Epic aim to make their UE 5 demo run at 60 FPS on consoles + Is the PS 5 demo ouperformed by a mid-tier PC?

[3rrant]

Just now, Delicieuxz said:

I guess I think of graphics card tiers in a way that normalizes the entire spectrum. I think of tiers as available performance, not cost, with the mid-tier being the middle of the spectrum of performance options. Otherwise, there's like 1 low-tier GPU and 1 mid-tier GPU from a manufacturer, and then the high-end has its own vast spectrum. And that seems awkward, to me.

 

As was clarified above, the origin of the '40 FPS on a 2080 Mobile PC' claim is not the video that was shown during the livestream. The source is the Epic Games China engineer who stated during that livestream that the demo (whether he was referring to the first part of it, or the demo as a whole) runs at 40 FPS on his notebook, and the comment that the UE 5 demo runs at 40 FPS in their editor.

 

The engineer's quoted comment is: "Our goal is that the graphic quality like this demo, we want to make it run 60FPS at next-gen consoles. But now we do not reach the goal. Now it is 30FPS. Our target is 60FPS, that is also why we can not release it now. And I can assure you that we can run this demo in our notebook, in editor , not cooked, it even can 40FPS."

The meaning of the original quote was lost in translation man, that's what Tim said in his tweets. 

[JoostinOnline]

11 hours ago, Delicieuxz said:

For reference, according to UserBenchmark

I stopped reading here. Why the fuck are people still using Userbenchmark?

[Delicieuxz]

7 minutes ago, 3rrant said:

The meaning of the original quote was lost in translation man, that's what Tim said in his tweets. 

Not really. What Tim said is that he doesn't know what the translation actually is, but that a claim of '40 FPS on a PC' can't be used to make a comparison of performance to the PS 5 based on its 30 FPS demonstration because the PS 5 demo itself doesn't run at just 30 FPS, but runs at 30 FPS and higher, in order to achieve a constant 30 FPS when vsync is enabled.

 

Tim didn't deny that the '40 FPS on a PC' claim is true and he didn't comment on its veracity. He only said that the claim can't serve to compare PC to PS 5 performance because the PS 5 demo wasn't simply running at 30 FPS.

[3rrant]

4 minutes ago, Delicieuxz said:

Not really. What Tim said is that he doesn't know what the translation actually is, but that a claim of '40 FPS on a PC' can't be used to make a comparison of performance to the PS 5 based on its 30 FPS demonstration because the PS 5 demo itself wasn't running at just 30 FPS, but was running at 30 FPS and higher in order to achieve a constant 30 FPS for the demo.

 

Tim didn't deny that the '40 FPS on a PC' claim wasn't true, and he didn't comment on its veracity. He only said that claim can't compare PC to PS 5 performance because the PS 5 demo wasn't running at simply 30 FPS.

"Not sure exactly what was said" + "The translated quote on 40 fps isn't a comparison between hardware capabilities" + ".. a video player"

= No it's not running on a PC. Dunno why you keep going with it. He said it clear: the demo was not running on a PC.

 

Other proof:

[Delicieuxz]

33 minutes ago, 3rrant said:

"Not sure exactly what was said" + "The translated quote on 40 fps isn't a comparison between hardware capabilities" + ".. a video player"

= No it's not running on a PC. Dunno why you keep going with it. He said it clear: the demo was not running on a PC.

You're not understanding what Tim said.

 

"Not sure exactly what was said. The translated quote on 40 fps isn't a comparison between hardware capabilities. To hit a 100% solid 30 fps with vsync in the demo, per-frame times vary from 30 fps up to much higher."
 

This is the breakdown for Tim's comment:

"Not sure exactly what was said."

 

That's not Tim saying that the meaning was lost in translation and that what the EGC engineer meat was that the video of the PS 5 demo was running at 40 FPS. That's Tim literally saying that he doesn't know what the EGC engineer said and so he can't comment on it.

 

"The translated quote on 40 fps isn't a comparison between hardware capabilities. To hit a 100% solid 30 fps with vsync in the demo, per-frame times vary from 30 fps up to much higher."

 

And that's Tim saying exactly what I just said:

37 minutes ago, Delicieuxz said:

... that a claim of '40 FPS on a PC' can't be used to make a comparison of performance to the PS 5 based on its 30 FPS demonstration because the PS 5 demo itself doesn't run at just 30 FPS, but runs at 30 FPS and higher, in order to achieve a constant 30 FPS when vsync is enabled.

 

Tim didn't deny that the '40 FPS on a PC' claim is true and he didn't comment on its veracity. He only said that the claim can't serve to compare PC to PS 5 performance because the PS 5 demo wasn't simply running at 30 FPS.

