Digital Foundry : PS5, Xbox Series X/S, PS4 Pro and Xbox One X under the microscope.
They can make all the excuses they want but the PS5 has shown in quite a few head to heads that its the stronger machine.
I think we need to be more than 6-7 months into the generation to make a call on that...
Not really? The only time Xbox has outperformed PS5 is in backwards compatibility. In every next gen game the PS5 has better performance. That's not going to magically change in a year.
@Magog I'm optimistic PS5 games will continue to look better, but we can't say it's a done deal yet. Remember that in the first 18 months or so of its life cycle, PS3 versions of games frequently looked and ran worse than their 360 counterparts, but by the end of the gen PS3 had the most impressive looking games. Also it hasn't been universally better on PS5 so far; Hitman 3 looked better in a few minor ways on XSX (though there were some associated performance issues, so it could be argued that it was pretty even).
Mag you might want to check those facts. Xbox has been finally showing to be the stringer system. I mean, it literally is the stronger system, and it's finally showing.
The Xbox One X predictably out performs the PS4 Pro The PS5 predictably out performs the Xbox Series X
@GamerRN Xbox has been finally showing to be the stringer system. I mean, it literally is the stronger system, and it's finally showing. Any information to backup your statement/opinion? Where has it been showing to be the definite hands down stronger system other than specs? You say finally like we should know what you are talking about that "finally" shows its the stronger system?
@jeece that's not comparable at all the ps3 had totally new tech that had to be learned to program for it. these new consoles are virtually the same tech and sony devs never had ab problem showing off first party games only third party games had problems because ps3 was hard to work with since it was brand new tech
@jeece well the 360 had a better GPU and more memory , the main downfall of the 360 was them using DVD Roma instead of going with BLU Ray
By the time it would even matter, the mid-gen refreshes will be launching.
the 360 did not have the graphical capability that the PS3 had, nor did it have more memory. It had unified memory and a more familiar architecture. if you JUST looked at the GPU vs. GPU, the 360 narrowly squeezed out the ps3. The problem though is that logic is flawed. The cell processor was meant to carry a portion of that weight when used correctly. and when it was used correctly, it walked all over the 360. as for the debate between the ps5 and xsx, they have very similar hardware, outside of sony's custom stuff. the thing is, sony's custom stuff is pretty brilliant and cerny is basically a wizard.
Seriously? At this point, the differences are so marginal for 90% of the games, it's ridiculous. Just get the console you want to get and enjoy it. Who cares what version is better? What matters here is gamers get more "choich" in what's available to them regardless of the platform. As long as they're similar, then they can be enjoyed. I don't think any Switch gamer would be upset that they got a port of Doom Eternal or Persona 5 Strikers.
@JEECE That was because the PS3 was a beast to develop for. The Series X and PS5 both use essentially standard PC parts but the PS5 is clocked higher which leads to FPS advantages.
Lmao people here are clueless. This is exactly why we are seeing more stable frame rates on ps5's. You got 5 disagrees and 0 agrees for straight up talking facts. Shows how much people know on this site.
You are correct on that - higher clocks will make it easier to get stable framerates in the short term/early in the gen, but having higher parallelism / CU count will pay off long term for performance & visuals, just like it did with PS3. Mid-late PS3 games destroyed 360 games. XSX is somewhat similar to PS3/CELL in that respect (albeit less complex and cheaper to manufacture) - it's going to take more work/effort to learn how to optimize for it and get the most out of it, but when one does, it will have a big advantage.
@Orchard " Mid-late PS3 games destroyed 360 games." No, they did not. Most multiplats favored 360 all gen long. Hell, Skyrim and RDR were late gen games that looked horrible compared to 360. "XSX is somewhat similar to PS3/CELL in that respect (albeit less complex and cheaper to manufacture) - it's going to take more work/effort to learn how to optimize for it and get the most out of it, but when one does, it will have a big advantage." Well, that may or may not pan out the way you think, but the comparison of XSX to Cell is way off. Cell was complex not because of the number of processing units, but the fact that you had one primary (PPE) and seven specialized units (SPE or SPU) that didn't even run the same code as the primary. This was completely different from the 360 Xenon that had three cores that were exactly the same and thus, much simpler. Cell was a different paradigm in programming entirely. This and the relatively weak RSX made PS3 suffer in multiplats. Regardless, game developers are able to optimize PC gaming GPUs that have more CUs than XSX just fine and increase the performance of the same exact games way beyond what XSX's 52 CUs can do. So they already have that capability. Where is all this extra "work/effort" that is supposedly needed for XSX?
