He is not another random, he is just crazy.I've never heard of this person. Thought they were just another rando. Can you fill me in a little on their history?
I know the game he described and simply can't say anymore then that.
I will admit he told me a couple of things about games I hadn't heard yet and since have indeed had what he said confirmed by someone I without a doubt trust.
I hope we see a couple of new games with this eventual PS5 showcase.
Late night guesses and all.
He is not another random, he is just crazy.
He usually is a leaker of Nintendo stuff, so you got an idea of how good its sources are.
Is not like what he says will be 100% accurate but as leaker is good enough to in the personal I don't follow him because
I don't want he/she ruined me what Nintendo is going to show.
https://www.wired.com/story/nintendo-leaks-crack-down-animal-crossing-e3/
He is a true leaker.
MS HQ after they saw Mark Cerny's presentation.
I appreciate the comedy in your comment but in reality those are poor Chinese workers who get a few dollars worth for their shift. My Pro also had that "toothpaste" badly applied but a very good one applied correctly afterwards didn't help much with the jet engine. We need a generally better cooling solution this time, or maybe Smartshift will do it.Perhaps Sony's cooling secret this time around is not having a blind person apply the thermal paste; and for the main constituent of said paste be actual thermal paste and not their current toothpaste/quick-dry cement combo.
Maybe it's because TFs aren't really useful metrics and game designers don't use them on daily basis? Here's Guerilla's GPU performance profiler.No it weren't BGs that did that you're quite wrong. BGs actually hinted that PS5 had lower tflops (but he said that PS5 had more performance at the start of the year), it was Tommy Fisher and Osiris Black that whipped people up into a frenzy with 11-13 tflops for PS5.
It might be the speed at which you need to stream assets to catch up with a typical VR FOW head turning speed at a certain resolution of PSVR2 and it's highest framerate. I think VR specs will be used as top metrics because VR is more demanding.At the same time, 40x faster is significant. Why did Sony feel the need to target 100x faster? Why not 50x? 80x? This is what I hope we get definite answers on soon.
The whole idea of Sony paying too much attention to a competitor which has less than half of their market share is a bit weird to me. I bet their primary concern is to satisfy their shareholders and that's a hefty task knowing that this generation was over 110 million consoles and way over a billion games sold.Might have been that Sony didn't EXACTLY know about Microsoft's SSD architecture and they went all out, balls to the wall? I mean, barring a mid-gen refresh in 3 to 4 years (IF it will happen this time around), what you come up with at the beginning of a new generation will be defining your sales; go half-cocked, you might regret it deeply down the line.
No idea, I am just shooting in the dark here. But as Microsoft went with the TF count, it's not crazy to think that Sony's team decided to put their eggs in a different basket.
I talked to him on several occasions at length and while what he was telling me was spot on about iterations of devkits (design wise) his number was always wrong to what I was being told.
I still to this day believe he really believed his numbers.
I do not believe he intentionally set out to basically ruin any credibility he had.
These are just my feelings though speaking with him directly.
Well ok. I'll try to section long replies next time!!!I agree with most of that but....
![]()
Maybe i should have worded it a little bit better but that's ok. I'll try to section my comments to make them simpler next time.Mate, i know you have something interesting to say but just quickly glancing at this wall of text puts me off majorly and my post is not intended in any way as disrespect.
I appreciate the comedy in your comment but in reality those are poor Chinese workers who get a few dollars worth for their shift. My Pro also had that "toothpaste" badly applied but a very good one applied correctly afterwards didn't help much with the jet engine. We need a generally better cooling solution this time, or maybe Smartshift will do it.
Maybe it's because TFs aren't really useful metrics and game designers don't use them on daily basis? Here's Guerilla's GPU performance profiler.
It might be the speed at which you need to stream assets to catch up with a typical VR FOW head turning speed at a certain resolution of PSVR2 and it's highest framerate. I think VR specs will be used as top metrics because VR is more demanding.
The whole idea of Sony paying too much attention to a competitor which has less than half of their market share is a bit weird to me. I bet their primary concern is to satisfy their shareholders and that's a hefty task knowing that this generation was over 110 million consoles and way over a billion games sold.
A bar mitzvah?is tomorrow finally the day i have been longing for? a sign on the horizon pointing to the promised lands
50%, we have waited enough.is tomorrow finally the day i have been longing for? a sign on the horizon pointing to the promised lands
There's something import that should be understood about XBOX Series X hardware and Microsoft's ingenuity (cough cough) that may or may not have been brought up here before...
