navierStokes
Boo
Just watched Geekerwan's video of the iPad Pro M4, I have some thoughts and doubts, but I'll spoiler it, because it's a lot and not completely relevant to the thread apart from being the first product made on the latest process node from TSMC (N3E).
BG3 and DD2 likewise are quite interesing ports, because of some of their technical hurdles or limitations.
However, I've seen how BG3 performs on the lowest end mac hardware (M1 with 8GB) and it's not pretty. I think it performs worse than the steam deck, especially due to its RAM limitations. With the 16GB configuration it gets a lot more headroom to play with and the frametime consistency is better, but not ideal.
Both of these titles also are continuously changing in performance starting from the launch build (or day 1 patch) and depending on how the state will be when the Switch 2 arrives, the perspective of what's possible with the hardware will shift as these games have hopefully completed their "final-state" .
Like Patch 7 of BG3 states;
But if you were to freeze the build of the game and let's say run it on theoretical T239 hardware, how would it do, I guess that's the question
. AAA 3rd party games are just a big ? at times.
Moreover, if Elden Ring can run on the Switch 2, I'm actually not that satisfied with how it is on the PS5, because it still fluctuates in framerate and the RT mode is just a no-go, so let's say if the devs make odd decisions, the game might be running in a state that's less ideal compared to the steam deck.
Any idea of what the settings are here, in city its sub-30, but the IQ looks fine (not amazing, but I imagine on the SD you won't have much complaints).
Geekerwan did a deep dive on the lastest M-series processor from Apple, M4. Made on the "2nd-generation of TSMC 3nm process", which is likely TSMC N3E.
It had 16GB RAM with 120 GB/s bandwidth, 9-10C CPU (3/4P + 6E) with peak frequency of 4.5 Ghz (?), 10-core GPU with a peak frequency of 1.47 GHz.
There's still some doubts in regards to their results, especially with the power consumption and just their overall metrics, because I presume it's peak value, which does not sustain throughout the test. Moreover, they try to provide results in "non-regular-use" conditions (for their CPU tests they literally try to cool the device with liquid nitrogen) to compare it with the previous generation (M3), but what's the baseline value....? It's not clear to me.
It's a really odd review and they should've just waited for the macbook pro with a fan if they wanted to appropriately estimate.
I'll cut to the chase, three games (Resident Evil 4, Death Stranding, COD Warzone) they were interested in, did not provide user-configurable settings to push the game and chip to its limit and see what's possible in this passively cooled form tablet.
Thus developers should update the app such that they can make use all the available power.
1 game, allowed to configure your own settings, RE Village. So they maxed it out with a resolution of ~2K (roughly close to 1440p) native, it resulted in an average of ~49 FPS. However, although we do not know the intensity of the scene they tested at. So let's give it a penalty and presume the average fps is ~30 FPS.
If I understand correctly, the total power consumption of the device is limited to ~10.5W when running these large games (sustained profile), to preserve battery life and heat generation. There's no exact data that you can easily capture regarding its clock speed of the CPU and GPU and I'd like to say you can estimate through their CPU/GPU synthetic curves, but because of their testing methodology I'm not sure how exactly they acquired each point. It's difficult to understand the validity.
Nontheless if these result were to be believed how does RE Village compare relatively to existing platforms?
On the PS4 Pro is ~1080p/60fps with its performance and with the resolution mode it's 2160p (checkerboarded) and a 30fps baseline (with unlocked fps). Moreover, although it's not actually estimated, but it's likely not running at the maximum presets. I do not know the exact internal resolution that the game has in its resolution mode and we need to have more results or better relative comparisons, bu when it comes to this title specifically you could say that Apple pretty much has a PS4 Pro level of performance in their tablet form factor, not synthethics only, but really from a results standpoint.
Compared to some mobile/desktop GPUs, at 1440p max preset, if I presume it's always around 30fps it's above the 1650 Mobile. For reference the RTX 3050 mobile as per notebookcheck's testing has an average of ~41.6 FPS.
How does that change the landscape of handhelds or portables?
Well, it's an Apple device and I am an enthusiastic user of their products, the reality is that not all games are going to be on it, there are going to be games released, but not all of them. They're also not going to release a gaming handheld, so the full performance of this SoC will be in their notebooks. Moreover the device configuration price they tested is well above any console.
If switch 2 is on TSMC 4N, I think it's going to fare quite well compared to this cutting-edge SoC, especially when docked. RE Village with DLSS to 4k and settings, which exceed what the PS4 Pro would be wonderful at the console price-range.
