PCSX2 for Android.

[willkuer]

Clank, the beginning of your sentence didn't match the end of your sentence.

You could have a look at git where people try to port pcsx2 to Win-x64, Linux-x64 and OSX(-x64) and put a lot of hard work into it to make the simple step from x86->x64. I am not sure but I think pcsx2 is far from being portable and adding another endsystem is much harder than just porting the x86 recompiler to arm.

[Nobbs66]

It probably wouldn't be too hard to port it to Android, the hard part is making it work on another architecture. Most Android devices have an arm processor, which would require quite a bit of new code for it to run.

That would require retargetting the recompilers for ARM. Good luck with that.

[Stevethegreat]

Android usually runs on relatively weak hardware(smartphones/tablets) that would have trouble running most PCSX2 games. So even if it was easy, this is not something worth your time right now.

This argument again? I've been gone for ~2 years, but I remember -clearly- how people jumped on me with the exact same argument 2 years ago. I told them that ARM processors are making strides and by 2016 they'd be as fast as the PCs from late 00s ... Guess I was wrong. We're there already, both Shield Console and Galaxy S6 can play certain Gamecube/Wii games with very few issues. 

What's more by 2016 ARMs are going to "catch" i5s: http://sharetech.biz/wp-content/uploads/...ttttt1.png
The blowout that's coming to Intels is unbelievable and I very much doubt that there will be many x86 devices by next decade. PCs are a dying breed...

I got derided for saying what is basically common sense: Things change and we actually have an arrow showing where things are going. Most "smart" devices are (already) smartphones, and since there is the money, *there* will be where performance lies.

PCSX2 is a great project, I love it, I really do, but its inability to change drove me away. I rarely use my PCs, and I don't expect them to be used by 2020 at all, it means people like me  are going to be left behind. It's a bit sad because I had donated in the past and I had been a very enthusiastic proponent of it in its first years (to both friends and family).

Anyway I understand it's a monumental task to make the migration and I can see why the problem may well be logistical after all (just not enough people to make the move). What I don't get is the reasons that are used to say that "it is not worth it" (au contraire it's very much worth it).

[gregory]

Are you aware that benchmark run in turbo mode that is only sustainable for 5 minutes. Find a test that show cpu/gpu clock during a game session, you will be surprise. Anyway the project is still evolving.

[omnikam]

it will be amazing playing ps2 portable ....
but i don't think it will happen :/

What do you mean? Play! emulator is a Ps2 emulator that already works on Android devices, I was playing FFXII on my Note 3, not the fastest granted, but alot faster than Pcsx2`s android version

[Stevethegreat]

Are you aware that benchmark run in turbo mode that is only sustainable for 5 minutes. Find a test that show cpu/gpu clock during a game session, you will be surprise. Anyway the project is still evolving./

http://wccftech.com/snapdragon-exynos-at...benchmarks

The performance difference over a period of time is only 10% on CPU and about 20%  (at most) on GPU.  Since emulators are typically CPU starved a 10% difference from peak performance is not something to be worried about. Especially since ARMs gain about 40% of real performance year over year (by comparison Intels in desktop gained 40% over 4 years).

Antutu is the benchmarks mostly used to cheat results, apart from being a very bad benchmark anyway, if you noticed I haven't used it in "my" charts at all.

I'm running GNU/Linux on ARM tablets for years and I can tell you that we're nearly there already (performance wise). I have a big PS2 collection that I used to replay. But since even booting to my PC is a hassle these days I mostly don't. I know this is a community project but I hope you can see the "discrimination" against people who don't use legacy devices (more and more I'm getting untethered from Wintel as others do as well) and it will become even more obvious in the coming years.

There are other emulators sure, but not with the level of quality that PCSX provides. I'd be as enthusiastic (as I'm sure others will as well) if I/we would see an effort to support ARM devices. In both practical terms (donations) and less practical ones (evangelizing the project). This issue is certainly *not* something to be taken lightly (as I told you 2 years ago, and I fear I will tell you in 2 years from now).

Anyway I understand there are logistical issues (a small team), I merely wish to show that the reasoning is well founded.

[gregory]

Perf increase was due to
1/ frequency => they can't sustainable 3ghz+. Transistor leaks so much they need to duplicate the cpu for normal operations. They manage 2gz on 28nm (was it 2.3 or 2.5?). Yet where are the 3ghz cpu on 14 nm?
2/ very poor ipc => so it is easy to increase when you're close to 0. Maximum is 1.

It is true that perf is better, I won't deny it. However those benchmark are likely far from real workload. I remembered some graph where the galaxy note 4 uses the idling frequency of the gpu after 5-10 minutes. (-66% of the raw power).

So maybe a couple of small game might run in an acceptable speed but it is lots of work. Besides will people enjoy a native vga game without upscaling on their 4k smartphone?

Anyway it doesn't mean we are against any contribution to improve the portability of the code that one day will lead to an arm port.

[FlatOut]

This argument again? I've been gone for ~2 years, but I remember -clearly- how people jumped on me with the exact same argument 2 years ago. I told them that ARM processors are making strides and by 2016 they'd be as fast as the PCs from late 00s ... Guess I was wrong. We're there already, both Shield Console and Galaxy S6 can play certain Gamecube/Wii games with very few issues. 

What's more by 2016 ARMs are going to "catch" i5s: http://sharetech.biz/wp-content/uploads/...ttttt1.png
The blowout that's coming to Intels is unbelievable and I very much doubt that there will be many x86 devices by next decade. PCs are a dying breed...

I got derided for saying what is basically common sense: Things change and we actually have an arrow showing where things are going. Most "smart" devices are (already) smartphones, and since there is the money, *there* will be where performance lies.

PCSX2 is a great project, I love it, I really do, but its inability to change drove me away. I rarely use my PCs, and I don't expect them to be used by 2020 at all, it means people like me  are going to be left behind. It's a bit sad because I had donated in the past and I had been a very enthusiastic proponent of it in its first years (to both friends and family).

Anyway I understand it's a monumental task to make the migration and I can see why the problem may well be logistical after all (just not enough people to make the move). What I don't get is the reasons that are used to say that "it is not worth it" (au contraire it's very much worth it).

There are already smartphones and tablets running Windows 10 and Linux that do support PCSX2, so those users aren't necessarily left behind and PCSX2 is therefore already future proof if PC's would be left behind.

[Stevethegreat]

There are already smartphones and tablets running Windows 10 and Linux that do support PCSX2, so those users aren't necessarily left behind and PCSX2 is therefore already future proof if PC's would be left behind.

Those are a tiny minority (especially in phones). Intel SoCs have about half the perf/w when compared to ARMs using the same process. It's almost a suicide to use x86 chip, at least on a phone, it's a severe disadvantage when compared to others. For example compare the ~5w Tegra chip in the leaked chart above to core-m. Same/similar performance to a ~10w x86 SoC

[FlatOut]

Those are a tiny minority (especially in phones). Intel SoCs have about half the perf/w when compared to ARMs using the same process. It's almost a suicide to use x86 chip, at least on a phone, it's a severe disadvantage when compared to others. For example compare the ~5w Tegra chip in the leaked chart above to core-m. Same/similar performance to a ~10w x86 SoC

This is right now. In a year or two that will all change for the better, when both Windows and Linux majority and mobile processors will improve and it will be more worthwhile to try using PCSX2 on mobile devices.