Nvidia’s RTX 5050 GPU starts at $249

That still doesn't really address your claim, but anyway. It's fine.

My claim is that the DCC in Nvidia GPUs is much better than in RDNA1/2.
So between the DCC and L2 cache, and the PS5 having to share and contend for memory bandwidth between the CPU and the GPU, they are probably even.
The problem with the 5050 is not memory bandwidth, it's memory capacity, as those 8Gb will be a huge bottleneck.
 
and feature 8GB

Life Fail GIF by MOODMAN
 
Why would you buy a card knowing that it wont run anything.

Sadly, it's largely for people who might not actually know that. Show up at a best buy, blue shirt tells you it's good - go home and wonder if something is wrong with the thing when you try to play harry potter.
 
Last edited:
So now you're just sidestepping and trying to distract me from your initial claims. Fascinating.

My claim was that the PS5 memory bandwidth and the RTX 5050 are not too dissimilar, due to Nvidia having the most advanced memory subsystem.
Something that I proved with 2 sources. So the 5050 will not be performance limited by it's memory bandwidth.

My claim that Nvidia has the best DCC wasn't even my main point, but it is still true, as the lead Nvidia had was huge.
But you are free to provide one single benchmark that shows that Nvidia lost the lead in this regard. Somehow.
 
It might match the PS5 GPU in rasterization. Though it will be severely limited in games that use more than 8Gb of vram.
In RT and AI, it does beat the PS5 very easily. And DLSS4 is a big advantage over the temporal upscalers that the PS5 can run.
Still, this is 2020 GPU performance, in 2025.
You think that thing will be as fast as the 4060? I doubt it. I see it being a good 20% slower than the console's GPU with only 8GB.
 
You think that thing will be as fast as the 4060? I doubt it. I see it being a good 20% slower than the console's GPU with only 8GB.

No, I think it's going to be as fast as a 3060, when it's not limited by it's 8Gb vram.

Something close to this:

4SpDGfgh4J7Lbtb3.png
 
Last edited:
Which is around 10-15% slower than a PS5 GPU.

The PS5 GPU is around the performance of a 3060.
Only when testing games that skew heavily to AMD's hardware, such as CoD and some Assassin Creed games, does it go above by that 10-15%.
Or when it's a DF test, with their broken test system.
 
The PS5 GPU is around the performance of a 3060.
Only when testing games that skew heavily to AMD's hardware, such as CoD and some Assassin Creed games, does it go above by that 10-15%.
Or when it's a DF test, with their broken test system.
The PS5's GPU is on par with the 6700/2070S. Both are 10-15% faster than 3060.
 
The PS5's GPU is on par with the 6700/2070S. Both are 10-15% faster than 3060.

The 6700 has 11.3 TFLOPs. While the PS5 has 10.3.
That is a 10% difference. I think that even DF did an underclock to the 6700, to better match the PS5 in comparisons.
The 2070S is almost 10% faster than a 6700 without underclocking. So it's not a proper comparison.
So the regular 2070 or the 3060 would be a better match for the PS5 GPU. Though then there is the question of some games being better optimized for AMD and some for Nvidia.
You are probably considering games like CoD and Assassins Creed, to compare with the 2070S.
 
eeehhh...

Yeah this and a low end am4 mobo bundle. And a case/fans. And storage. And a psu. And a couple knick-knacks to finish. Ehhh, indeed. ~$650 to *almost* handle new games with that 8gb. Far from "killer".

This isn't supposed to be super high end. But access to dlss 4 for less than $300 (again...that price is a pipe dream) would be a great edition to extremely small form factor pcs.
Dropping $600 bucks to get a rig like this for someone just getting into PC gaming looks* like it would be fantastic.

I have a 280TI super in a nzxt H1 micro case and it crushes games at 1080p 60fps-120fps...well it crushes Elden Ring, Fortnite, PUBG, and R6. All the stuff my nephew and his friends play.
Ill try to grab one at MSRP (looooooooool) and give it a go, my nephew won't care esp if his fps goes up.
 
The 6700 has 11.3 TFLOPs. While the PS5 has 10.3.
Yes, so?
That is a 10% difference. I think that even DF did an underclock to the 6700, to better match the PS5 in comparisons.
How many more TFLOPs does the SX have over the PS5 again?
The 2070S is almost 10% faster than a 6700 without underclocking. So it's not a proper comparison.
So the regular 2070 or the 3060 would be a better match for the PS5 GPU. Though then there is the question of some games being better optimized for AMD and some for Nvidia.
You are probably considering games like CoD and Assassins Creed, to compare with the 2070S.
No, I'm going based on the myriads of benchmarks we have. There is almost no game in which a PS5 doesn't easily beat the 2070, come on.
 
