Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I would argue that 5050 is not low-end, it’s a 50W+ GPU after all. One should be looking at the MX series for a low-end product, but it appears discontinued. And sure, since this argument is likely to come - yes, it’s the lowest end current gen GPU that Nvidia sells, but it doesn’t make it a same class product.
From a gaming PC perspective the 5050 is low end on the Nvidia stack. They don’t sell a PC part that is lower.
That would depend on CP2077 being updated to take advantage of it. They show off Blender because Apple contributes to it, and can make sure it’s optimized for Apple Silicon.
Why would CP2077 need to be updated to take advantage of the additional RT performance?
 
  • Like
Reactions: Queen6 and maflynn
I'm saying it runs in the same power budget, and that if you want to compare GPU performance of a low end integrated part, you should compare to a low end integrated part. In PC land it is playing in the same space as machines with intel/amd integrated graphics.
*emphasis added

I think most people looking for a gaming machine, does not choose a laptop because its wattage consumption is significantly less then another laptop. Rather getting the most performance from one's budget and in this case, for less money. Do you think someone who's interested in faster GPU horsepower is going to care that the GPU is only drawing 20 watts, or that their game will run at 80FPS instead of 40?
1760873760687.png
 
  • Like
Reactions: Queen6
apple do not have an equivalent to high end discrete parts.
No they don't but apple has in the past compared apple silicon vs Nvidia, so I think in a gaming forum, talking about gaming performance, its a valid comparison. The OP opened this thread up with discussions of FPS, so when gaming, it doesn't make sense to compare a PC with an igpu, but rather against other laptops that actually can game.
 
Last edited:
  • Like
Reactions: Queen6
No they don't but apple has in the past compared apple silicon vs Nvidia, so I think in a gaming forum, talking about gaming performance, its a valid comparison
Maybe this thread should be moved to the Apple Silicon forum and title changed to M5 GPU performance.

From a gamer perspective the M5 seems okay, nothing too jaw dropping. It isn’t suddenly allowing us to play games on the MacBook at native resolution, even with light upscaling. So that means the argument will shift back to developers needing to optimize for Apple hardware and we aren’t seeing the true power of the platform because developers are lazy, or some other similar excuse.
 
  • Like
Reactions: Plutonius
From a gamer perspective the M5 seems okay, nothing too jaw dropping.
That's my point, compared to the M4, its a remarkable improvement, but I think for gaming its only fair to see how it lines up against other laptops that can game.
 
  • Like
Reactions: Queen6
Metal 4, desnoising.
Does that impact “regular” RT also or is it a PT performance improvement?

I am curious to see if the new Denise has performance impacts like Ray Reconstruction does. In games like Starwars Outlaws I see performance dips if I turn it on/up.
 
Does that impact “regular” RT also or is it a PT performance improvement?

I am curious to see if the new Denise has performance impacts like Ray Reconstruction does. In games like Starwars Outlaws I see performance dips if I turn it on/up.
So you do see how an update could change Rt performance?
 
  • Like
Reactions: M4pro
Maybe this thread should be moved to the Apple Silicon forum and title changed to M5 GPU performance.

From a gamer perspective the M5 seems okay, nothing too jaw dropping. It isn’t suddenly allowing us to play games on the MacBook at native resolution, even with light upscaling. So that means the argument will shift back to developers needing to optimize for Apple hardware and we aren’t seeing the true power of the platform because developers are lazy, or some other similar excuse.
How do you know this? What results are you basing this statement on? The only result we know of is a 44% increase (according to Apple). That seems very significant to me.
 
  • Like
Reactions: M4pro
How do you know this? What results are you basing this statement on? The only result we know of is a 44% increase (according to Apple). That seems very significant to me.
I'm basing it on their own comparisons stats that @Homy posted? They were comparing performance at 1280x800 with upscaling from balanced, so like a render resolution of 464p. As a gamer, cause that is where this thread is, that isn't great. We should not be excited over these "numbers" at such a low resolution.


Though I will be honest and say I am confused as to why CDPR hasn't updated the macOS version of CP2077 to include all the new things Apple released with macOS 26. They are usually the first to update the game with features for Nvidia (technically Alan Wake 2 beat them to the punch with mesh shaders, but that is neither here nor there).
 
