Official Thread Pillow Fight that nobody wins with MOAR Jackie Chan and guys comfortable with STRETCHING their sexuality!

Status
Not open for further replies.
Depends. Do you rank up to master through quantity or only when an attempt is at least moderately competent? If see him as a yellow belt tops, certainly not a master.


tenor.gif
lol at gif. You want me to show you tough? I'll show you tough.
 
I think that the hardware discussion is a fun one to have but not the end all be all to game development. For example, Ryse on the Xbox one is still a graphical stunner IMO on relatively weak hardware. The Order running on the base PS4 fall into the same category. Good devs will make use of the tools given no matter what.
 
That’s a bold title.


Xbox Series X will crush the PS5 based on these benchmarks
By Richard Priday

The PS5's performance can't compete with the Xbox Series X, AMD benchmarks show


We heard last week about the PS5's total GPU teraflop performance, and how at 10.3 TFlops, it's behind the Xbox Series X's 12 TFlops. Sony said that it had other ways of offering leading performance, despite the difference in these numbers, but some new benchmarks suggest otherwise.

As pointed out by Notebook Check, benchmarks for AMD's Radeon RX 5000 line of graphics cards can give us an indication of how the two consoles might perform in real life. The RX 5000 cards, also known by their codename "Navi", use the same RDNA 2 architecture as the chips in both the new Xbox and new PlayStation, which is why they're a good comparison.

It all comes down to clock speeds: the speed at which processors can deal with a list of instructions. First off, Sony itself has said that the PS5's performance output scales depending on usage. That means that while the Xbox, as far as we know, will deliver 1.8 GHz consistently, the PS5 may sometimes give less than its maximum 2.23 GHz speed.

Sony's key claim to ignoring the numeical gap between the PS5 and the Xbox Series X is that the PS5 GPU runs at a higher clock speed. Because 2.23 GHz is around 24% faster than 1.8GHz, it follows that the PS5 will perform at a 24% higher rate than expected — right?

Not from what Notebook Check has seen. Notebook Check claims that overclocked RX 5000 cards do not see as big of a jump in performance from the speed as you might expect. As an example, an RX 5700 XT card overclocked by 18% results in only 5 - 7% higher frame rates. The improvement is there, but it isn't linear, as Sony would have us believe.

The PS5 still has some potential advantages in the form of fast 3D audio, SSD storage and ray tracing, but the Xbox Series X has the latter two features as well, and it's debatable whether better visuals or better sound is more of a selling point to gamers. It's far too early to say for sure which console will win this generation's duel for supremacy, but the PS5's numbers aren't looking great so far.
 
16gb is greater than 10gb
5.5 is greater than 2.4
2.26Ghz is faster than 1.825 GHz
PS dev skills > XB dev skills

Not sure assuming the PS5 not using several gigs to run the os is a good idea. I think it's unlikely they are using the SSD to run it. Audio probably still needs ram too.

People pretending an SSD is going to solve all their problems are delusional.

I notice your little list ignores all the ways the SeX out classes the 5. Just sayin.

I've been through too many move engines and cloud sauces to buy this without proof.
 
  • Like
Reactions: Mcmasters
Not sure assuming the PS5 not using several gigs to run the os is a good idea. I think it's unlikely they are using the SSD to run it. Audio probably still needs ram too.

People pretending an SSD is going to solve all their problems are delusional.

I notice your little list ignores all the ways the SeX out classes the 5. Just sayin.

I've been through too many move engines and cloud sauces to buy this without proof.
With It all being available(unified) and running at it's top speed they can put it wherever they want
OS or wherever and it would be the top speed.Let me add...
I wouldn't be surprised if they used the super fast SSD 5.5 raw-8-9 gb/s uncompressed to handle some of the OS.

In overall bandwidth/brute force the XSX has the advantage and I think it should=more higher resolutions i.e later in the gen when/if games can't run true 4K they will be more likely true 4K or closer to 4K on XSX than PS5.

The ram advantage of PS5 I can see PS5 having more details in games.