 

To be clear, Tim did not say that the '40 FPS on a PC' claim comes from the demo shown on the PC during the livestream or that the claim isn't true. And I'm not claiming that the demo shown during the livestream was running in real-time on the laptop. The video shown of the demo during the livestream has nothing to do with the '40 FPS on PC' claim - they aren't related. The '40 FPS on a PC' claim comes from the EGC engineer's statement during the livestream, which Tim was unaware of when he was first asked about the claim and shown a photo of the livestream.

[3rrant]

1 minute ago, Delicieuxz said:

cut

I'm gonna stop responding because this is turning in a 1on1 discussion.

You did not read the context of his discussion.

 

This entire discussion was translated by a random user who claims to know chinese but also claims that has VERY POOR ENGLISH knowledge. The user is ゆめしん@wangxingyu1999 on twitter. He made the claims, which made no sense, and Tim simply pointed out that he has no clue about what the guy in the original stream said and why, but it's not about how the benchmark runs on PC because the benchmarks never run on a PC in the first place! It's not hard to get that.

 

 

[Delicieuxz]

59 minutes ago, 3rrant said:

I'm gonna stop responding because this is turning in a 1on1 discussion.

You did not read the context of his discussion.

 

This entire discussion was translated by a random user who claims to know chinese but also claims that has VERY POOR ENGLISH knowledge. The user is ゆめしん@wangxingyu1999 on twitter. He made the claims, which made no sense, and Tim simply pointed out that he has no clue about what the guy in the original stream said and why, but it's not about how the benchmark runs on PC because the benchmarks never run on a PC in the first place! It's not hard to get that.

You're wrong. I read the entire context of Tim's comments, and they clearly say something very differently than what you have claimed. And the livestream has been translated by more than 1 person, and the EGC engineer's statement that the demo runs at 40 FPS on his notebook has been confirmed by multiple people.

 

But why did you post this? As I've already told you, multiple times, it has nothing to do with the topic. And so, it isn't proof of anything:

1 hour ago, 3rrant said:

Other proof:

 

It has never been debated that a video of the PS 5 demo was shown during the livestream. During the livestream, multiple things happened, including: A video of the PS 5 demo was shown, and the EGC engineer talked about the performance the demo gets on his notebook.

 

So, yes, a video of the PS 5 demo was shown on a laptop during the livestream, and that screenshot you posted is of a video of the PS 5 demo and not the demo running in real-time on the laptop. That's a given and not at-all disputed. And that's what confused Tim when he was first asked about it while being shown a screenshot. But then it was clarified to him that the question and claim are not about the video that was shown during the livestream, but about the EGC engineer's comments, such as:

 

"Our goal is that the graphic quality like this demo, we want to make it run 60FPS at next-gen consoles. But now we do not reach the goal. Now it is 30FPS. Our target is 60FPS, that is also why we can not release it now. And I can assure you that we can run this demo in our notebook, in editor, not cooked, it even can 40FPS."

 

And Tim accepted the clarification and said he couldn't comment on the EGC engineer's statement because he doesn't know exactly what was said. Why can't you also accept the clarification like Tim did?

 

For some reason, you're not getting the point that I've told you over and over, which is that the video shown during the livestream has nothing to do with the '40 FPS on a PC with an RTX 2080 Mobile' claim. They're completely unrelated. Yes, a video of the PS 5 demo running on a laptop was shown during the livestream. I haven't once argued against that. But it's completely irrelevant to the subject of the EGC engineer stating that the demo runs at 40 FPS on his notebook.

[Tech Enthusiast]

37 minutes ago, JoostinOnline said:

I stopped reading here. Why the fuck are people still using Userbenchmark?

Because the relative performance graphs are still accurate.

No one forces you to read their blog posts, if you don't like them.

 

There simply is no faster way to check relative performance of all the products out there. You simply won't find a review of "insert my current gpu" vs "all the GPUs i consider for an upgrade".

 

Once you settle on one that fits your cost / performance needs, you can go ahead and double check with reviews if it really is good for you. But the first glance on userbenchmark can save plenty of hours of research and no one else offers this sadly.

[Doobeedoo]

I'd expect them too. For early demo it's quite good, but yeah 1440p 30fps though. Consider that in actual gamr there will be a lot more going on with assets and AI and other things, it will be quite interesting to see games made with this. Especially 60fps should be the minimum target, no reason not to. Only some dumb devs opting going 30 by choice as a "design choice" limitations ahm incompetence. Hopefully not or better have mode options too. Sucks that 60fps is not mandated at this point. 

Also, cool to compare for sure, but I love how everyone are so engraved in stone compare current PC hardware to consoles that are half a year from release as the only thing. Before consoles release we'll have next gen CPUs and GPUs as well so that will be more of a compare to see. 

But yeah it's early stages, can't wait to see new and improved dems and games. 

