martino
Member
watch the video....the solution is in it.
the problem is known for long....so make stick easily replacable ....it doesn't seems an impossible task if the focus is put on it during conception.
Last edited:
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
- Status
the problem is known for long....so make stick easily replacable ....it doesn't seems an impossible task if the focus is put on it during conception.
FrankWza
Member
Oh they changed that. QUICK
Riky
$MSFT
> Full RDNA2 on console is only available on Xbox Series X|S.
And yet loses the performance crown in game after game. I guess full RDNA2 features set, which I think is a questionable claim, is a bit of Pyrrhic Victory.
They already had the claim of world's most powerful console. Now we moved on, as if that was not good enough, and now full RDNA2 feature set.
Please MS take your victory lap with RDNA2 held high over your head, while Sony enjoys the spoils of war. Lots of wine, women and song.
Most of the launch games have been patched now and run just as well or better on Series X, then we have the biggest third party release post launch in Hitman 3 where Series X smokes PS5.
It was a marathon not a sprint but Series X took the lead after about 200m.
TheThreadsThatBindUs
Member
I'm not sticking up for any company. I'm just being realistic.
The very same video linked above clearly spells out that the potentiometers aren't really the issue, it's the fact that they aren't made to be easily replaceable (you basically need to re-solder new ones onto your controller, and even then you may need advanced calibration software to make the replacement part work).
So the potentiometer isn't the issue. It's the fact that as a part with a relatively short lifespan, it's too deeply integrated into the controller design, instead of the controller design treating it as a consumable part that should be able to be easily replaced.
I can sympathise with the frustration around controller prices going up when obvious flaws in the controller assembly design mean they're prone to failure well within the warranty period.
I watched the video and recognise the suggestion iFixit makes. It's a reasonable solution and one the console vendors should adopt.
I'm not sure why you seem to think anyone was arguing otherwise?
DJ12
Member
martino
Member
you seems to be defending console maker like there is no real alternative with post like this one
Last edited:
Nikana
Go Go Neo Rangers!
Perhaps you dont see it but your words very strongly suggest that the controller isn't the issue but that consumers have unrealistic expectations because a better product doesnt exist.
The potentiometer is indeed the issue. That would imply it the longevity of it cant be improved. Things can always improve.
Yeah, looks like the FUD is back for PS5. It’s honestly pretty sad.
I mean it’s the PS5 that’s been beating or equaling the SX in multi platform games and outselling both models. I don’t understand the endgame for Xbox fanboys when reality paints a vastly different picture and always will. They’re trying to climb with a torn rope. Get out of that 2013 payback mindset already. It isn’t working and you’re going to continuously embarrass yourselves because nothing is going to change in this console war the way you think it is.
Keep fighting the good fight and all that though.
I mean it’s the PS5 that’s been beating or equaling the SX in multi platform games and outselling both models. I don’t understand the endgame for Xbox fanboys when reality paints a vastly different picture and always will. They’re trying to climb with a torn rope. Get out of that 2013 payback mindset already. It isn’t working and you’re going to continuously embarrass yourselves because nothing is going to change in this console war the way you think it is.
Keep fighting the good fight and all that though.
Krisprolls
Banned
Only in your alternate reality. In the real world, PS5 runs better in nearly every game until now. Even Cyberpunk now runs a lot better on PS5. Valhalla, Dirt 5, Black Ops... Every game outside Hitman 3 actually.
And that's before 1st party games start smoking everything on Xbox by a wide margin as usual. This will be another long generation for what's left of green team fanboys.
Last edited:
DJ12
Member
To be fair, you forgot photomode on Control also, but as the actual game runs better on PS5 does it really count, I think Riky will say yes to that one.
Are you saying you find this thread is Ech or something?
sircaw
Banned
And if you go into the main Xbox channel, it's alive with people crying about how this website is a Sony outpost and that their members are in constant fear of being banned or have been chased away.
sigh
fish wants an easy life, he does not want to come out again to fight the likes of sleep doctor and eastwood again.