@RazzerRedux Yes they did, late gen PS3 games for the most part, looks great compared to 360 equivalents. Lots of great looking exclusives towards the end too which showed off things the 360 couldn't even dream of rendering. "but the comparison of XSX to Cell is way off" Not really. The extreme parallelization & non-unified memory pool made the PS3 hard to work with. XSX also has extreme parallelization (on the GPU side), and has two effective pools of memory. Writing SPU programs/jobs was complicated at first and caused me many headaches, but once you got comfortable with it, it was a lot more powerful than 360. "that didn't even run the same code as the primary" You wrote the same code on PPU and SPU, it was more that you had to work out what you wanted to offload in terms of jobs, and then share relevant data back and forth via DMA. It was hard to design systems where one task/core wasn't dependent upon the results of another, causing stalls within the current frame and/or full out hangs. "Regardless, game developers are able to optimize PC gaming GPUs that have more CUs than XSX just fine and increase the performance of the same exact games way beyond what XSX's 52 CUs can do." This is a very naive view from a non-developers mind. Devs don't just pick up hardware and immediately write heavily optimized code for it - especially custom hardware like a games console. Like every gen in the past, it will take time for devs to max things out for both consoles, but there's no denying that XSX is harder to develop for than PS5, far more CU's to offload jobs to, and a non-unified memory pool. PS5's architecture on the GPU side is brute force, and by extra work, I meant that devs will have to learn the Xbox architecture and how to exploit those extra CU's, as well as the fact that there's some extra work to optimize your game on Xbox vs PS5 (which devs may or may not be willing to do... let's see). Brute force power will always be easier to optimize initially than having to do smart optimizations by offloading work - ironically, that is also why a lot of software & games don't benefit heavily from multi-core CPU's on PC but benefit more so from higher clocks and less cores. Effective parallelization is hard - but it's obvious to anyone that 12TF, even when implemented in a complex way, will eventually outperform 10TF.
"@RazzerRedux Yes they did, late gen PS3 games for the most part, looks great compared to 360 equivalents." Such as? Here's RDR that I mentioned. "Despite already compromising in terms of resolution and frame-rate, it's clear Rockstar had to cut and pare back the PS3 version of the game still further in order to maintain performance. We can see a fairly obvious array of smaller changes between the two versions that almost always favour the 360 rendition of the game." https://www.eurogamer.net/a... "But, despite their best and most well-meaning efforts at achieving parity on both consoles, we can confirm that the 360 does command a clear advantage over the PS3 when it comes to both image quality and performance. " https://www.eurogamer.net/a... So show me the comparisons where PS3 games "destroyed" the 360 versions? "You wrote the same code on PPU and SPU" They even had different compilers. "It was hard to design systems where one task/core wasn't dependent upon the results of another, causing stalls within the current frame and/or full out hangs." That was a blatant rewording of this: "Such dependency chains require one task to pause and wait for the results of another, creating stalls." From the introductory section on the cell processor "programming model" within a white paper on ray tracing in the cell. https://www.sci.utah.edu/pu... If this were a college course you would be suspended for plagiarism. "This is a very naive view from a non-developers mind." Then I am in good company. I never claimed to be a game developer, but I do read quite a bit for a layman's perspective. Obviously I don't buy this bullshit that you are a game developer. Especially when I can damn near reproduce the technical jargon you stole with a simple google search. "Effective parallelization is hard - but it's obvious to anyone that 12TF, even when implemented in a complex way, will eventually outperform 10TF." And yet, real developers don't put much stock into teraflops at all. This, above all else, is why it is obvious you are a fraud. When you yapp about teraflops it is clear that you are nothing but yet another wannabe from the multitude of fanboys roaming the planet.
