-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
-
We have made minor adjustments to how the search bar works on ResetEra. You can read about the changes here.
PS5 and Xbox Series speculation launch thread |OT9| - For flops sake!
- Thread starter Mecha Meister
- Start date
- OT
- Status
The "big deal" is always going to come from a small subset. It was the same with the info that PS5 was stronger during E3 and it's the same after the GitHub shenanigans. I assume that when people ask "what's the big deal?!" they're talking to the people that it would be a big deal for, no matter how large or small the group may be.
I don't know how involved you were last go around but as someone who was checking forums daily the ratio of "Secret Sauce hidden GPU" to actual people that believed it is so lopsided that it's hard to imagine why it's still brought up and referenced 7 years later, even in this thread. Or how Tidux was brought up a few weeks ago. Or how MisterX is brought up here....ever. Point is, people tend to notice when they hear something crazy no matter how quiet the voice is. Mix that in with good ol' fanboyism and you've got some nice ammo for a console war. It's why my biggest hope for this gen is for the consoles to be as close as possible. Not because I care that XSX is going to be more capable than my PS5, but so the bickering that has to exist can at least be about the games more than the machine.
That's pretty much my understanding on it, as much as I'd like the PS5 to hit the double-digits, if the GitHub info is correct, it just doesn't seem like there is a solid case for Sony skipping CU redundancy.
people might confuse it with the xbox one x and think they copied the design ;)
Sony would have visibility of AMD roadmap way before some of it is public. So they'd have known about rdna 2, RT etc. Is it possible when they were first looking at specs that AMD over promised thermal and power performance for Navi and Sony got painted into a corner by committing to a narrow/fast solution and now the power curve doesn't work so they're needing to downclock/ spend on cooling/spend on low yields?
He can draw the line to what ever he wants if he thinks it has better chance of outing his source.
Or if he knows which company has the harsher NDAs.
You would think they have a idea of the power curve with how early they were testing Ariel .
yes ps5 will probably
-not push 2ghz (heat and reliability issues)
- of course have CU redundance
- not have a different form factor from ps4 (no greater cooling than today)
- have a mainstream price
all this combined will mean 8-9TF with current gpu tech. So it WILL be below XBX pretty much certain imo
If it was because the navi thermals where worse then expected then wouldnt we see a highly clocked part from MS aswell?. I feel like because we aren't that something else is going on instead.
TBF Klee has provided us with the XSX specification for the GPU (slightly above 12tf RDNA) as recently as December 13th, but not for PS5 for some reason. Although I agree I'm not sure why leaking the XSX tf number is fine but leaking the PS4 tf number would pose an issue. I'm sure Klee has his reasons though, doesn't want to get his source in to trouble.
Don't think cell was weaker than the 360 cpu .
It was just more of a pain in the ass to program for .
But it has been a long time and i can be wrong .
Well it seem no one want to be the person to leak hard PS5 specs since no one is doing it.
Cell was much faster then Xenon at FP math but branchy code was slower as the SPUs didn't really handle branching too well and where more suited towards graphics like computations.
Everyone's assumption that the GitHub leaks are the most recent development is the main crux of the discussion. We know these numbers are real. In the grand scheme of the PS5 itself and where it fits is an entirely different discussion and one that I feel won't get answered until Sony unveils the specs or we get a huge leak.
+1 -on top of slowly unpicking why they Went for the solution they did (I'd they have), I'm also looking forward to seeing how devs adapt to RT, vrs` and a fast ssd. Turing RT is too slow even on 2080Ti so brute force RT will almost certainly need to be replaced by custom applications of that RT hardware which will be fascinating to see evolve
This is why I can't reconcile a 36cu high clock chip in my head. The idea that Sony would do that relies on believing that they didn't known the road map and/or had issues when moving to 2020 release. And that Cerny decided that a high clock 36cu chip was the most effective way to get to their desired specs.