Sinthor
Gold Member
Yeah, me too. Don't know if it's the same things that are disappointing you, but for me....the game is so technically good (almost PERFECT) that it's just a shame to me that I just can't make myself play thru it more and enjoy it. They just killed it for me. Which to be fair...they said their goal was not to have people enjoy the game and to be 'uncomfortable.' Mission accomplished! Really irritated that I bought it digitally so I can't even trade it in to get a little credit towards something I might enjoy more. I should have gotten Demon's Souls as digital, for example. Oh well. Live and learn!
SportsFan581
Member
You must be very selective with the modes you use to compare with or the settings that things are running at, but if it makes your day, that's great. Enjoy your PS5.
Sinthor
Gold Member
I don't get all the furor about this. From some people's arguments, a company can't use ANY piece or design of an earlier product if they're making a new revision and want to rename it. Like RDNA 2.0....can't use ANYTHING from 1.0 or else it isn't REALLY RDNA 2.0. Just silly. That's like saying an 8 cylinder car engine can't be called a 'V8' because after all it DOES have 4 cylinders....so REALLY it's just a double-stacked 4-cylinder! Call it 4x2 cylinder engine instead! Just....wow.....such a stretch in arguments! LMAO!
Oh the gameplay is sublime. I'm so scared they're gonna lose this gameplay engine by throwing it out with the bathwater. Or it gets tweaked into something different feeling, like Uncharted gameplay.
And the control customization/accessibility settings, are literally progressive. I was literally able to re-enable a button combination to allow punching whilst in aim. That's my goto panic move in TLOU1 mp.
TheThreadsThatBindUs
Member
Nah, you just need to learn how not to think in such an absurd binary fashion.
My post was trying to ascertain from the other poster why he seemed to think the potentiometer was shitty, when the iFixit video itself doesn't suggest so.
That you read my posts as a defense of anything says more about your own myopia.
Where in my post does it suggest this? Was it this part?
Because to me this suggests the opposite.
Well, we've clearly established you're not the expert in the matter, and the folks that are, i.e. iFixit, proffered a solution to the controller drift issue that doesn't at all include "use a better potentiometer".
Of course there are probably joystick modules with more robust potentiometer assemblies.... they're probably just more expensive, and console makers very likely aren't willing to spend the extra on them. Hence, iFixit's more realistic solution is to make the joystick module a consumable portion of the controller design.
It's all well and good to say, "just make something better", but the very fact that all major console controller manufacturers are all using the same tech based on the same design points to the very real fact that the existing potentiometer design (albeit flawed) is the best available tech for the price these companies are willing to pay.
This isn't a defense of the design, it's merely a commentary on the very real observed facts of the situation. ALPS is the company who makes this component and they supply it to everyone. If they have no interest in improving the product because they have no competition, what do you want the console makers to do? "Just make a better potentiometer" clearly isn't an option for the console companies.
It's called being pragmatic, rather than idealistic which is what I'd argue you're being.
Last edited:
BattleScar
Member
All the people who are discussing which console is and isn't "full" RDNA2, can you please explain to me exactly what RDNA2 actually is?
martino
Member
okay....then i will let you keep fighting the pragmatic fight with insult and certainties
DeepEnigma
Gold Member
Whatever the clever marketing word it as. We play specs & marketing buzzword trademarks around these parts, not games.
Last edited:
TheThreadsThatBindUs
Member
Insults?
If you're gonna post on a web forum like this, then you should really learn to have a thicker skin.
Nothing I said was intended to be insulting. If you've taken it that way, you're probably being a tad too oversensitive.
Negotiator
Member
Something that the average Joe doesn't give a toss about it.
You're supposed to buy PS5 for exclusives and/or XBOX Series X/S for Game Pass.
PaintTinJr
Member
Sorry for replying to you
P
Panajev2001a
in this thread, but figured it's too much speculation from me to be for the PS5 Die Shot thread.