@Razzer “They even had different compilers.” Different compilers doesn’t mean different code. It just means that code may be compiled into different processor instructions (and some intrinsics may differ). PS5 and XSX have different compilers but both compile C++ code. Syntax is the same for those consoles, but different libs and headers are available (also due to one being Windows based and one being FreeBSD). The PPU code and SPU code were both C++ run through the SN systems compiler. “That was a blatant rewording of this” Except it came from my head, because I’m an experienced engineer in the industry :) Also CPU hangs are just a common problem/part of working in a multithreaded environment. You can probably find similar quotes all over the internet - even for other platforms/devices. RDR is just one game, look at any of the ND games towards the end of the gen - they mastered the hardware and the 360 could only dream of rendering similar results. “Obviously I don't buy this bullshit that you are a game developer.” Shrug, I’m more than willing to prove to anyone on here that I am a dev. Send me a DM if you want proof - but I don’t expect that you will because having the truth confirmed wouldn’t suit you :) And any real engine dev understands teraflops are important. Anyone saying otherwise is talking BS - there’s a reason why we aren’t still gaming on GeForce FX :)
"RDR is just one game" I included links to Skyrim as well. There are plenty more. You did not provide one example while your claim stated mid-late PS3 games. That's a lot to choose from. Remember, they "destroy" 360. "look at any of the ND games" lol...what? You explicitly said "late gen PS3 games for the most part, looks great compared to 360 equivalents" and now you want to yapp about Naughty Dog? Hilarious. "Different compilers doesn’t mean different code. PS5 and XSX have different compilers but both compile C++ code. Syntax is the same for those consoles, but different libs and headers are available (also due to one being Windows based and one being FreeBSD)." C++ is the language, not the code. Code is the implementation of the language. Different compilers are created to compile different instruction sets of code. SPUs have their own distinct instruction set and do not even share all the same data types as the PPU. From just some basic google research, PPU and SPU used entirely different templates for compiling for PS3. "Shrug, I’m more than willing to prove to anyone on here that I am a dev. Send me a DM if you want proof - but I don’t expect that you will because having the truth confirmed wouldn’t suit you :)" You aren't blocked from sending me anything. If you want to send me "proof" then why didn't you just send it? Hell, send it now.
@Razzer ND are a perfect example of devs who learned the PS3 hardware inside out and made the best looking games of the entire generation. Simply because the PS3 was a powerhouse if you could master it - and they did. Devs who put the effort into PS3 yielded good results - the problem was a lot didn’t put the effort in or did cheap outsources for the PS3 version (e.g. the orange box). Will devs be willing to put in the extra effort to max out the extra perf capabilities of the XSX? That remains to be seen. PS5 is the lead platform for a lot of games due to the sales so they’ll likely just get XSX to similar levels of visuals and perf and leave it there. C++ is indeed a language, but code is just C++ functions called in a certain order basically. The compiler converts those calls and each line of code basically into hardware instructions/assembly. SPU’s did have their own instruction set, but that’s not really unique. Working on PS5 and XSX, they both have unique instructions which sure you could say are optional, but if you really want to max out performance you’re going to have to know how to exploit the hardware. Most standard data types were available to both SPUs and the PPU. The process of sharing data between them was a bit convoluted though. Not sure what you’re talking about with compiling templates - both PPU and SPU support templates - most C++ compilers support templates. Well do you accept a dev kit pic with custom message / drawing / whatever as proof of being a dev? If so, send me a message or drawing or whatever you want on the post-it and you’ll get your proof as soon as I’m at work.
Orchard The pinnacle of graphical fidelity in the 7th gen wasn't any PS3 exclusive. It was multiplatform game called Crysis 3, a game that on PC rivaled early 8th gen games and the 360 had the better version on 7th gen consoles. The 360 outperformed the PS3 throughout the whole generation and the people that at the time paid attention and came to websites such as this one know this to be a fact so don't give us your crapola about the PS3 rendering things that the 360 couldn't.
