Well you know you’re not going to get that, so why assume either way. If you’re going to require profiling results to determine if a game is CPU bound, it makes sense to require those same results to determine if a game is not CPU bound, right?
Just like any other situation where we assess game technology, we can only make educated guesses with the knowledge we have on hand. We know the jag CPUs sucked and we know only three open world games ran at 60fps on last gen consoles, with two of them being racing games (Forza Horizon 4 and Burnout Paradise remaster) while the third was a cross gen game (MGS5). Please correct me if I’m wrong on this. We also know that a lot of NPCs puts a major burden on CPUs due to the number draw calls, AI, physics, etc that needs to be run on the CPU. Last we know that there would be no good reason to withhold a 60fps mode if the developers could lower the resolution, so that the load is lower on the GPU, and offer such a performance mode.
Based on what we do know above, it’s far safer to assume the game was CPU bound on last gen systems.
Edit: besides, as I also said above, we see a similar modest improvement to load times on Days Gone, even though it’s an enhanced BC title.
Series X is a technological marvel in terms of compute power, I/O, thermal efficiency, passive cooling - everything is just aimed for a balanced and sustained performance profile.
Speaking of performance, it’s quite bad in Borderlands 3 now. Today a new update release and i can’t help but wonder how did this pass certification.
Every ten or so seconds, sometimes more, the game just freezes for like a second, it’s like a hiccup ,like something is loading but nothing is loading. This happens just out of the blue and often. Used to be smooth with few framerate dips here and there. Made a thread about it on Reddit and the Gearbox forum and several XSX and S players have the exact same issue.
Hope DF takes notice and they can let Gearbox know, like with Titanfall 2 and One X back then.
Fair enough. Though something in the lines of “Ps5 ssd is a 15x increase from Pro HDD” was reported by those that attended the demo. Hard to tell exactly what Sony meant with this figure, but it would be strange of them to gave a performance multiplier focusing only on the devkit and not actual specs, even more so when the information that it was slower than the final one only came almost an year later.
(But maybe I’m just reading too much on Ms already have working demos of insane streaming scenarios with SFS, while all Sony had was Cerny talking about how their raw ssd speeds could potentially make the scenarios Ms was already showing a reality)
For all these results you don’t need games to be CPU bound And unlocking framerate in BC to 60fps mode doesn’t contradict different BC modes and frequencies between PS5 and Xbox.
A simple scenario: PS5 CPU in BC mode runs with 2Ghz, Xbox with full 3,8 Ghz. Could explain all results we see: Xbox decompresses faster and loads faster, PS5 BC games run in 60fps in CPU bound games.
What I don’t understand is why automatically assume this about the PS5 BC mode?
Also we know devs have the option to run components on PS4 clocks. It’s safe to assume there is a similar option to run components on PS4 Pro clocks. So why would they bother adding options to lock clocks lower if it didn’t run at full clocks by default?
Fact is: Xbox games load usually faster in BC mode.
And in BC mode game use normal storage API like OpenFile and decompress data with zlib on the CPU. On Xbox Series the CPU in BC mode runs at full speed. PS5 has different BC modes with probably less speed (more like PS4 or PS4 Pro, see Cernys presentation). So loading and decompressing takes a lot more time than on Xbox. This explains our observations with different games in BC mode. No API magic needed.
Yeah and Cerny’s BC slide also had a PS5 mode. Many, including me, assumed that meant the mode the system was on based on the PS5, PS4 Pro, or PS4 game running on the system. However now it seems like these are performance/clock options when running games in BC. So your theory doesn’t make sense either.
Don’t thin k it’s just the CPU delta, Cerny explicitly said they needed to limit the SSD speeds in BC games.
I don’t know why exactly it’s like that, but a lot of Sony BC relies on things the exact same time as it used to. They even added timers on the soc to make sure every instruction call on the cpu/gpu takes the exact amount of time as it did on Ps4/Pro if the game running is expecting this to be the case.
It’s worth mentioning that Xbone do have hadrware compression blocks too. Could be the reason as well with SX hardware blocks being able to chew more data
My assumption from that image was that PS4 games that never got a Pro update ran at normal PS4 settings on PS5, that games that did get a Pro update run at Pro’s settings on PS5, and that games where devs go in and make changes can run natively on PS5 using whatever compute/cycles the devs wanna tap into.
Hardware emulation vs software virtualisation the latter of which Microsoft (Windows & Azure) excels in and they make billions selling that enterprise tech.
PS4 Games made after July 2020 have to be certified by the developer to have no issues running under the Unlocked PS5 Clocks BC mode. The exact cutoff date was listed in the Sony presentation. It may not be July, but was months before the PS5 release.
Next-Gen PS6 BC modes will be interesting since the PlayStation still has no software layers to assist so the hardware will have to do all the lifting again.
I’d hope it would be as easy for them to just flip a flag if it works without issue as Native BC and then it would run at those speeds. It would be disappointing if it required a recompile with the later PS SDK to get the highest clocks.
I don’t think Sony will allow PS4 releases going forward after that cutoff date if it can’t run at PS5 Native BC clocks.
And unlocking framerate in BC to 60fps mode doesn’t contradict different BC modes and frequencies between PS5 and Xbox.