After all the discussion about the FPU and AVX2, I googled about AVX2 use for games, and quickly found myself looking at a 2011 paper on Ray packet tracing acceleration with SIMD AVX(2?). The packet in this situation being the 256bit vector unit to packet up 4 coherent rays - rays starting at similar place and on a similar/same vector - to process in a single instruction for intersection test for triangles - sounding a lot like the 4 triangles per pixel reasoning in UE5 nanite .
There was also explanation/analysis of RAM being the true bottleneck for CPU RT, not processing speed, because they said the critical element is the cost of work lost from not indexing the correct geometry in time. The critique of the RAM was largely made through algorithm performance for misses, when compared to optimised use of L1 cache in different ways, including benefits of prefetching L1 cache lines, and the bigger performance from data being correctly aligned in the L1 cache - even at the expense of having less data in the cache line. The importance of the indexing the geometry database in terms of latency and speed read like the problem the IO complex solves in the PS5.
The more recent research paper (2016?) I skimmed was about 5D ray packet tree hierarchy technique for packet-ing up random (Secondary bounce rays) and efficiently processing them in AVX2/GPU as best can be done IIRC - because AVX2 width utilisation still remained too low for grouped random rays. Further in the past (2001) the ray packets were being used in RT research with Intel SSE SIMD, so it seems there's been nearly two decades of research in using SIMD for high quality real-time RT on a CPU and the resolutions, triangle counts, RT and frame-rates look like they converge to the RT we see in spider-man morales 30fps or UE5 demo.
So from that, I actually think Cerny's posed question on 256bit was alluding to the PS5 CPU being setup for heavily using 256bit instructions, and Spider-man RT (and UE5 nanite) is already using it with coherent ray packet tracing on Zen2 CPU cores(along with GPU BVH ), and either it is new 256bit fully functional FPU or RF blocks - that doesn't fit with existing Zen2/Zen3 blocks shown - or that the pre-existing block/s have been streamlined around ray packet tracing needs, and to reduce power consumption while being heavily used, or the 256bit stuff is being done elsewhere (CUs, GE, Tempest, an unrevealed block)
Cycling back to the 256bit instruction usage question, Cerny posed, maybe it is that way because of power draw in the XsX GPU, as using one or the other, should mean half the heat/power generated over the same area.
After all the discussion about the FPU and AVX2, I googled about AVX2 use for games, and quickly found myself looking at a 2011 paper on Ray packet tracing acceleration with SIMD AVX(2?). The packet in this situation being the 256bit vector unit to packet up 4 coherent rays - rays starting at similar place and on a similar/same vector - to process in a single instruction for intersection test for triangles - sounding a lot like the 4 triangles per pixel reasoning in UE5 nanite .
There was also explanation/analysis of RAM being the true bottleneck for CPU RT, not processing speed, because they said the critical element is the cost of work lost from not indexing the correct geometry in time. The critique of the RAM was largely made through algorithm performance for misses, when compared to optimised use of L1 cache in different ways, including benefits of prefetching L1 cache lines, and the bigger performance from data being correctly aligned in the L1 cache - even at the expense of having less data in the cache line. The importance of the indexing the geometry database in terms of latency and speed read like the problem the IO complex solves in the PS5.
The more recent research paper (2016?) I skimmed was about 5D ray packet tree hierarchy technique for packet-ing up random (Secondary bounce rays) and efficiently processing them in AVX2/GPU as best can be done IIRC - because AVX2 width utilisation still remained too low for grouped random rays. Further in the past (2001) the ray packets were being used in RT research with Intel SSE SIMD, so it seems there's been nearly two decades of research in using SIMD for high quality real-time RT on a CPU and the resolutions, triangle counts, RT and frame-rates look like they converge to the RT we see in spider-man morales 30fps or UE5 demo.
So from that, I actually think Cerny's posed question on 256bit was alluding to the PS5 CPU being setup for heavily using 256bit instructions, and Spider-man RT (and UE5 nanite) is already using it with coherent ray packet tracing on Zen2 CPU cores(along with GPU BVH ), and either it is new 256bit fully functional FPU or RF blocks - that doesn't fit with existing Zen2/Zen3 blocks shown - or that the pre-existing block/s have been streamlined around ray packet tracing needs, and to reduce power consumption while being heavily used, or the 256bit stuff is being done elsewhere (CUs, GE, Tempest, an unrevealed block)
I ran some guesstimate numbers before speculation the above - but probably messed them up somewhere.