@Orchard You failed to provide an example of a single PS3 game where it "destroyed" its 360 equivalent. You don't know what the hell you are talking about and you know it. The fact that you keep going back to ND when it doesn't fall into that criteria at all just further damages your bullshit claims. "You wrote the same code on PPU and SPU" "SPU’s did have their own instruction set, but that’s not really unique." Your story keeps evolving. Having its own "instruction set" means it has "different code". I've looked documents detailing the spu_instrinsic.h header. PPU has its own functions as well. They are not interchangeable at all. You should know this. "Working on PS5 and XSX, they both have unique instructions which sure you could say are optional, but if you really want to max out performance you’re going to have to know how to exploit the hardware." That's really your take on the difference between developing on XSX and PS5? No discussion on DX12 or PSSL at all? Somehow you simplified those APIs as "unique instructions"? Seriously? "Most standard data types were available to both SPUs and the PPU. " Most, yes, but not all. That's why they had different compilers. You should know this. "Not sure what you’re talking about with compiling templates" Obviously I'm talking about IDE templates. Compiling to different processors requires different templates. Templates designate the targeted processor. You should know this. You have demonstrated absolutely no knowledge that cannot be obtained by google alone and you still got most of it wrong. When I read what real developers have to say about game programming, it wouldn't matter how much I googled, I'd have no clue what they are talking about. But as a non-gaming developer, I know more than enough to call out your bullshit. "Well do you accept a dev kit pic with custom message / drawing / whatever as proof of being a dev? If so, send me a message or drawing or whatever you want on the post-it and you’ll get your proof as soon as I’m at work." Please. Being able to take a picture of a dev kit is not proof that you are a game developer. Considering I've already shown you to be a plagiarist (which you do not deny) then it is obvious you'll go to any length to "prove" this bullshit fantasy. @Sol4ris lol...oh stop your bitching. I was highlighting the article as it related to performance. If you feel that paragraph was pertinent then fine. Get over yourself.
razzer, it wasnt the RSX. that was a capable chip and very close to the performance of the 360s gpu. the cell was meant to carry a significant amount of burden away from the RSX. it was the unified memory + development choice that lead to some multiplats running better on 360. games that were basically designed for PC ran a little better on the 360. games that were designed for the ps3 did things visually that the 360 just couldnt do. Raz, the games that absolutely put the 360 to shame werent available on that console. pit the 360s best visuals against the ps3s best, and its not a contest. 360 was great for easily developing multiplats(not a bad thing) but just did not have the muscle the ps3 had. period. this arguing is really dumb because we already covered this what, 10 years ago?
The game performs a little better on PS5. Regardless you’d think it would be a locked 60fps on both PS5 and Series X.
It would have been quite the surprise if Sony have developed a game that ran better on their competing hardware.
Actually, Sony did just that. "I can well believe that the conspiracy theorists will be full-throated in pointing out performance differentials between PS5 and Xbox Series X that are somewhat at odds with the technical specification of the consoles, but the fact that Xbox One X gives a smoother ride than PS4 Pro hopefully underscores that San Diego Studio gave the Xbox ports a good shot."
It hit or miss either way. Studios haven’t been given much time to work with titles solely on the next gen system’s especially with covid delaying things. It’s going to take a while for things to shake out.
Are you delusional? How many DF vids will it take until you realize that it’s not? The only thing Sony has going for it is it’s exclusives, VR, and the dual sense controller. I like the PS5, but let’s not kid ourselves, here.
Lol....the 'only things' PS5 has going for it are exclusives, VR and a revolutionary controller. Any one of those is reason to buy a PS5 over the competition.
and sound, and loading..
Xbox has shown it's stronger in a bunch too, and is stronger on paper.
PS5 XSX Native 4K. "Basically identical" "PlayStation 5 and Series X render proceedings at native 2160p resolution and barring some very minor differences (a slight difference in select filtering? An odd water rendering bug in one scene on Series X?) they are basically identical." XSS 1080p Lower detail level Performance -Frame rate dips "harder" on XSX than PS5, sometimes up to 9-11 frames per second. -Series S at 1080p operates much the same as Series X, sometimes a little smoother but more often than not, 2-3fps slower. Interesting tidbit: "I can well believe that the conspiracy theorists will be full-throated in pointing out performance differentials between PS5 and Xbox Series X that are somewhat at odds with the technical specification of the consoles, but the fact that Xbox One X gives a smoother ride than PS4 Pro hopefully underscores that San Diego Studio gave the Xbox ports a good shot."