MS have documented that the XSX uses 'asymmetric' memory setup of 10GB at 560GB/s and 6GB at 336GB/s - 'faster 10GB primarily to be used with the graphics system and 6GB to be used for the other computing functions' (taken from wiki)
For those who ain't discovered, this asymmetric setup is the MOTHER of all bottlenecks cos it happens to be the backbone of the whole system! Why you may ask is it such a major problem, it has 560GB/s feeding it's 52 CU's right? Put it simply the XSX will really need 2 dedicated memory bus's to get the dual bandwidths they claim, that would be crazy 560GB/s + 336GB/s would make 896GB/s total! No, what they have is 320 bit bus at 14ghz. That 320bits used for 10GB fast access and then 192bits of the same bus to access the 6GB pool on a separate cycle of the 14ghz. If you ain't wrapped your head around that yet then think scenario CPU needs to access the 6GB and GPU accessing its 10GB at same time, this will happen on alternate cycles of the 14ghz so in effect they would be getting half the bandwidth will be 7ghz x 320bits will be 280GB/s for the GPU 10GB (fast?), an 7ghz x 192bits is 168GB/s for the CPU 6GB. Well that sounds really bad, but of course that's extreme conditions, BUT it's not great either any which way you chop up the bandwidth!!
To elaborate they have 5 x 64bit memory controllers accessing 16GB. Well 16 into 5 doesn't go evenly so how the fook did they make that fit. The GDDR6 memory are 32 bit wide, so they have physical 6 x 2GB and 4 x 1GB chips, 10 chips in total. Memory is written striped across ram chips, doubt MS would have done it differently otherwise it would be a lot slower. Now would assume... Fast 10GB access would be 320bits connected 4 x 1GB + 6 x lower half of the 2GB chips, slow 6GB would be 192bits (out of the 320bits) switching to the 6 x top half of the 2GB chips
Some may argue the cpu would not need to access the ram as much cos its cache will keep it busy for a relatively while. But guess what, Microsoft have reduced the L3 cache substantially on their Zen2 (probably to make room for their precious CU's and/or cost saving), so it will be taking even more trips out to ram than usual. And then take into consideration all the other access that will be going on by the IO sound dma's and such. All this reducing the bandwidth ratio when the full 320bits will able to grab data making it even rarer for the gpu to keep all CU's firing
I had discovered the above info from Duoae at :
Analyse This: The Next Gen Consoles (Part 9) [UPDATED]
He's done a great job in breaking this down! I would guess it has been posted here before somewhere
Not sure how the heck XSX is able to lever any benefits of HSA (heterogeneous system architecture) as the XSX memory pool of the gpu and cpu are segregated. One of the main benefits of using APU architecture is HSA but Microsoft have missed the point of it. Maybe its gpu cpu can access some same memory but the way they're going about it seems to defeat the purpose of hsa
Microsoft touting they've made the 'World's Most Powerful Console' being super efficient and they've eliminated every bottle-neck is an extreme LIE. The XSX is not designed by talented engineers, otherwise they would have not made such stupid decisions and compromises in the architecture. But by incompetent engineers made to order by Phil Spencer to meet their PR narrative they have a 12TF console that is twice as powerful as their last generation, that was their goal no matter what sacrifices they made to get it! It's got 52 CU's do the maths you get 12TF. But it won't deliver as its being discovered right now by developers working on it... if they ain't got anything good to say about it then don't say anything at all!
Any of you wanna call this fud, fine it will be your loss when you buy your XSX and discover it underperform at everything, not because of early releases but its entire life cycle. Don't get angry at me, get angry at Microsoft for selling you a lie
I would happily give my foreskin for a ps5, but i already did for far lessA bar mitzvah?
Truth is, a lot of us did. It's hard to say but the next-gen tech really exposed a lot of our ignorance on gaming architecture and systems, a lot us Sony fans fell for the teraflop power narrative a long time ago thinking it was the only metric for performance, it's why so much of us were adamant on jumping on the 13 TF train (hype train memories lol). I remember so many people on this thread went into a meltdown because the PS5's tf was lower than the Series X. it was so bad I actually took a break from this thread for a few weeks until things cooled down.
Luckily we've had so many respected developers, tech analysts, trusted insiders and journalists come out and tell us that the PS5 is actually a more than capable console and more importantly that it would be a game changer in how games are designed. I remember one game designer even reacting to the UE5 trailer and saying that this sort of capabilities will actually change the workflow of game studios since they won't be wasting time on creating LOD's, maps and solving sufficient data stream problems, all of which are mostly caused by slow hard disk drives. It's actually going to cause small shifts in the gaming industry.