I'm basing it on their own comparisons stats that @Homy posted? They were comparing performance at 1280x800 with upscaling from balanced, so like a render resolution of 540p. As a gamer, cause that is where this thread is, that isn't great. We should not be excited over these numbers at such a low resolution.
You’re making the assumption that because they quoted those settings, only those settings will benefit. It seems more sensible to say the performance improvements are unknown at this point.
Though I will be honest and say I am confused as to why CDPR hasn't updated the macOS version of CP2077 to include all the new things Apple released with macOS 26. They are usually the first to update the game with features for Nvidia (technically Alan Wake 2 beat them to the punch with mesh shaders, but that is neither here nor there).
Because it takes time. Tahoe has barely been out a month.
 
  • Like
Reactions: Rnd-chars and M4pro
Base RT performance should improve without feature designed to improve IQ.
It depends on the source of the improvements. If there are more cores, higher frequency or better utilization, then sure. If the improvements come from newer features within the hardware, then no.
Much the same way DLSS 3 is faster than 4 at the cost of image quality (at least in CP2077 for this comparison).
This is an example of newer technology being slower than old technology if you meant to write that.
 
  • Like
Reactions: M4pro
Maybe this thread should be moved to the Apple Silicon forum and title changed to M5 GPU performance.
OR how about the OP or maybe @throAU, or @leman open a more generic, general M5/Pro/Max/Ultra GPU discussion in the Apple Silicon forum. With the OP starting the discussion in terms of gaming, FPS, and ray tracing a more generalized thread may make more sense. The M5's GPU has some significant improvements that fall outside of gaming and having a more general thread has benefits.

1760877559491.png
 
  • Like
Reactions: diamond.g
You’re making the assumption that because they quoted those settings, only those settings will benefit. It seems more sensible to say the performance improvements are unknown at this point.

Because it takes time. Tahoe has barely been out a month.
In PC land, we don't use such low resolutions for performance comparisons because it mostly becomes a CPU test. So if anything the performance gap at higher resolutions should be higher if the GPU performance has really improved.

Was Metal 4 not included in the beta's of Tahoe back in June? Better yet has Apple not sent engineers to CDPR to incorporate these features? Seems like an easy layup to me.
 
This is an example of newer technology being slower than old technology if you meant to write that.
Yeah it was on purpose, DLSS 4 transformer model is slower than the CNN model but looks better. From a marketing perspective Nvidia shows that you can make your upscale more aggressive to gain performance back and still have a good looking image/game. If Apple marketed their denoiser that way, CP2077 not being updated would be fine because we wouldn't be getting a like for like performance improvement.

I also would concede that Nvidia comparing DLSS 3 to 4 also includes them enabling MFG on DLSS4 which also makes comparisons to 3 not like for like unless done on the same hardware (which it usually isn't).
 
In PC land, we don't use such low resolutions for performance comparisons because it mostly becomes a CPU test. So if anything the performance gap at higher resolutions should be higher if the GPU performance has really improved.
Pc land is so vast that it simply isn’t true to say similar resolutions aren’t used. Again you’re basing your judgement on one score. This seems very unwise.
Was Metal 4 not included in the beta's of Tahoe back in June? Better yet has Apple not sent engineers to CDPR to incorporate these features? Seems like an easy layup to me.
These things take time. Judging by some comments I’m not sure Metal 4 is even stable yet.
 
  • Like
Reactions: M4pro
Pc land is so vast that it simply isn’t true to say similar resolutions aren’t used. Again you’re basing your judgement on one score. This seems very unwise.

These things take time. Judging by some comments I’m not sure Metal 4 is even stable yet.
I'm not aware of any reviews using resolutions lower than 1080p for comparing PC GPU's without the test being for CPU performance comparisons, steam deck and its buddies notwithstanding. If you could point me to some that do that would be great.
 
This thread is in the gaming forum and we’re discussing gaming performance. Comparing the M5 to nvidia is apropropriate

From a gaming PC perspective the 5050 is low end on the Nvidia stack. They don’t sell a PC part that is lower.

Oh, I am not contesting any of this. It's obvious that if you want best gaming performance and compatibility in that price range (or any price range, really), you'd go for Nvidia.