[CarlBar]

5 hours ago, 3rrant said:

I'm gonna stop responding because this is turning in a 1on1 discussion.

You did not read the context of his discussion.

 

This entire discussion was translated by a random user who claims to know chinese but also claims that has VERY POOR ENGLISH knowledge. The user is ゆめしん@wangxingyu1999 on twitter. He made the claims, which made no sense, and Tim simply pointed out that he has no clue about what the guy in the original stream said and why, but it's not about how the benchmark runs on PC because the benchmarks never run on a PC in the first place! It's not hard to get that.

 

 

 

Please provide a screenshot quote of tim's posts saying what your claiming.

[3rrant]

6 minutes ago, CarlBar said:

 

Please provide a screenshot quote of tim's posts saying what your claiming.

Click on any of the tweets on the first page > Tim's profile > tweets and answers > read.

[RejZoR]

17 hours ago, SupaKomputa said:

Who cares? The laptop with the RTX 2080 probably cost way more than 500 bucks (assumed ps5 price).

Why would you compare a severely thermally and power restricted laptop? Compare it to a freaking PC...

[xAcid9]

17 hours ago, Delicieuxz said:

According to that, the UE5 demo gets 33% more FPS when running on a PC with an RTX 2080 Mobile GPU, where it experiences 40 FPS, than it does when running on the PS 5, where it experiences 30 FPS.

I thought the PS5 demo they shown was capped at 30fps? Nobody know how many fps it can get w/o frame limiter.

Anyway, i'll just wait for them to release the demo on so i can try it on my low-tier PC. 

[Sauron]

@ryao clarified that Sweeney's statement on the throughput was believable.

 

Even so, don't forget, these are basically just PCs and even if you can load assets much faster the GPU can only render so much.

[ryao]

10 minutes ago, Sauron said:

@ryao clarified that Sweeney's statement on the throughput was believable.

 

Even so, don't forget, these are basically just PCs and even if you can load assets much faster the GPU can only render so much.

I do not think I said that. I said that the performance should be comparable to a GeForce GTX 1080 Ti (based on the reported GFLOPS numbers) before considering ray tracing and for you to tell me. I don't doubt what they are saying, although my area is storage, not games/graphics, so I am not the best person to ask.

[Sauron]

Just now, ryao said:

I do not think I said that. I said that the performance should be comparable to a GeForce GTX 1080 Ti before considering ray tracing (based on the reported GFLOPS numbers) and for you to tell me. I don't doubt what they are saying, although my area is storage, not games/graphics, so I am not the best person to ask.

You did say that the throughput is probably higher, or did I misunderstand? The second line of my post was my opinion in case that wasn't clear

[CarlBar]

1 hour ago, 3rrant said:

Click on any of the tweets on the first page > Tim's profile > tweets and answers > read.

 

18 hours ago, 3rrant said:

Fake news.

Tim Sweeney already specified on Twitter that the laptop in question was running A VIDEO of the demo running on a PS5. 

https://twitter.com/TimSweeneyEpic/status/1261779320101572609?s=20

 

Is this what your referring to? Because that's not even referencing the thing your discussing.

 

The thing your discussing is the statement by Eopic china regarding how it performs when the demo is run on a laptop. Not how the video of the PS5 demo runs on a laptop. 

[ryao]

10 minutes ago, Sauron said:

You did say that the throughput is probably higher, or did I misunderstand? The second line of my post was my opinion in case that wasn't clear

How about I tell you what I actually think.

 

"The Xbox One S and All-Digital Edition are both capable of up to 1080p graphics running at 60fps (as on games such as Forza Horizon 4)."

 

https://www.pocket-lint.com/games/buyers-guides/xbox/140841-xbox-one-x-vs-xbox-one-s-what-s-the-difference

 

Going to 120fps requires double the framerate while going to 4k requires quadrupling it. The XBox One S GPU does 1.4Tflops according to the specifications. 8 times that would give you a 11.2Tflops. The PS5 GPU does 10.28TFlops. If you consider archiectural improvements between the pre-Polaris GPU in the Xbox Series S and the RDNA2 GPU in the PS5, you likely reach what is required for 4k at 120fps. If you increase detail, you won't get to 120 fps, but 60 fps is probably attainable as long as the detail increase is only a factor of 2 (although architectural improvements might allow for more than that). Also, the addition of raytracing, if done without imposing bottlenecks, could also result in better detail as GPU resources would be freed from calculating shadows, lightning, reflections, etcetera through traditional techniques.

 

Note this is not my area of expertise, so take my napkin math and speculation as being from an enthusiast rather than an expert.

[sof006]

Well regardless of what has been discussed here I am just excited for the PS5 and new Xbox as the release of them brings "new" technology to the masses which allows for devs to improve their engines to leverage the tech fully. UE5 looks amazing and I can't wait to see it officially come out so we can start to see games use the technology. Seeing that demo has gotten me very hyped for "next-gen". 