EDIT: I would just like to add, regardless of what spec sheets say, the developor praise for the PS5 is enough for me, I fell in love with Playstation first party games and if the creators of those games are highly praising the PS5 than who am I to complain. I'm sold on it already.
is tomorrow finally the day i have been longing for? a sign on the horizon pointing to the promised lands
I am letting the game do the talking. Look at the Video I posted. That is a game thats coming to ps5 running on better hardware than a ps5
WALL OF TEXT
The whole idea of Sony paying too much attention to a competitor which has less than half of their market share is a bit weird to me. I bet their primary concern is to satisfy their shareholders and that's a hefty task knowing that this generation was over 110 million consoles and way over a billion games sold.
Cerny in road to PS5 said that you need at least 7,5GB/s to reach standart PS5 SSD. and how he said it, body language tells me that number could be even higher. we'll see.You will use 7GB/s SSDs on PS5 after launch and it will probably improve performance over the default SSD that come with PS5.
Sony choice for 12 lanes is due to use cheaper NAND modules to reach 5.5GB/s.
All the key custom hardware IO Sony did is inside the APU with the I/O controller, cache, co-IO controllers and de decompression units.
BTW all new 7GB/s M.2 SSD to be released this year (that will be compatible with PS5) uses NVMe protocol version 1.4.
Here's Guerilla's GPU performance profiler.
Why tomorrow specifically?is tomorrow finally the day i have been longing for? a sign on the horizon pointing to the promised lands
It's a wall of text yeah. And even i admit that I could i have done it better by splitting it into paragraphs (sections). That would've made it better. There's always a next time though!!!
Can't you edit?It's a wall of text yeah. And even i admit that I could i have done it better by splitting it into paragraphs (sections). That would've made it better. There's always a next time though!!!
There's a 5 minute limit after making a post. After that limit passes, you can't edit the post.Can't you edit?
I would happily give my foreskin for a ps5, but i already did for far less![]()
There's a 5 minute limit after making a post. After that limit passes, you can't edit the post.
I did get promoted to member 3 weeks ago. Let me try it and see if it works.Thats for Neo-Members, Members don't have that limit.
There's a 5 minute limit after making a post. After that limit passes, you can't edit the post.
Yep, i just managed to edit the post and make some slight alterations. Thanks!Only for Neo-Members.
Edit: sorry for being late^^
Thats for Neo-Members, Members don't have that limit.
A lot of us expected PS5 at 12-13 teraflops. When it was revealed at 10.28 tf not sustained it was a letdown. You contributed to it.
Nobody cares about GCN conversions. We wanted to know RDNA teraflops.
My guess is you didn't actually know the teraflops number.
You will use 7GB/s SSDs on PS5 after launch and it will probably improve performance over the default SSD that come with PS5.
Sony choice for 12 lanes is due to use cheaper NAND modules to reach 5.5GB/s.
All the key custom hardware IO Sony did is inside the APU with the I/O controller, cache, co-IO controllers and de decompression units.
BTW all new 7GB/s M.2 SSD to be released this year (that will be compatible with PS5) uses NVMe protocol version 1.4.
I would happily give my foreskin for a ps5, but i already did for far less![]()
I thought it was already confirmed.Question, CAN we confirm RDNA2 is a 25% increase in performance per clock/FLOP over RDNA1?
Ah ok, you were summoning him.A Medium or Fortune Tellers script covering all their bases.
I think he knew the teraflops number but the issue was mostly that he applied those GCN conversions only for 1 system. If he would've said that the same calculation can be applied to XSX the story would be different. But now he hinted that PS5 is at 12-13TF (using GCN conversions but not telling this), would have people believe the PS5 GPU is on par or even better than the XSX, while now we know this isn't the case, which caused a letdown.A lot of us expected PS5 at 12-13 teraflops. When it was revealed at 10.28 tf not sustained it was a letdown. You contributed to it.
Nobody cares about GCN conversions. We wanted to know RDNA teraflops.
My guess is you didn't actually know the teraflops number.
You have a PM. I hope that this leaves the issue settled for eternity.
Why tomorrow specifically?
edit: Oh it's a Tuesday all right
Cerny in road to PS5 said that you need at least 7,5GB/s to reach standart PS5 SSD. and how he said it, body language tells me that number could be even higher. we'll see.
I don't think it can be higher, I'm sure Mark said 7GB would be near to saturating the PCI Express link. From a quick google, PCI Express 4.0 has 2GB/sec per lane, so 2*4=8 tops.
If that be the case, it's likely we'll see a higher storage model maybe a year or two down the line because of demandAgree 8 is probably max transfer then decompression, think Cerny is being vague as they need to find if a faster raw speed makes up for missing the 6 priorities.