My point is about managing expectations. M5 is a premium ultrabook platform, a product of kind that Nvidia does not currently offer*. I'd love an entry-level Mac to match a mid-range gaming GPU, I just think it's a tall order given the design constraints. After all, everyone is cooking with the same transistors. Although, given how quickly Apple progresses in terms of compute efficiency, who knows, maybe they can actually pull it off.

*Although with new Intel-NV partnership, it is possible that we will see an Intel+RTX SoC.
 
  • Like
Reactions: maflynn
I'm not aware of any reviews using resolutions lower than 1080p for comparing PC GPU's without the test being for CPU performance comparisons, steam deck and its buddies notwithstanding. If you could point me to some that do that would be great.
Lol. Moving goal posts. You’re genuinely not aware of gaming reviews of integrated gpus of Lunar Lake, Strix Point that test at 720p at all? This was a while ago the last time I looked. I don’t have time to look for others.
From Steams' hardware survey Its clear that 1080p is still largely the favored resolutions and sub 1080p occupy tenths of a percent in many cases
View attachment 2569871
Ok. Again this doesnt relate to the point in discussion. It was stated that no one tests at 720p. You’re showing me display resolution stats.
 
Last edited:
  • Like
Reactions: M4pro
Lol. Moving goal posts. You’re genuinely not aware of gaming reviews of integrated gpus of Lunar Lake, Strix Point that test at 720p at all? This was a while ago the last time I looked. I don’t have time to look for others.

Ok. Again this doesnt relate to the point in discussion. It was stated that no one tests at 720p. You’re showing me display resolution stats.
Thanks, for some reason I didn't take into account folks using iGPUs on the PC side for comparisons. I was wrong, clearly there are tests. I do find their CP2077 chart interesting as it does show what I was talking about with lower resolutions having smaller performance gaps than higher ones. So maybe Apple is sandbagging the performance increases.
 
  • Like
Reactions: OptimusGrime
The changes to the GPU in M5 are in fact quite substantial:

- A second FP16 pipe for doubled FP16 compute (used for lighting calculations, but the game needs to be optimized for it)
- Doubled the integer multiplication performance (can be occasionally useful)
- Doubled the special function performance (exp/log — useful for advanced shading)
- Yet unknown changes to shader memory management, will likely improve occupancy
- More memory bandwidth, larger caches
- Maybe: a second FP32 add pipe for improved mixed FP32 calculations (haven't properly investigated this yet)
- Maybe: additional stuff like improved memory compression etc...

They've come a long way since M1 — this is a state of the art GPU core, and industry leading in some key areas. All these improvements will certainly translate to gaming, and the improvement will be likely larger for games that utilize complex shading. We already see a 40% increase in performance on non-RT benchmarks such as Steel Nomad for A19.
As the older generation retires (looking at you Cook) and the younger gamer generations fill those seats I can see Apple putting a genuine focus on gaming like they haven’t before. I think we’re already starting to see that. We’re in a transition that is mostly being fueled by the need to keep up with AI generational leaps but People still want to game.

I’ve watched my 17 year old on his M4 mini and 1440p 144Hz HDR monitor play games with his friends online and he swaps between spaces as he’s playing to see a message or check something quickly and then swaps back within a second. He much prefers macOS’ UI to Windows and he was a Windows gamer before he decided to take the trade off in performance as he got more serious about becoming and adult and his future. He still loves gaming. Just wants the best of both worlds (don’t we all?).
 
  • Like
Reactions: Homy and M4pro
We already game on our Macs and now with the improvements to M5 GPUs our games will play at noticeably higher frame rates.

A 44% increase in frame rates in a popular AAA game is big.

Pretty simple.
 
Last edited:
As the older generation retires (looking at you Cook) and the younger gamer generations fill those seats I can see Apple putting a genuine focus on gaming like they haven’t before. I think we’re already starting to see that. We’re in a transition that is mostly being fueled by the need to keep up with AI generational leaps but People still want to game.

I’ve watched my 17 year old on his M4 mini and 1440p 144Hz HDR monitor play games with his friends online and he swaps between spaces as he’s playing to see a message or check something quickly and then swaps back within a second. He much prefers macOS’ UI to Windows and he was a Windows gamer before he decided to take the trade off in performance as he got more serious about becoming and adult and his future. He still loves gaming. Just wants the best of both worlds (don’t we all?).
Apple taking gaming 'seriously' is always just around the corner.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.