 

I use "" because obviously new and next-gen is all held back purely by consoles and the technology they ultimately decide to use. 

[Raptor_Fawr]

17 hours ago, SupaKomputa said:

1080 can't do ray tracing, what do you think the lighting comes from?

Guys why are you so overwhelmed by ray tracing? It's a very cool feature to have in a game, but it's not a defining characteristic of a good game. 

 

Also, playstation 2 was capable of ray tracing as well so to have it on ps5 it's good but it's not really a "better than pc" feature

[ryao]

44 minutes ago, Sauron said:

You did say that the throughput is probably higher, or did I misunderstand? The second line of my post was my opinion in case that wasn't clear

I just read the original article. It sounds to me like they have the streaming asset loads working on a PC. Perhaps you meant in terms of the PC being able to support the bandwidth. That is possible, although I would not expect the average PC to be able to do it. 

 

The SSD that they used in that laptop supports 3400MB/sec reads. If they are efficiently doing their own compression/decompression on top of that, the bandwidth probably isn't far from what the PS5 does and could be faster than the XBox Series X. Anyway, Epic would need to tell us things like how fast of a SSD is enough and whether compression is being used.

[Raptor_Fawr]

As for the "mid tier" discussion, I agree that calling a 1080 mid tier is not really a match in our mind, but let's face it guys if you have to look at what options a user can buy BRAND NEW I'd call a 1080 mid to high tier, not at all high tier. 

 

A costly mid to high graphic card, but still... 

[Commodus]

1 hour ago, Raptor_Fawr said:

As for the "mid tier" discussion, I agree that calling a 1080 mid tier is not really a match in our mind, but let's face it guys if you have to look at what options a user can buy BRAND NEW I'd call a 1080 mid to high tier, not at all high tier. 

 

A costly mid to high graphic card, but still... 

Here's the thing, though: when the PS4 and Xbox One came out, people complained that they had graphics equivalent to recent mid-range GPUs from computers.  They still produced great results.  And this time around, the systems will be quite current in terms of feature support.  If you could get equivalent performance to a GTX 1080 Ti with hardware ray tracing thrown in, I'd say that's a pretty good value for a console.

[Delicieuxz]

Obviously people have some different opinions on this. But I think that classifying GPU tiers without considering for how they fit into the performance spectrum just doesn't make sense. If someone lists a used GTX 1060 GPU for $1,000, that doesn't make it a high-tier GPU. And if Nvidia decide to increase the prices on their GTX 1660 GPUs so that they cost the same as an RTX 2080, that doesn't make them high-tier GPUs and comparable to an RTX 2080.

 

This is how I currently consider Nvidia's gaming GPUs.

 

Low-tier: GTX 1060, all forms of GTX 1660,

 

Upper low-tier: GTX 1070, RTX 2060

 

Mid-tier: GTX 1080, RTX 2060 Super, RTX 2070

 

High-tier: GTX 1080 Ti, RTX 2070 Super, RTX 2080

 

Top-tier: RTX 2080 Super, RTX 2080 Ti

 

 

And when the RTX 30XX series releases, everything will likely get bumped down a tier.

 

I hear the view about price-points and it just doesn't make sense to me. I don't need to rationalize my GTX 1070 as being a mid-tier or high-tier GPU to make myself feel better about owning it rather than an RTX 2070, 2080, or 2080 Ti, even though I'd love to have any of those cards. I don't know if that's a factor for everyone, but I wonder if that's the only way the price-point barometer makes sense as the default one to use when describing one's GPU. It seems a lot more functional in a discussion to me to describe tiers as normalized across the spectrum rather than have most everything bunched into the high-tier and above category.

 

Also, adjusting GPU tiers when Nvidia inflates their prices encourages greater exploitation from Nvidia: If people adopt the view that a mid-tier GPU is a high-tier GPU, then Nvidia will feel justified in charging high-tier prices for a mid-tier GPU. If you don't acknowledge that the card Nvidia is charging $500 for is actually just a mid-tier card, then you let them get away with price-gouging - which is exactly what Nvidia is doing with their hyperinflated prices.

 

So, acknowledging the performance levels of graphics cards also maintains some semblance of accountability of GPU manufacturers. Lowering your expectations to conform to Nvidia's price-gouging is letting Nvidia trample over you without any protest.

[Tech Enthusiast]

Im with you on that one.

Tiers are not dictated by an artificial price limit, but by what is available on the market.

 

A 1080 is no longer high-end, because even the lower end current gen GPUs offer similar performance.

And while a 1080 is enough for most 1080p gamers, it hardly makes 4k@60 viable, let alone with amazing settings. So no, it is not more than a mid tier GPU.