I can't believe any of that really.
i think the form factor of the ps5 will reveal a lot.. if it is a pc like box (like xbx) it can push 12TF... if it has a standard flat design it just wont exceed 9TF (laws of physics, heat etc)
I dont think they will stay the same, the high clocks and high core counts for both graphics cards and the CPUs are likely to require major changes to cooling on both sides.
For example from what i can tell the 5700 (non xt) draws roughly 165 watts alone, that's the same as the PS4 PRO on the GPU alone and only at 1.725Ghz not 2.00Ghz.
Last edited:
If Sony only have one sku then being more powerful (and likely more expensive) than xsx might not be practical. You'd cede the 'value' market to a possible Lockhart. Still doesn't explain the crazy clocks though
i would be ok with a high price and performance but for the overall market i would guess a lower powered 399/499 would do better than a higher powered 499/599
i would be ok with a high price and performance but for the overall market i would guess a lower powered 399/499 would do better than a higher powered 499/599
Could someone confirm the veracity of the info provided at below link:
www.techpowerup.com
Has this been posted before?
AMD Playstation 5 GPU Specs
AMD Oberon, 2233 MHz, 2304 Cores, 144 TMUs, 64 ROPs, 16384 MB GDDR6, 1750 MHz, 256 bit
srry for the late quote, but do you know where this was orginally posted? sounds like a dope response
Has been posted. It's considered as placeholder data.Could someone confirm the veracity of the info provided at below link:
Has this been posted before?AMD Playstation 5 GPU Specs
AMD Oberon, 2233 MHz, 2304 Cores, 144 TMUs, 64 ROPs, 16384 MB GDDR6, 1750 MHz, 256 bit
Matt was upset. Don't quite remember the context.
He only provide that information after Microsoft annoaunced the console, before it was 10+
Sony will be fine at $499 at least for the first year .
As for the value market i not certain how much there is in the early days .
If we look at the PS3\360 gen i don't remember the 360 arcade edition selling much more then the normal one early on .
its placeholder.Could someone confirm the veracity of the info provided at below link:
Has this been posted before?AMD Playstation 5 GPU Specs
AMD Oberon, 2233 MHz, 2304 Cores, 144 TMUs, 64 ROPs, 16384 MB GDDR6, 1750 MHz, 256 bit
Not at all likely. The presence of Cerny precludes such terrible decisions.
My impression is that V looking dev kit really runs at 2GHz but retail PS5 will have lower clock, in 1.8-1.9GHz range so overall a bit below 9TF GPU.
Do you guys think PS5 will be the first Sony system to launch simultaneously worldwide? Or will a region be pushed back to early the following year again?
For comparison per wikipedia:
PS1 12/94 Japan->9/95 NA/EU (9 months)
PS2 3/00 Japan->10/00(NA) and 11/00(EU) (7/8 months)
PS3 11/06 Japan/NA->3/07 PAL (4 months)
PS4 11/13 NA/PAL->2/14 Japan (3 months)
For comparison per wikipedia:
PS1 12/94 Japan->9/95 NA/EU (9 months)
PS2 3/00 Japan->10/00(NA) and 11/00(EU) (7/8 months)
PS3 11/06 Japan/NA->3/07 PAL (4 months)
PS4 11/13 NA/PAL->2/14 Japan (3 months)
If anything it's the Series X that I expect will really push the clock on the GPU/CPU, that combined with some great cooling just like the One X.
And stop fooling yourselves talking about dual-GPU's please.
I think the highly optimistic thoughts (yet somewhat realistic) about PS5 at this point is:
- The leaked tests are real but doesn't show the final architecture in terms of number of compute units etc. and being way overclocked.
- lower clocked variant of the series X configuration (same amount of CU's or near).
- 16GB GDDR6.
- Notably faster SSD implementation with end-to-end customization (hardware and software).
And stop fooling yourselves talking about dual-GPU's please.
I think the highly optimistic thoughts (yet somewhat realistic) about PS5 at this point is:
- The leaked tests are real but doesn't show the final architecture in terms of number of compute units etc. and being way overclocked.
