That's the TFs difference, but without knowing what are the gains for higher clock on RDNA2, for example, we can't assume how much will compensate for the 40% difference in CUs, Idem for smartshift.
That makes no sense at all IMHO. If it was fill-rate, then I would agree on the limited info, but Flops have no bearing on why every AMD card gets owned by Nvidia in the same bracket. With limited info of only knowing two things are in the same tier, fill-rate is usually the decider, and with a raised clock, PS5 is much closer than 20% to the XsX on that metric unless the picture of the XsX chip isn't representative.
Some days back people were speculating on the ROPs for the XsX, but I don’t recall anyone comparing the small 36CUs of the Pro and the – presumably to scale – much larger PS5 36CUs from the slide in Cerny’s talk, so we can make speculation about the ROPs count for the PS5. According to an old Pro vs One X DF article the Pro has 64 ROPs (One X 32 ROPs), and a clock of 911Mhz. But the article states the Pro ROPs just couldn’t be feed by memory bandwidth, so maybe a comparison of 911Mhz to 2230Mhz clocks and a slide that makes the illustration of the PS5’s 36CUs look 2-2.5 times bigger in area than the Pro 36CUs – with no way of knowing if scale difference normalizes for 16nm to 7nm lithographies.
An important point of contention is that while lower budget (AA) nintendo games use simplistic visuals and scope they deliver on all the other aspects making a quality end product
Ok 9I stumbled up to this interesting picture, 46k people have answered this question on some french youtube channel. ( no idea which channel is it or is it biased to one or another, but numbers sound realistic and comparable of stuff I have seen about sony vs ms before)
I guess that Xsex powah aint sexy enough to appeal Europeans. Saw similar share of votes at Finnish gaming magazine query too.
On US this would be closer "battle" I think. Just to show how dominant PlayStation can be on many places, and being faster wont cut it.
I saw some people comparing PS5 ssd to xbox one esram.
Like wtf, they are completely different cases. Funny how far some people are fetching to prove that anything PS5 offers is poop.
If fast SSD on PS5 would be a weird/bad thing like esram, then logically xsex ssd would be 2x worse as it is slower.
ESRAM was used because ddr3 were too slow, so kind of "fix a leak with a tape" situation.
If anybody thinks that SSD (of PS5) is some kind of bad engineering or is used to "fix" something that is too slow, they must be joking or nuts.
esram tried to fix slow ram, ssd just makes things fast and removes all/most of limits of hdd. Dunno how some can even twist their mind to compare them
I guess many people still live in psychosis and think that "if ps5 ssd is faster, that can only mean that bad bad ps5 fanboys think it is faster console than xsex, not possible! RRREEEEEEEEEEEE"
This is IGN, more american site I guess
Backup processing unit? Caution, man, you're entering Misterxmedia's hidden power brick GPU territory here...
The Tempest engine is an AMD GPU-based CU with other customization.
So, what about this one :
The GPU has reduced the “time-to-triangle” from PS4’s 1-2 months to less than a month.
Can we expect faster game development or making bigger games for less amount of time?
4k asset creation takes a lot of time from studios, no wonder Sony created a studio just for that in Asia.So, what about this one :
The GPU has reduced the “time-to-triangle” from PS4’s 1-2 months to less than a month.
Can we expect faster game development or making bigger games for less amount of time?
So, what about this one :
The GPU has reduced the “time-to-triangle” from PS4’s 1-2 months to less than a month.
Can we expect faster game development or making bigger games for less amount of time?
The two best this gen in my opinion. I loved them both!
Look I get snide comments and cynical approaches and all that console warring and FUD stuff
.
X1X's 32 ROPS is linked with 2MB render cache and has up to 326 GB/s memory bandwidth.
And now to respond to your second sentence which is also patently wrong, 5.5GB/s raw data is fast enough to fill streaming asset portion of a game that is generally taking up 2-2.5GB; which means it will fill in half a second, and that is within parameter of what Cerny said in his presentation -> 'a second of streaming cushion instead of 30 seconds' meaning that the GDDR6 on the system will be used much much more efficiently.
1. 2.5 GB of XSX's 6GB is allocated towards low BW intensity OS.
Maybe for a year if it worry's some people that don't understand scaling of a game but the PS5 will have to share graphics with 9.2 TF for 5- 7 years.
holy shit u came wtf what happened
That article was incorrect and really should no be used as an source for anything about Rops and Bandwidth on PS5.
Backup processing unit? Caution, man, you're entering Misterxmedia's hidden power brick GPU territory here...
Interesting.
That's any subscription service in a nutshell. But I agree that eventually, quality will suffer.
Because the demo was meant to show the difference in data stream speed, not rendering capabilities. Spiderman is 100% of the time on screen and loaded in either case, so his presence or absence wouldn't mean anything anyway for the purpose of the demo.