It had 16GB RAM with 120 GB/s bandwidth, 9-10C CPU (3/4P + 6E) with peak frequency of 4.5 Ghz (?), 10-core GPU with a peak frequency of 1.47 GHz.
There's still some doubts in regards to their results, especially with the power consumption and just their overall metrics, because I presume it's peak value, which does not sustain throughout the test. Moreover, they try to provide results in "non-regular-use" conditions (for their CPU tests they literally try to cool the device with liquid nitrogen) to compare it with the previous generation (M3), but what's the baseline value....? It's not clear to me.
It's a really odd review and they should've just waited for the macbook pro with a fan if they wanted to appropriately estimate.
I'll cut to the chase, three games (Resident Evil 4, Death Stranding, COD Warzone) they were interested in, did not provide user-configurable settings to push the game and chip to its limit and see what's possible in this passively cooled form tablet.
Thus developers should update the app such that they can make use all the available power.
1 game, allowed to configure your own settings, RE Village. So they maxed it out with a resolution of ~2K (roughly close to 1440p) native, it resulted in an average of ~49 FPS. However, although we do not know the intensity of the scene they tested at. So let's give it a penalty and presume the average fps is ~30 FPS.
If I understand correctly, the total power consumption of the device is limited to ~10.5W when running these large games (sustained profile), to preserve battery life and heat generation. There's no exact data that you can easily capture regarding its clock speed of the CPU and GPU and I'd like to say you can estimate through their CPU/GPU synthetic curves, but because of their testing methodology I'm not sure how exactly they acquired each point. It's difficult to understand the validity.
Nontheless if these result were to be believed how does RE Village compare relatively to existing platforms?
On the PS4 Pro is ~1080p/60fps with its performance and with the resolution mode it's 2160p (checkerboarded) and a 30fps baseline (with unlocked fps). Moreover, although it's not actually estimated, but it's likely not running at the maximum presets. I do not know the exact internal resolution that the game has in its resolution mode and we need to have more results or better relative comparisons, bu when it comes to this title specifically you could say that Apple pretty much has a PS4 Pro level of performance in their tablet form factor, not synthethics only, but really from a results standpoint.
Compared to some mobile/desktop GPUs, at 1440p max preset, if I presume it's always around 30fps it's above the 1650 Mobile. For reference the RTX 3050 mobile as per notebookcheck's testing has an average of ~41.6 FPS.
How does that change the landscape of handhelds or portables?
Well, it's an Apple device and I am an enthusiastic user of their products, the reality is that not all games are going to be on it, there are going to be games released, but not all of them. They're also not going to release a gaming handheld, so the full performance of this SoC will be in their notebooks. Moreover the device configuration price they tested is well above any console.
If switch 2 is on TSMC 4N, I think it's going to fare quite well compared to this cutting-edge SoC, especially when docked. RE Village with DLSS to 4k and settings, which exceed what the PS4 Pro would be wonderful at the console price-range.
BG3 and DD2 likewise are quite interesing ports, because of some of their technical hurdles or limitations.
However, I've seen how BG3 performs on the lowest end mac hardware (M1 with 8GB) and it's not pretty. I think it performs worse than the steam deck, especially due to its RAM limitations. With the 16GB configuration it gets a lot more headroom to play with and the frametime consistency is better, but not ideal.
Both of these titles also are continuously changing in performance starting from the launch build (or day 1 patch) and depending on how the state will be when the Switch 2 arrives, the perspective of what's possible with the hardware will shift as these games have hopefully completed their "final-state" .
Like Patch 7 of BG3 states;
DD2 likewise has Capcom saying that they will bring patches in the future with CPU-specific optimisations;
But if you were to freeze the build of the game and let's say run it on theoretical T239 hardware, how would it do, I guess that's the question
. AAA 3rd party games are just a big ? at times.Moreover, if Elden Ring can run on the Switch 2, I'm actually not that satisfied with how it is on the PS5, because it still fluctuates in framerate and the RT mode is just a no-go, so let's say if the devs make odd decisions, the game might be running in a state that's less ideal compared to the steam deck.
Any idea of what the settings are here, in city its sub-30, but the IQ looks fine (not amazing, but I imagine on the SD you won't have much complaints).
/cdn.vox-cdn.com/uploads/chorus_asset/file/21696466/acastro_200804_1777_arm_0003.0.jpg)