Appreciate df does this , but were we really expecting big difference in a baseball game this early in the covid generation of crossgen...?
@RazzerRedux "Remember, its engine was developed on PlayStation hardware and when it comes to Series consoles, we're still very early on in the generation - and development tools are immature, especially on the Microsoft side. PlayStation 5 does have sparing DualSense controller support which is not matched on Xbox, while loading times are faster: a game loaded in eight seconds on PS5 up against 12 seconds on Series X. Initial boot times are also very swift on PS5 but only incrementally faster on Series consoles, " Your attempt at objectivity has failed spectacularly by not including the quote above. It is understandable, since it doesn't fit with your biased narrative.
Given they're basically identical, the most interesting part of the article is actually: "It's a fascinating, if not overly exciting face-off when it comes to the game code, but the battle of the console business models is where the sparks really fly. I'm not a baseball fan so can't really comment on the quality of the simulation but it's a no-brainer that Game Pass will see the title more widely played - if you've got the sub, you have nothing to lose by sampling it whether you're a die-hard fan of the sport or not. Microsoft is certainly marketing it heavily, awareness is high and the novelty of playing a PlayStation game on an Xbox console will be compelling. Meanwhile, on the PlayStation side, this is a full-price game with arguably niche appeal - certainly outside of the USA - and no real sampling mechanism. Charging £60 for the product certainly won't convince anyone who isn't deeply invested in the game to give it a go. Meanwhile, charging a premium for a PS4/PS5 double-pack sits uneasily with me when this is the epitome of a cross-platform development."
The "most interesting part of the article" rehashes a subject that has been covered thoroughly for weeks now?
Well the rest of the article is basically 'everything is the same', it's not exactly the most riveting DF analysis they've done. SD Studio did a good job on consistent performance across so many platforms - especially since over half of them were first time platforms/new hardware for them. And clearly it is an interesting topic, we've never been in this situation before - and probably won't be again for a while.
'everything is the same' Not true. PS5 has more stable frame rates than XSX and Xbox One X outperforms PS4 Pro.
Exactly. Pretty much all the same, and you can refer to my earlier statement: "it's not exactly the most riveting DF analysis they've done"
"PS5 has more stable frame rates than XSX and Xbox One X outperforms PS4 Pro." "Exactly. Pretty much all the same" Translation: https://media0.giphy.com/me...
@RazzerRedux You should go back and correct your post then, because you also quote the same sentiment: "Basically identical" "PlayStation 5 and Series X render proceedings at native 2160p resolution and barring some very minor differences (a slight difference in select filtering? An odd water rendering bug in one scene on Series X?) they are basically identical." Minor framerate differences isn't huge by DF standards when compared to previous comparisons, so yes, not that interesting of a comparison.
"@RazzerRedux You should go back and correct your post then, because you also quote the same sentiment:" That was clearly referencing the resolution. You apparently ignored the Performance section: "-Frame rate dips "harder" on XSX than PS5, sometimes up to 9-11 frames per second." You are being willfully ignorant. Edit: Ah....I see you edited your post. "Minor framerate differences isn't huge by DF standards when compared to previous comparisons, so yes, not that interesting of a comparison." 10fps drops are exactly the types of differences DF typically highlights so you can pretend everything is the "same" but that is factually not true. And it is hilarious that in one post you say I said everything is "identical" and then reference where I show the difference right after.
I actually wonder if it's true in all cases that being on game pass results in more people actually playing a game (as opposed to just testing it out). I know I'm much more likely to ignore the games I got for cheap in steam sales than ones I paid full price for, and it wouldn't surprise me if the same dynamic frequently plays out with people on Gamepass. Like MS is always quick to cite number of total players for games like Sea of Thieves but they never say how many have played the game for less than 2 hours.