So again a summary...

PS5 games faster load times
More detailed games.

XSX (potential for) Higher resolutions
Better RT(?)

(assuming devs use PS5's strengths) PS5 exclusives should look significantly better.
PS5 third party games aren't likely to take advantage of the ram/SSD advantages so I can see XSX having a slight edge.

I see the sound advantages as a wash.

You or anyone underestimating the huge advantages of more available high speed ram and much faster SSD is ignorant.
 
Quick lesson for Val or anyone curious about this stuff. Anyone who's ever built a PC knows that texture quality is most impacted by the GPU. The only exception is if you have a RAM bottle-neck...because you're weakest component will hold everything back. If you're not bottle-necked on memory bandwidth and are getting the most out of your GPU, then expanding your RAM will not do a thing to improve your visual performance. At that point, your GPU is now the most limiting factor in visuals.

As it relates to memory, graphics are impacted by your speed. CPU instructions and things like OS are not as much. So if you are measuring whether you have enough memory for graphic intensive stuff, it's determined after you subtract the usage for everything else. If CPU tasks and OS take up 6 GB of memory, then visuals are impacted by the speed of the remaining 10 GB.

The Series X 10 GB of fast memory is really fast. It will be used for graphics. CPU and OS will use the slow stuff. You cannot ignore CPU and OS usage. It always subtracts from the overall pie. Never will 100% of the memory be dedicated to the GPU. Even though the fast and slow memory are the same pool, GPU related tasks will primarily be riding the fast highway. Even if you want to argue that devs may sometimes change that balance and for example, use 11 GB of overall memory for graphics, it's still less than 10% riding the slow highway in that instance.

Between the fast 10 GB of memory for graphics, VRS, more advanced texture compression techniques and fact that the SSD can act as an extension of the memory, Series X should see full use of it's GPU with no memory bottle-necks. At that point, it doesn't matter who has more or less overall RAM. The thing that matters is that the memory isn't a limiting factor. So far I've not heard a single reputable and established tech outlet nor any developer leaks indicate that it might be. It's a poor value of expense and thermal management to increase the speed of the 6 GB if that's usually going to be occupied by CPU and OS and memory isn't a bottleneck.

So future Series X customers, don't let Val make your scurred. The point with the memory isn't about spinning numbers against the competition. This is an attempt to find a problem so that the other plastic can have a better solution. If there's no problem, there's no advantage. It's a matter of not being the bottle-neck and limiting factor. Same thing goes for I/O speeds. There's no indications that it is.
 
Developers always say that new hardware removes restriction and limitations. The only limitation is their imagination. Etc....

Show the games.
 
  • Like
Reactions: Frozpot
Quick lesson for Val or anyone curious about this stuff. Anyone who's ever built a PC knows that texture quality is most impacted by the GPU. The only exception is if you have a RAM bottle-neck...because you're weakest component will hold everything back. If you're not bottle-necked on memory bandwidth and are getting the most out of your GPU, then expanding your RAM will not do a thing to improve your visual performance. At that point, your GPU is now the most limiting factor in visuals.

As it relates to memory, graphics are impacted by your speed. CPU instructions and things like OS are not as much. So if you are measuring whether you have enough memory for graphic intensive stuff, it's determined after you subtract the usage for everything else. If CPU tasks and OS take up 6 GB of memory, then visuals are impacted by the speed of the remaining 10 GB.

The Series X 10 GB of fast memory is really fast. It will be used for graphics. CPU and OS will use the slow stuff. You cannot ignore CPU and OS usage. It always subtracts from the overall pie. Never will 100% of the memory be dedicated to the GPU. Even though the fast and slow memory are the same pool, GPU related tasks will primarily be riding the fast highway. Even if you want to argue that devs may sometimes change that balance and for example, use 11 GB of overall memory for graphics, it's still less than 10% riding the slow highway in that instance.