Should have been the price of the pathetic RTX 5060 8GB while this rtx 5050 becomes 200 bucks or less. And the RTX 5060 TI 16GB is 350 at most.
 
Yes, so?

How many more TFLOPs does the SX have over the PS5 again?

We are comparing the PS5. Which has 10.3 TFLOPs.
And the Series X has lower clock speeds that affect how fast units can work, resulting in lower performance.
The 6700 doesn't have that issue.

No, I'm going based on the myriads of benchmarks we have. There is almost no game in which a PS5 doesn't easily beat the 2070, come on.

Does it? Show me the benchmarks.
Or are you going to show some DF video, where they can't even configure the system memory and the Windows installation.
 
Yes, so?

How many more TFLOPs does the SX have over the PS5 again?

No, I'm going based on the myriads of benchmarks we have. There is almost no game in which a PS5 doesn't easily beat the 2070, come on.
yup i can support your claim

something happened after 2023, outliers increased. to a point i cannot call them outliers anymore

gow ragnarok runs horribly on 3060 compared to ps5. spiderman 2 too. i accepted these as outliers as they were ps5 1st party games

but i cannot call ac shadows and doom dark ages outliers anymore. not even with the assumption that assassin's creed games favoring AMD or something. the game has heavy ray traced GI.

ps5 runs doom dark ages with base ray traced gi at 4k fsr performance (around 1080p internal resolution average) at around 50-60 FPS. most of the time 60 but i had to mention

even if we assume ps5 runs the game at low preset (i highly doubt, it is probably a mix of lows, mediums and highs similar to most console games)



this is what 3060 gets you. way below what a ps5 is capable of. if we assume ps5 gets 4k fsr performance 55 fps at low settings here (it really doesn't drop below 1000p and 50 fps and that is with the worst case scenario possible) there's still a massive performance gap there. you would literally need a 3070 or 3060ti to get to 55-60 fps in this scene

situation with ac shadows is even more dire. ps5 gets you ray traced GI between 1250p and 1500p at 30 fps. meanwhile 3060 gets you that similar performance... with ray tracing disabled and dlss performance.



i don't care what anyone says. these are two big 3rd party AAA games with baseline ray tracing implementation. i don't care if nvidia GPUs destroy consoles in nitpicked alan wake 2 and cyberpunk. i'd rather make assumptions on doom and ac shadows rather than alan wake 2 or something

this is not to say you can't turn things around by using DLSS 4. but that is another discussion.
 
Last edited:
Does it? Show me the benchmarks.
Or are you going to show some DF video, where they can't even configure the system memory and the Windows installation.
You mean the ones where they used a freakin' 13900K somehow was suboptimal for a 6700?



Now please, show us the benchmarks where the 3060 matches a PS5 in overall performance.
 
i don't care if nvidia GPUs destroy consoles in nitpicked alan wake 2 and cyberpunk.
So you don't care if games that utilise feature set of Nvidia's tech the most significantly outperform console hardware. Hmm interesting.
Very interesting.

You would think AW2 and Cyberpunk are 2 games you'd want to look at more often.
 
You mean the ones where they used a freakin' 13900K somehow was suboptimal for a 6700?



Now please, show us the benchmarks where the 3060 matches a PS5 in overall performance.


You did notice that in several of those games, the 6700 beat the PS5? As expected by a GPU that ha 10% more TFLOPs.
But in some games, the 6700 won by a huge margin.
For the most part, it took bad PC ports, for the PS5 to win.
 
You did notice that in several of those games, the 6700 beat the PS5? As expected by a GPU that ha 10% more TFLOPs.
Yes and in others, the PS5 beat it. Furthermore, MHR and Hitman 3 are both PS4 games.
But in some games, the 6700 won by a huge margin.
For the most part, it took bad PC ports, for the PS5 to win.
Cyberpunk with RT and A Plague Tale Requiem are bad PC ports? Point is, the PS5 sticks dangerously close to the 6700 and 2070S, usually within 5-10% and it can actually beat them in some cases. The 3060 and 2070 aren't on that level. The PS5 in those tests consistently beat the 3060 and was much closer to the 2070S/6700.
 
Last edited:
So you don't care if games that utilise feature set of Nvidia's tech the most significantly outperform console hardware. Hmm interesting.
Very interesting.