It depends for me. I too have a bunch of Steam games I've never even booted up (probably 80% of my total games...), that seems to be a common pitfall of Steam (and Steam sales). On Game Pass, I play a lot of games I wouldn't otherwise buy / risk spending money on. I certainly try to avoid buying annual sports games like MLB, Madden, FIFA etc.
@Orchard Exactly. And my question with those games is whether you actually end up playing them, or do you just boot them up to check them out and then end up moving back to things you either spent money on or would have paid full price for but for Gamepass.
I'm still waiting for xbox to have superior performance than the cheaper PS5.
It's had stronger performance in a lot of recent head to heads.....
O boy, xbox fanboys and their little cute lies. Stronger performances on what games? Avengers? Ok well done, that's kind of/ sort of a next gen game (but not really) that runs better. Mortal shell? The ps5 is patched and running on par. Overall, ps5 stills seems to have the upper hand.
It's like ps5 has 20+ new Gen games better performance as opposed to one or two on the series x and one of them is in photo mode on a specific hallway in one angle, but the rest of the game runs better on ps5.
Another win for the world's most powerful console
the main thing we can take away from this is, the "s" is a mess
I'm not interested in this game, but I love me some DF breakdowns
No surprise, PS5 stronger.
I’m still confused as to how gamepass games get paid. Does MS pay them a certain amount up front or do they get paid by the download or what?
I don't think there is a clear explanation of how they do it but most people claim that MS pays a lump sum to get them onto the platform if they aren't owned by them. Sony with PS+ I believe credits the developer a certain amount for each download or add to library they get from that month on PS+.
I don't know 100%, but I'm assuming a tiered plan based on players who install and play it. More money to MLB/Sony based on how many install and play it. There's also MTX. I honestly think the *only* reason it's on GamePass is that Microsoft made a great deal to stick it to Sony in a vie to advertise Game Pass in the way that hurts Sony the most. The game itself doesn't have P2W microtransactions like other sports games, so it's not likely to bring in a ton of MTX money unless they get a ton of lazy people who just want to get through it fast.
I know a few people that are sport game centric and that was a huge deal. Baseball fans had to buy a PS4. They now have access to that series x version instead of finding a console or playing the PS4 version. That’s not even getting into the benefits of gamepass holders which is another big win for baseball fans in general. Regardless of the intent I’d say it’s a good thing. That Sony team is about to make an ass of money. Everybody is happy.
Very close performance , pretty much on par when playing the game, but when animations/videos show up, the nod goes to Sony. Sometimes up to almost 10 FPS. These systems are closer than ever and pretty the same performance wise. I do wonder if there is a bottle neck on Microsofts side somewhere, since they have the TFLOP advantage. Surprised DF hasn't really looked into it further?
I don't think DF want to look into anything further the point is the 12TF means nothing. People said this from the start but n4g experts claimed ps5 will be anihalated
higher clocks + similar chips = less drastic dips.
It's made by a PS studio wtf you expect?
It's ignorant to take a comparison like this and summerize that one console is more powerful than the other. In this day and age that talking point is moot, and any differences you see are from optimizations or lack there of. Given the same effort from a studio they can get indentical performance between ps5 and xbox. Differences now a days are not due to minor differences in hardware but with the underlying code and how well optimized it is for a platform. What do I know, Ive only worked as a dev for multiple AAA titles to xbox and playstation over the last 10 years.
But but but do the online players record pitching signs in the Show 2021 like the Houson Astros?
There are not going to be big differences. PS5 10.2TF That only counts CU Performance when filled 100% It's easier to Fill 36cu's then 52cu's. Thats why cerny went with 36 at high clocks and a shit hot cooling system. Other money went on a New 12lane ssd and i/o structure with additional custom chips, cache scrubbers etc. New 3d sound system. Completely new Controller with custom features. Microsoft spent most money on a Bigger GPU. Velocity Architecture, quick resume.
N4G is a community of gamers posting and discussing the latest game news. It’s part of NewsBoiler, a network of social news sites covering today’s pop culture.