Between the fast 10 GB of memory for graphics, VRS, more advanced texture compression techniques and fact that the SSD can act as an extension of the memory, Series X should see full use of it's GPU with no memory bottle-necks. At that point, it doesn't matter who has more or less overall RAM. The thing that matters is that the memory isn't a limiting factor. So far I've not heard a single reputable and established tech outlet nor any developer leaks indicate that it might be. It's a poor value of expense and thermal management to increase the speed of the 6 GB if that's usually going to be occupied by CPU and OS and memory isn't a bottleneck.

So future Series X customers, don't let Val make your scurred. The point with the memory isn't about spinning numbers against the competition. This is an attempt to find a problem so that the other plastic can have a better solution. If there's no problem, there's no advantage. It's a matter of not being the bottle-neck and limiting factor. Same thing goes for I/O speeds. There's no indications that it is.
I built many PC's. The ASSuming is funny on your part. I doubt you have or have a clue about anything we are talking about though :)
Texture's are based on the GPU's ram not so much speed.
While nobody knows how much either is using for the OS I would hope it's not 6gb,id bet that the PS5 uses less and with more high speed ram available.
^
10gb of the XSX ram is fast but it's less than what the PS5 has to work with.
Using the SSD will help both but again PS5 has a huge advantage in this aspect as well.
It sure does matter who has more memory.

Actually More have said PS5 is better and I think the strengths of the PS5 are the reason.
 
I built many PC's. The ASSuming is funny on your part. I doubt you have or have a clue about anything we are talking about though :)
Texture's are based on the GPU's ram not so much speed.
While nobody knows how much either is using for the OS I would hope it's not 6gb,id bet that the PS5 uses less and with more high speed ram available.
^
10gb of the XSX ram is fast but it's less than what the PS5 has to work with.
Using the SSD will help both but again PS5 has a huge advantage in this aspect as well.
It sure does matter who has more memory.

Actually More have said PS5 is better and I think the strengths of the PS5 are the reason.

My point was to talk about bottle-necks. The GPU is the most expensive and most important hardware differentiator for graphics. If your GPU can't draw it fast enough, then it doesn't matter how much memory or throughput you have. Ideally if your console or PC is put together right, your GPU is your most limiting factor when it comes to graphics. The idea that Microsoft invested a ton into a monster GPU only to have it bottle-necked by memory doesn't make sense and I haven't heard a single dev state otherwise.

Here's an analogy I heard from a developer. You GPU is your car. Memory and throughput is your highway. If your car can drive 100 mph but you can only travel 60 mph due to highway congestion, then your car is no faster than a car that can drive 60 mph on that highway. If there's no congestion and the speed limit is 150mph, then you can still only go 100mph due to the speed of your car. Basically your weakest link is your limitation.

As a company that's going to mass produce this hardware, their goal is to get that balance as close as possible. There's no point in wasting money and try to manage thermals for a 200mph freeway with a 100mph car. On the other side, they don't want to invest in mass producing a 100mph car that can't drive over 60mph due to poor highway infrastructure. If I/O speed and memory bandwidth of the Series X are just enough that the GPU can travel at it's top speed, then that's all they need to be. Anything more is overkill.
 
Last edited:
I built many PC's. The ASSuming is funny on your part. I doubt you have or have a clue about anything we are talking about though :)
Texture's are based on the GPU's ram not so much speed.
While nobody knows how much either is using for the OS I would hope it's not 6gb,id bet that the PS5 uses less and with more high speed ram available.
^
10gb of the XSX ram is fast but it's less than what the PS5 has to work with.
Using the SSD will help both but again PS5 has a huge advantage in this aspect as well.
It sure does matter who has more memory.

Actually More have said PS5 is better and I think the strengths of the PS5 are the reason.

In one of the break down videos they stated that the os will use 2.5 gigs...
 
I believe the no spin “real” number for PS5 is that’s a 9.2 TF console just like the github leak stated.
 
