JAK II CRASHING

[Rogar]

With an Intel Mac and a Vega 56, it normally runs full speed at 4K resolution using Metal rendering.  A couple of specific spots in the city cause some minor slowdown (at night).  Also the entrance to the oracle causes a big slowdown when the door is open, but inside it's normal.  It does crash unless mipmapping is set to "basic".  As with Jak 1, I have EE cycle rate set to 130% which fixes a few places where an actual PS2 can't maintain 60fps, but that doesn't help the rendering slowdown in Jak 2. (And using software rendering doesn't help.)  99% of the game is fine but it's not flawless like the first one.

           

[Ugo88]

I can confirm that, after another couple of missions, I got slowdowns again, even using software rendering. 

IMHO there is some compatibility problem with the M1 chip, I started the game on my other PC (which has an Intel iGPU) and it runs fine at 1080p. I don't think the problem is computing power, the M1 (Pro in my case) is supposed to be a high-end chip.

[jesalvein]

I can confirm that, after another couple of missions, I got slowdowns again, even using software rendering. 

IMHO there is some compatibility problem with the M1 chip, I started the game on my other PC (which has an Intel iGPU) and it runs fine at 1080p. I don't think the problem is computing power, the M1 (Pro in my case) is supposed to be a high-end chip.

high end compared to what ?

[Ugo88]

From what Apple says, anything else (except the M2, of course...).

In my case, I'm comparing the M1 (Pro) to a cheap Intel Core i3.

[jesalvein]

From what Apple says, anything else (except the M2, of course...).

In my case, I'm comparing the M1 (Pro) to a cheap Intel Core i3.

There are still people who believe what Apple says ?

I should try to study religions a bit more, I guess

[Rogar]

That's not particularly helpful.  The Dolphin developers were pretty impressed a couple years ago or so when they made a M1 native version: "Compared to an absolute monstrosity of a desktop PC, it uses less than one-tenth of the energy while providing ~65% of the performance."  I've seen it mentioned elsewhere, as well as in this thread, that M1 Mac performance in PCSX2 is normally fine, aside from this particular game. There's some edge case here that's causing significant performance issues that aren't present in other games, and are also (mostly) not present in a desktop Intel Mac. Just from a technical standpoint it could be kind of interesting to know what's going on exactly. Dolphin isn't immune to this either: "It’s not all perfect though, as the team notes that “full MMU” GameCube and Wii games like the Rogue Squadron titles and Spider-Man 2 do suffer from some performance issues. This is seemingly evident in the first graph, showing Star Wars: Rogue Squadron 2 running better via Rosetta than via the native version of the emulator." If that's somehow related, it might be worth trying to run Jak II using the x86 version of PCSX2 via Rosetta rather than the native version.

[jesalvein]

That's not particularly helpful.  The Dolphin developers were pretty impressed a couple years ago or so when they made a M1 native version: "Compared to an absolute monstrosity of a desktop PC, it uses less than one-tenth of the energy while providing ~65% of the performance."  I've seen it mentioned elsewhere, as well as in this thread, that M1 Mac performance in PCSX2 is normally fine, aside from this particular game. There's some edge case here that's causing significant performance issues that aren't present in other games, and are also (mostly) not present in a desktop Intel Mac. Just from a technical standpoint it could be kind of interesting to know what's going on exactly. Dolphin isn't immune to this either: "It’s not all perfect though, as the team notes that “full MMU” GameCube and Wii games like the Rogue Squadron titles and Spider-Man 2 do suffer from some performance issues. This is seemingly evident in the first graph, showing Star Wars: Rogue Squadron 2 running better via Rosetta than via the native version of the emulator." If that's somehow related, it might be worth trying to run Jak II using the x86 version of PCSX2 via Rosetta rather than the native version.

ok. you're clearly comparing apples and oranges.

But at this rate, i'd suggest trying to use Aethersx2 instead

[Ugo88]

There are still people who believe what Apple says ?

I don't, actually. I do believe in benchmarks, though.

But at this rate, i'd suggest trying to use Aethersx2 instead

On a MacBook? 

[jesalvein]

I don't, actually. I do believe in benchmarks, though.

they rarely use pcsx2 as a base

On a MacBook? 

With an M1, yes