I was thinking a Zen2 SMT core - even if downclocking to 3GHz and being 40% efficient(SMT should be more like 60% IIRC) - if it was doing 3/4 filled L1 cache packet intersections on average, and did 10 intersections for every valid intersections - even though coherent rays should be far more successful than 1 in 10 - then if we say AVX2 is a single clock cycle then how many a second - and frame can a core do.
3x10^9 (GHz) X 0.4(40% efficient) X 0.75(3/4 filled cache = 3 or 4 triangles when AVX2 packs 4 triangle) X 0.1 (1 in 10 intersections is a hit) x 1 ( AVX2 is 1 cycle per clock) = 90M intersections per Core.
Then if we say UE5 demo showed 2560x1440 and did 4 triangles (hits) per pixel, per frame = 14,745,600 ~= 15M
Then if we say 30fps = 442,368,000 ~= 450M
Then if we say 60fps = 884,736,000 ~= 900M
For 30fps 5 Core would be needed - leaving 3 Cores over for game logic, physics, etc, which seems to fit for UE5 Demo.
For 60fps, 10 cores would be needed, which is more than the 8 available in the PS5, and would have still left nothing for running the game engine.
I was thinking a Zen2 SMT core - even if downclocking to 3GHz and being 40% efficient(SMT should be more like 60% IIRC) - if it was doing 3/4 filled L1 cache packet intersections on average, and did 10 intersections for every valid intersections - even though coherent rays should be far more successful than 1 in 10 - then if we say AVX2 is a single clock cycle then how many a second - and frame can a core do.
3x10^9 (GHz) X 0.4(40% efficient) X 0.75(3/4 filled cache = 3 or 4 triangles when AVX2 packs 4 triangle) X 0.1 (1 in 10 intersections is a hit) x 1 ( AVX2 is 1 cycle per clock) = 90M intersections per Core.
Then if we say UE5 demo showed 2560x1440 and did 4 triangles (hits) per pixel, per frame = 14,745,600 ~= 15M
Then if we say 30fps = 442,368,000 ~= 450M
Then if we say 60fps = 884,736,000 ~= 900M
For 30fps 5 Core would be needed - leaving 3 Cores over for game logic, physics, etc, which seems to fit for UE5 Demo.
For 60fps, 10 cores would be needed, which is more than the 8 available in the PS5, and would have still left nothing for running the game engine.
Nikana
Go Go Neo Rangers!
You entire tone and post suggests that. Simply because I don't have the engineering background to "come up with something better" doesnt suddenly mean I am not qualified to ask for something better and doesn't automatically mean I am being idealistic.
Trying to suggest that none of these hardware vendors dont have the capabilities to create a better product if asked is pretty rich. Can they snap their fingers and have it done? Not at all. It would take time and R&D most likely, but it can most certainly be done. There is nothing to suggest it cant be other than the lack of ambition and/or or lack of necessity as people just keep buying controllers even as they continually go up in price for no real reason.
Microsoft and Sony both have products that were made from the lack of existing materials/components in the tech sector of a product they deemed fit for their vision. No reason to think that cant be applied to a controller. You do not have to use existing tech simply because its the only thing available.
Asking a company to fix a fault in a design thats been there for two generations of machines, now going on three, is not being idealistic. We can create better and better chips for these machines with complex cooling systems while partnering with these companies to create better tech for the future but improving a 50 cent piece is suddenly idealistic? Beg to differ.
THEAP99
Banned
Desperation
martino
Member
Sure thick skin is what your to prove your point....are you sure analog module were not better or that better doesn't exist ?
Look at x360, dual shock 2 (or 3) compare them to the new thing.
Also look at thrust-master did with eswap analog module (both the module and how it is delevered in the pad) ? cannot console maker selling 100M of controllers do the same at least for the module ?
