closely with AMD and building DX12U to be platform-agnostic (insofar as Xbox and PC is concerned) will be beneficial to ensuring a lot of the new features are integrated into projects going further.
1 Like
I think the key differentiator from Sony’s previous situation and that of the current Xbox scenario is that the Pro’s features were only present on the Pro; compared to the RDNA2 and DX12U implementations that are also present in the PC space, thus opening up the possibility of utilization infinitely more than they were in Sony’s position.
6 Likes