I think your talking about the higher tflops the PS5s higher clocks produced, e.g 2230mhz gives 10.28tf and 2000mhz gives 9.2tfs, so yes in that sense the ps5s high clocks did provide a boost , but we are talking about cernys and others that have claimed that a gpu with less cus but higher clocks performs better then a gpu with more cus and lower clocks but both gpus have the same tflops, in the PS5s case this has yet to be demonstrated, then again now that I think about it the higher clocks may be giving it a slight boost because the PS5 is performing like a 10tf machine even though the performance is variable due to varible clocks.
2 Likes
yeah i was talking about clocks i think Series S needed it the most for longer lifecycle. i donāt know it was Comparison about more CU vs Less CU yeah more CU from same architecture should win
1 Like
SeriesS will be fine, the switch received ports and the seriesS is a far more powerful machine compared to the ps5/xsx then the switch was to x1/ps4.
To put thinks into perspective the seriesS is like if the switch had 8jag cores, 5gb ram @88gb/s and a 614gflop gpu, if that was the case the switches x1/ps4 ports would of faired much better.
2 Likes
yeah i know this it should hold up for 5 years. i was just stating there was a room for improvements on both Series X and Series S designs before they decide to lock with these clocks and that it
Oh, ok, but I dunno why your quoting me about this.
2 Likes
The clock speeds are irrelevant to this discussion and are already factored in to teraflops calculations. The point is that the PS5 is performing completely in line with a 10.2Tflop machine. Good for it. There is no magic going on though.
The Series X is not performing as a 12Tflop machine which shows something in its dev environment or APIs are not working as expected.
The video answers the question as to whether the PS5 is performing better for some secret sauce reason and the answer is: NO.
Xbox just need to iron out whatever wrinkles there are that is stopping the full performance from their card being achieved.
2 Likes
You have it the wrong way round when it comes to designing a chip.
Microsoft has already said that when they set out designing the Series X they wanted to double the raw teraflops of the One X and that is what Microsoft achieved, they hit their design goals with the Series X and henceforth the Series S because that was designed to scale in-line with the Series X.
2 Likes
i was just station there is a room from improvement and that it
You donāt understand what you are talking about. The clock speeds are part of the teraflop calculation. The Xbox one had higher clock speeds than the PS4 but fewer CUs.
Clock speeds are pretty irrelevant as digital foundry have shown previously.
2 Likes
NO Buddy what i mean RDNA 2 GPUs can handle high clocks speed and MS played it Safe with clocks. it doesnāt matter if the improvements were %5 or %15 on the game. as long itās improvements. MS decided it was not worth it and iām not talking here about variable clocks they can push high clocks like 2000MHz on the X and 1700Mhz on the S
A 5700XT has around just a 10% performance uplift over the 5700, it may go higher on some games but in general a lot closer to 10 than 15-20 range. Though that was for older games. Benchmarks for 5700 became harder to come by after 5700XT launched because everyone defaulted to that one.
TBH the ps5 result in here impressed me. Wasnāt expecting the settings to be that high (and previously Alex mentioned they were using the same settings as X, but that doesnāt seem to be the case), and the performance to be that high comparatively.
Hope they do more comparisons like that. It doesnāt seem itās reaching that level across other games but it didnāt seem it was that high in here at first either.
The only logical explanation that I can think of at least about the Ubisoft games is that the PS5 version was the lead platform aka the games were designed with 36 CUās in mind and due to the more immature dev tools of MS/time constraints devs just ported this version to the Series X and didnāt use the extra CUās to take advantage of the extra horsepower. That might also explain the surprisingly poor state of the One X versions when compared to the PS4 Pro versions as well.
I donāt buy at all that thereās a bottleneck on the SXās APU, the system looks extremely well designed. As you already pointed out from various benchmarks on AMD cards we know that the higher clocks of the PS5ās GPU isnāt magically a secret sauce and having more compute units are clearly an advantage (Teraflops DO matter on the same architecture as much as some want to believe otherwise) so something is fishy here and itās most likely not a hardware problem.
Iād be interested to know how much work is required to build a game around velocity architecture.
definitely extra work they need to learn SFS Efficiently and how to manage data on split-up Ram
I will note one interesting fact here that should be kept in mind with the earliest talk about comparisons. Sony has not jumped on anything publicly talking up some supposed hardware advantage based on any of this, but more importantly where are the developers out there either publicly or even anonymously saying that the PS5 is really punching above its weight and the Series X has some type of real ābottleneckā in the architecture? The answer is that there isnāt that type of chatter because everyone knows that Series X is more powerful and will show itself proper as tools improve and devs get more time with the system on development. Developers had no problem talking about the ESRAM complexity for Xbox One around launch for example.
Every console has a power and heat budget. MS went with a wide GPU with 12 teraflops of compute power, Sony went with a narrower GPU with 10 teraflops or GPU power.
Given the Series X is likely very close to its heat budget and already at the top of the power budget saying they played it safe with the clocks is nonsense. They have a more powerful GPU - the clock speeds are an irrelevance. The Xbox One had faster clock speeds than the PS4ā¦
You design the machine to use up your heat and power budgetā¦There is nothing conservative about the clock speedsā¦and itās nonsense, complete nonsense to suggest you can just increase the speeds. The PS5 has a much larger physical form and its heat management is questionable already so with more CUs and a smaller physical form itās ridiculous to think you can increase clock speeds. Especially when you already have a more powerful GPU.
I think itās very clear that the Series GDK has less than performant APIs. Itās the only explanation possible hereā¦if the issue had anything to do with CU counts or clock speeds youād see that reflected in the latest Digital Foundry performance video comparing PS5 and PC yet you donāt see that - everything scales relative to its teraflop performanceā¦
Do you really think XSX canāt handle 2000Mhz or 1900MHz with itās cooling ? lol no it can but they decided not do it because the fan will get louder. there is no nonsense about this. also this is not about PS5. iām comparing XSX to RDNA 2 GPUs for example RX 6800 XT 72CUs wider and faster 2,015 MHz.
1 Like
Isnāt it strange that PS5 is performing upto 5700xt?
I mean both have same CUs but PS5 have faster clocks plus RDNA2 efficiency and cores.
shouldnāt it perform better then 5700xt?
Good thing then that SFS is part of DX12 Ultimate, which will be used on PCs as well as consoles.