Now with recent consoles drift happen so soon now we have class action in the 6 first month !
Did you had all that mind or was your pragmatism ?
Last edited:
Ass of Can Whooping
Gold Member
9tflop RDNA 1 GPU beating a 12tflop RDNA 2 GPU
Cool right
Cool right
TheThreadsThatBindUs
Member
Read the words of the post, not what you project onto it.
No. But when you ignore all the important context, i.e. APLS who produces the potentiometers is the sole vendor for all major console manufacturers, they have no competition and thus no incentive to improve their product, and just jump to "Sony/MS/Nintendo should just make a better potentiometer", you're being idealistic.
The console makers didn't make the existing joystick module. APLS did. So APLS has the responsibility of making a more robust unit. But guess what, there's little incentive for them to do so. Not just because they supply all major consumers of the part, but also because it's an extremely low tech, low value part (less than a $1 in BOM), so the impetus to spend in R&D to improve the longevity of the part simply isn't there (because the market pressure to do so isn't).
So Sony/MS/Nintendo, are better off treating the module as a consumable component, just as iFixit suggests. They clearly don't have much choice on 3D Joystick potentiometer modules, otherwise all three companies would do what most large companies do and source components from multiple vendors to create competition and ensure you're always getting the best deal.
Well sure. It is. But I never once did this. You keep projecting this shitty strawman onto my posts.
If APLS could have made a better product they would have. If someone else could, they would have and would have stolen away the supply contract for the parts from APLS. The fact this hasn't happened says a lot.
Maybe you need to simply stick to the facts rather than pie in the sky idealism hoping that one day in the indeterminate future a better potentiometer would be emerge.
Well, there you go. Now you're starting to be more pragmatic. Now you're starting to see why "just make a better potentiometer" isn't a likely outcome to fix this controller drift problem. The reasons APLS hasn't already rectified this are pretty self-evident. Could they? Sure. Will they? Probably not.
I'm struggling to parse these sentences. What are you trying to say here? Surely you can't be suggesting they use tech that isn't available?... Surely you don't see the oxymoron in that?
That's not why I called you idealistic. Keep up. You were being idealistic when you reduced the problem down to "just fix the potentiometer". iFixit already proffered a far more workable and realistic solution, and you continue to ignore it.
Stop arguing an untenable position and just acknowledge that the potentiometer low lifespan issue isn't likely to be fixed any time soon, however, there are clearly other ways for the console makers to address the problem. Both iFixit and I agree, that their suggestion is a sound one and Sony/MS/Nintendo should adopt it.
You seem to be the only one arguing something different.
Looool, the irony of this statement is hilarious. The reason you don't see large expensive company R&D departments working on improving a 50 cent part is precisely BECAUSE it's only a 50 cent part. It's a low value, low margin, not worth the time or effort, piece of tech. There's no commercial reason or impetus for APLS to redevelop it unless they lose the supply contract or the console makers sue them.
Last edited:
Nikana
Go Go Neo Rangers!
Your inability to have a conversation without insulting people is hilarious.
Good day.
TheThreadsThatBindUs
Member
What are you even arguing? You're first reply to me was you highlighting the iFixit suggestion, but all of a sudden you're now arguing for a better potentiometer? You're not consistent.
Look at the facts: i.e. a single company, APLS, is supplying this part to all current existing console manufacturers. The same part has been used for at least two generations now, i.e. that's at least the last 10 years.
Arguing about what other controllers over a decade ago may have used is fruitlessly moot. It's over a fucking decade ago. The suppliers of those components could have gone under, discontinued those parts, or any other number of possibilities.
3rd party peripheral device makers may very well be using more robust potentiometers. Or they may not. Original platform controllers from Sony/MS/Nintendo sell 10s to hundreds of millions more units than 3rd party devices, so will get more use... so will get more visibility... thus class action lawsuit.
Fundamentally, if Sony, MS and Nintendo have been being supplied these parts from a single vendor for the last 10+ years, there is a very good reason for it.