You would think AW2 and Cyberpunk are 2 games you'd want to look at more often.
alan wake 2 was released in 2023
cyberpunk 2077 was released in 2020

doom and ac shadows, this year

there were literal games where 2060 super seemed to match xbox series x in ray tracing WITH increased base ray tracing settings.



all of this points to the fact that optimizing ray tracing around consoles were not a priority back then. with doom and ac shadows, it feels like things have changed and that massive advantage is not there anymore. and i don't see any outlier games in 2025 that favors RTX GPUs like alan wake 2. only if you run path tracing you truly get an astounding experience but then again you won't be running path tracing on a ps5 or a 3060 so it's not exactly relevant for this comparison

this is coming from an actual rtx 3070 user who enjoyed the "faster" ray tracing experience in alan wake 2 and cybeprunk 2077 back in the day. playing doom dark ages and ac shadows, it literally felt I was getting the same performance as a PS5. it's not even VRAM, I've checked rtx 4060ti 16 GB benchmarks and my performance is on par with that as long as I keep texture settings in check. thankfully 1440p dlss balanced with DLSS 4 looks clear and gets really high framerates which something PS5 cannot do
 
Last edited:
Yes and in others, the PS5 beat it. Furthermore, MHR and Hitman 3 are both PS4 games.

Cyberpunk with RT and A Plague Tale Requiem are bad PC ports? Point is, the PS5 sticks dangerously close to the 6700 and 2070S, usually within 5-10% and it can actually beat them in some cases. The 3060 and 2070 aren't on that level. The PS5 in those tests consistently beat the 3060 and was much closer to the 2070S/6700.

RT is something that we have spoken about in RDNA2, between PC and the PS5. RT on PC is severely bottleneck by DirectX DXR API, something that even DF spoke about.
So in this situation, it's not the hardware that is faster, it's Sony's API that is better.
And the results in Plague Tale, are very similar between the PS5 and the 6700.
But then there are the games where the PS5 loses to the PS5. Such as the Hitman, Monster Hunter and Avatar.
 
RT is something that we have spoken about in RDNA2, between PC and the PS5. RT on PC is severely bottleneck by DirectX DXR API, something that even DF spoke about.
So in this situation, it's not the hardware that is faster, it's Sony's API that is better.
And the results in Plague Tale, are very similar between the PS5 and the 6700.
But then there are the games where the PS5 loses to the PS5. Such as the Hitman, Monster Hunter and Avatar.
Changes nothing to what I said. They're very close and the 2070S/6700 are the closest PC GPUs. The 3060 gets beaten 90% of the time by a margin of 10% or more.
 
Changes nothing to what I said. They're very close and the 2070S/6700 are the closest PC GPUs. The 3060 gets beaten 90% of the time by a margin of 10% or more.

Did you notice that in those DF benchmarks you showed, that the 3060 is losing by a bigger margin than it should, compared to the 2070S.
Techpowerup benchmarks show a 8% difference between these 2 GPUS.
But in the DF benchmark images of avatar that you posted, the 3060 loses by 15%. And in Plague Tale, by 17%.

The 6700 is 10% faster than the PS5. And the benchmarks you posted, on average show something similar.
Only when you cherry pick a few games does that change.
And the comparisons to the RTX cards seem very flawed, as several games are skewed to AMD.
And there seems to be some results that simply don't match up for the 3060, as this should be 8-10% slower than a 2070S. But some of your pictures show a difference of 15% or more.
 
This isn't supposed to be super high end. But access to dlss 4 for less than $300 (again...that price is a pipe dream) would be a great edition to extremely small form factor pcs.
Dropping $600 bucks to get a rig like this for someone just getting into PC gaming looks* like it would be fantastic.

I have a 280TI super in a nzxt H1 micro case and it crushes games at 1080p 60fps-120fps...well it crushes Elden Ring, Fortnite, PUBG, and R6. All the stuff my nephew and his friends play.
Ill try to grab one at MSRP (looooooooool) and give it a go, my nephew won't care esp if his fps goes up.

I don't know what the next card up that's going to fit in that case is, but disregarding that, a new build with this thing is in a no man's land. 150 bucks more on the card, and can actually count on playing games coming out for the next several years. On the other hand, that's still 20 or 25% more on the build.

This thing might make sense as an upgrade from something really bad to get some more frames in fortnite or whatever? Or maybe if you've already got that small form factor build from a long time ago and want to give it a little more kick?
 
Or maybe if you've already got that small form factor build from a long time ago and want to give it a little more kick
That's my thought about this card as well. I personally wouldn't have a main rig with any card in this line (outside of the 5090fe...at initial msrp) but I am curious how it will perform in mid-low to low end builds. My nephews will be a great experiment. I was able to get the new intel card into this case but I didn't close it up.

I would need to get a FE version of the card (I am assuming they have one...need to verify that) for it to fit in that case.
 
Top Bottom