I built many PC's. The ASSuming is funny on your part. I doubt you have or have a clue about anything we are talking about though :)
Texture's are based on the GPU's ram not so much speed.
While nobody knows how much either is using for the OS I would hope it's not 6gb,id bet that the PS5 uses less and with more high speed ram available.
^
10gb of the XSX ram is fast but it's less than what the PS5 has to work with.
Using the SSD will help both but again PS5 has a huge advantage in this aspect as well.
It sure does matter who has more memory.

Actually More have said PS5 is better and I think the strengths of the PS5 are the reason.

More who though? Lol, unnamed "developers ". Two or three. Beyond that, I saw a few of these quotes, and they were along the lines of it does some things better, or its hella powerful (not mentioning anything about be more than MS' machine), it's great, etc.

Its these fanboys that extrapolate it into a vs thing. There are also "developers " who say the SeX crushes PS5. It's all BS until we see the result.

All I know is that MS has been really transparent in talking about its features (the game stack stuff has gone into great detail), and all the Sony stuff is conjecture. Even then, the only games we've seen are minecraft (actual footage of raytracing) and a Gears 5 upgrade that runs at higher settings than a 2080ti.

They both need to start leaking some real s***.
 
I'm already bored of the fake bottlenecks. Let's talk about ML. Think it's one of the more under-rated aspects of next gen. Could drastically change the opportunity for smaller and medium sized devs.
 
More who though? Lol, unnamed "developers ". Two or three. Beyond that, I saw a few of these quotes, and they were along the lines of it does some things better, or its hella powerful (not mentioning anything about be more than MS' machine), it's great, etc.

Its these fanboys that extrapolate it into a vs thing. There are also "developers " who say the SeX crushes PS5. It's all BS until we see the result.

All I know is that MS has been really transparent in talking about its features (the game stack stuff has gone into great detail), and all the Sony stuff is conjecture. Even then, the only games we've seen are minecraft (actual footage of raytracing) and a Gears 5 upgrade that runs at higher settings than a 2080ti.

They both need to start leaking some real s***.
Still more have said it than on the XSX side.
I don't see 3rd party devs saying XSX crushes PS5.

Thought it was almost clear cut XSX over PS5 but now I dunno as they both have advantages over each other.

I stand by how I think it will turn out.
 
  • Haha
Reactions: Swede
Do you think developers would be honest and say PS5 sucks if it really does? Of course not seeing as though developers have a vested interest in seeing PS5 succeed.
 
When games are released you really expect EA for example to say the variable rate architecture of the PS5 gave us fits. So instead of figuring out this power management thing behind it we decided to target 9TF and call it there because the XSX version was done a month ago and we have to ship.
 
When games are released you really expect EA for example to say the variable rate architecture of the PS5 gave us fits. So instead of figuring out this power management thing behind it we decided to target 9TF and call it there because the XSX version was done a month ago and we have to ship.
If that's the case, and even if they don't say anything, then the multiplats might show a gap. Who knows though. People are completely melting down on Twitter over the variable clock/overclock, only ~9 TF at the base, whatever. I'm not sure just what the difference will actually be in most games, and I don't want to downplay it anymore because if it doesn't work and sustain at full clock most if the time, then we will probably know on certain games.

Still, it's all relative. Graphics on PS5 should still trounce current gen X1X, and Pro. Exclusives might look fantastic too. In fact, I will probably only get one console at launch. If they both have amazing 1st party games out if the gate, I actually may get both... if my financials allow in whatever economy we end up with by November. 🤞

What the world needs to see in this crisis is a bit of hope! We need an online E3 now! We need to see the games! LET US SEE THE GAMES!
 
I don't know how assembling PC has to do with knowledge of how graphics works, but okay.

The GPU determines the number of polygons that it can draw/unit time, the shading (like metallic, texturing etc). The ram determine how much graphical data it can hold to do all the shading.

As extreme case, a fast GPU but low ram graphic card can render a very high res (polygon) scene with low res textures or have many repeating textures.
A slow GPU but lot of rams can produce a lower res (polygon) scene but with more detail texture or less repeating textures. I wanted to point out ram is also used to store vertice data as well.
 
  • Informative
Reactions: karmakid
Status
Not open for further replies.