We can speculate all we want about there being other better options available, but unless you can point to real hard examples, you're just spouting hot air.
Grow a thicker skin. There was nothing insulting about my response to you. I've made no personal attacks. Just criticised the points you're arguing. If you can't handle that, then maybe a discussion forum isn't really the place for you.
Last edited:
Nikana
Go Go Neo Rangers!
akaFullmetal
Member
Hey what's going on guys.....
Setsuna Mudou
Member
Things can improve, but they won't so much because people are hellbent on not paying more than $500 for a product that will last them about 7 or 8 years under significant use (regardless of inflation). That's the issue.
So yeah, companies will need to cut all corners they deem acceptable in order to hit that target price.
Basically the consumer gets what they paid for.
As an example I just upgraded my sim racing gear from something that combined cost $500. My new gear costed about $3.5k. That can be the difference between using top of the line components and mass market, entry level stuff, which is what consoles are.
SportsFan581
Member
And if you go into the main Xbox channel, it's alive with people crying about how this website is a Sony outpost and that their members are in constant fear of being banned or have been chased away.
sigh
fish wants an easy life, he does not want to come out again to fight the likes of sleep doctor and eastwood again.
Seriously, it does seems like a lot of people have been banned recently, hopefully the politics section doesn't clean the whole site out.
Shmunter
Member
Last edited:
kyliethicc
Member
That 1 XSX chip could be used in a server to run just 1 XSX game for streaming too, of course. For now, I bet its like a 720p stream of Xbox One games going to phones. Plus they will always need the servers to be able to run Xbox One games for xcloud. They're not gonna remove their Xbox One games from Game Pass anytime soon.
Cause xCloud is mainly for streaming games to phones.
They can't get to their stated goal of 2 BILLION customers with just consoles & PCs lol.
“We believe there is going to be 2 billion gamers in the world, and our goal is to reach every one of them,” explains Choudhry.
/cdn.vox-cdn.com/uploads/chorus_asset/file/10399581/xboxonex.jpg)
Microsoft’s new gaming cloud division readies for a future beyond Xbox
Cloud services seen as the future of games
www.theverge.com
Last edited:
Setsuna Mudou
Member
yes, that's exactly what it is... and it's not even commercially available yet
assurdum
Banned
Sony fans? Lol really? Every thread which talked about AMD are literally invaded by posts about ps5 is no full RDNA2 shit
Last edited:
IllegalLemon
Member
Things you could do.
On the main ps UI:
1.Hover over your game and click options on the controller and select check for updates OR
2. Use the same facility to uninstall your game, then reinstall it and if that still won’t work
Alternatively
You can:
Rebuild the systems’ database (although that is done if the slow start manifests in other games too)
Try any of those and see if it works for your case.
Last edited:
Bo_Hazem
Banned
Last edited:
Bo_Hazem
Banned
All that is good, but when a $320-450 controller like the Elite 2 from xbox is using the exact same cheap-ass components with much higher fail rate that when you have a serious problem. For a $60-70 controller? It's understandable when the vast majority don't face those issues.
Last edited:
Rea
Member
I really wish that Cerny and his team, or Sony's marketing team, should give PS5 gpu a name, instead of calling RDNA2 custom. There are only a few things that are RDNA2, which are Ray tracing and CUs. Many features are Cerny's idea requested AMD for PS5 gpu. In fact, PS5 gpu can't compare with anything else, the gpu is fully customized and Unique. It should have its own unique name. People too hung up on Amd and Xbox marketing buzzwords and keep comparing them with PS5, fanboys spreading FUDs based on another person's guess work. Tired of seeing this again and again.
Nikana
Go Go Neo Rangers!
I think the elite controller proves there is a market for people who will pay more for a quality controller.
But nobody is asking for the controller to be a tank and full of expensive high end materials.
Everything on the controller standard parts, like face buttons and triggers, list millions and millions of presses but the joy sticks last a fraction of the time and are used just as much.
It's inconsistent and lazy. If any of these lawsuits actually amount to significant damage, which I doubt they will, you better believe things will improve.
Last edited:
- Status