- lower clocked variant of the series X configuration (same amount of CU's or near).
- 16GB GDDR6.
- Notably faster SSD implementation with end-to-end customization (hardware and software).
I think it'll be properly worldwide. Maybe that could be one of the reasons behind Jim Ryan's fast transition statement?
Even at 1.8Ghz your still looking at probably like ~230-250watts of power draw by the SoC.
Interestingly, the Xbox article has been updated and now states 96 ROPS.
And PS5 has changed from 8GB to 16GB RAM.
Why would we assume that 9.2TF was the PS5's performance target?
Did Sony know AMD's roadmap when they chose to have a 1.84TF GPU in the PS4 while AMD had 5.6TF GPU in the market before PS4 came out? Why would you assume Sony had chosen their parts out of ignorance and not out of intent? PS4 was designed to be as powerful as possible while priced at 399$ and sold for break-even day-one (which didn't happen because they've upgraded to 8GB). PS4 was not designed around the best GPU offering AMD had in 2013, not even close to that.
I'm going to do this every time this graph is posted, quote why it's irrelevant to consoles in every way :)You mean the graph that had the X/Y axis as GHz and CUs? Let me see if I can find it.
modiz This one?
Can anyone let me know a broad countdown? I found a site that said 12pm AEST Monday the 6th. But that doesn't seem right, as that's less than 24 hours away. I assume it is actually 12pm AEST Tuesday 7th. That would line up with 5pm PT on Monday 6th.
But from everything i can tell a wider lower clocked chip would have better yields, and a cheaper cooling solution making the console cheaper overall.
PS US,EU and JP already all announced holiday 2020 launch of the PS5 on twitter.
That's 100% true, wider and slower will probably be cooler and less expensive. Thing is, who to say PS5 was targeting 9.2TF when it was designed? Around the end of 2018, the silicon was locked, the CU count was locked. Whatever CU count they had in late 2018 / early 2019 is probably the final CU count the PS5 will have. Clocks? They change according to yields, how efficient your cooling solution is and how much you are willing to spend in order to push your console's power a bit further in the endgame.
I don't think it's much different than the memory case in the PS4. Sony went for a clamshell so they can achieve 8GB, they wouldn't have if that was their intent from the get-go because clamshell has its disadvantages and it was a very odd design choice. But they reacted to MS and to developer feedback, so they had to think on their feet, they had no choice. I'm 100% sure Cerny wouldn't have designed a PS5 which has 36CU running at 2Ghz from day one. If the PS5 is 36CU @2Ghz, and that's a big if, it probably happened because of reasons out of Sony's hands.
Oh, and I forgot to tell you that the TD was a really great find, good job!
It's one thing to use this chart to show how clocks affect power consumption (which is great), it's another thing to use it in order to claim RDNA clock sweat spots or console APU TDP.
Because give the options of getting to 9.2tf, why on earth would they pick a 36cu high clock scenario? Given the general assumption on power,/heat, it doesn't make sense.
I think his theory is, and frankly it sounds very reasonable to me, that Sony from the start wanted ~ 8TF PS5 and when they found out Anaconda will be ~12TF Sony decided to raise the clocks (cause they didn't have any other option after 36CU design was locked) in order to close that gap a bit, right DrKeo ? :)
i think regardless of the system parity its going to be a much more respectful tone this gen; both companies are doing as good a job at appeasing their fans as they ever have. the problem last gen was a feeling that the entire content focus of one side was going downhill, but here everyone's playing the hits. version differences will once again be minor curiosities for most people unless they're gamebreakingly severe.
They freestyling it [or getting inspiration from this thread...]
Hey guys, saw there was a new patent by Mak Cerny published 2nd January 2020.
That makes no sense at all. You get slight increase in performance for potential $100 increase in price, ultimately failing to come anywhere close to the competitor in power and match them in price.
- Status