• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Windows 95 running on a New Nintendo 3DS

LordOfChaos

Member
Does this also mean we have more detailed information on it's internal spec update now? I was interested in that a while ago but the exact updates seemed dodgy. I heard all of either a twice as high clocked dual core, same clocked quad core, or two of the old CPU cores and two with a new architecture entirely? The last being based on hints from an insider on neogaf itself, not sure how reliable.
 

LordOfChaos

Member
that's right
PSM only give 400 mhz of the vita processor speed ( maybe 1 of the 4 cores only ? )

it's not complete , and it's very complicated.

edit: i think i got some wrong infos , the CPU of the Vita is clocked to 333 or 2 GHz ?

the tests by hackers says it's 333 , but maybe because they are running from Rejuvenate ?



It runs at 333MHz regularly with a max of 444. You don't divide core count by clock speed for per-core clock speed, clock speed is a maximum across all cores, all four will run at that 333MHz if needed. The 2GHz max was only a shot in the dark based on what the Cortex A9 could be clocked at; even in the iPhone 4 it was only at 800MHz though, 2GHz in the Vita form factor was never likely with it's meatier GPU needing more of it's own juice, 4 cores to run, etc. The GPU iirc was something like an iPad 3-4?

http://bitparade.co.uk/article.php?id=3859
 

onken

Member
I had no idea 3DS had a homebrew scene beyond those hacked memory cards with a bunch of pirated games on them. What's the best resource for finding out what's available?
 

2+2=5

The Amiga Brotherhood
Don't confuse generic PSM titles for PSM-for-vita titles. Unity-for-PSM was practically on-par with native Unity feature- and performance-wise, but lacked native debugging and profiling support and allowed only C# in a VM (which did pose a performance hit on its own).
Well yeah i forgot about that, but even those had big limits, in fact PSM unity games used to run really bad, and they were ugly too.

That's not how processors work... Each processor has its own frecuency and they don't sum.
You are right lol.
 

blu

Wants the largest console games publisher to avoid Nintendo's platforms.
Well yeah i forgot about that, but even those had big limits, in fact PSM unity games used to run really bad, and they were ugly too.
That's irrelevant to the subject. Read p. 21 of the slides I linked a few posts back.
 

2+2=5

The Amiga Brotherhood
Don't confuse generic PSM titles for PSM-for-vita titles. Unity-for-PSM was practically on-par with native Unity feature- and performance-wise, but lacked native debugging and profiling support and allowed only C# in a VM (which did pose a performance hit on its own).

From the slide 23 it seems that unity under PSM runs even better than unity natively on vita and that didn't seem to be the case at all.

In any case unity or not we started all this discussion because of the SNES emulation performance on vita, unless an emulator isn't written natively on vita we can't make comparisons with the N3ds(but imo vita would clearly beat it).
 

blu

Wants the largest console games publisher to avoid Nintendo's platforms.
From the slide 23 it seems that unity under PSM runs even better than unity natively on vita and that didn't seem to be the case at all.
How did you reach that conclusion? Here's the page in question:

Code:
Performance of Unity for PSM
• PSM === PS Vita, except
    • Just-In-Time compilation of scripts to native code, “on demand”
        • When a method is accessed
    • Runtime compiled shaders
        • When loading a level (currently)
            • But cached - subsequent runs will be (much) faster
• Any optimizations done for PS Vita will benefit PSM automatically


In any case unity or not we started all this discussion because of the SNES emulation performance on vita, unless an emulator isn't written natively on vita we can't make comparisons with the N3ds(but imo vita would clearly beat it).
The discussion started with the premise that PSM somehow did not expose full CPU functionality or performance to Vita-PSM compared to native vita apps. I brought up Unity-PSM as a counter-argument.

The one disadvantage of Unity-for-PSM was in the mandatory managed C#, shaders and scripts, which got JIT'd to native only upon execution. Yes, that could pose a performance hit to some types of code. No, that was not because the OS somehow throttled the CPU or GPU for PSM apps. Your original assumption was wrong.

There never was a "true native (tm)" vita mode where apps were given access to a fabled 2GHz CPU. The quoted 444MHz is the most a vita app could get out of a single core, and that was only under certain concessions; 333MHz was the nominal CPU speed.
 

entremet

Member
Pretty neat.



Krejlooc

Dude is very knowledgable, but he just doesn't know how to let go of a minor argument lol.

He always get banned for that, which is a shame since he contributes a lot of good info.

Hopefully, some self awareness is in order.
 

c0de

Member
There never was a "true native (tm)" vita mode where apps were given access to a fabled 2GHz CPU. The quoted 444MHz is the most a vita app could get out of a single core, and that was only under certain concessions; 333MHz was the nominal CPU speed.

Huh? That would be some big limitation for the Vita hardware, in my opinion, and would mean that it got outpowered by smartphones sooner than I thought.
 
Apparently, the Win95 build now runs Dosbox and has mouse support
EWmof.gif
 

c0de

Member
Apparently, the Win95 build now runs Dosbox and has mouse support
EWmof.gif

Huh? So it runs dosbox in dosbox? Wow, hackers nowadays! I think they should try to run Windows 95 in the dosbox that runs dosbox in the Windows 95. And then they could try to run dosbox! Endless possibilities.
 
Huh? That would be some big limitation for the Vita hardware, in my opinion, and would mean that it got outpowered by smartphones sooner than I thought.

Mm, it was cutting edge (on par with what Apple was shipping in the iPad 3) on release, but by the time the iPhone 5, iPad 4 and similar devices with Snapdragon S4 Plus/Pro were released in 2012, it was surpassed and then some.

The Swift cores in the Apple A6 SoC and Krait in S4 Pro/Plus were a whole generation ahead of the Cortex-A9 in Vita/iPhone 4s/iPad 3.

And then Apple shipped the first SoC with ARM v8 in 2013 with the iPhone 5s' Apple A7, which also had PowerVR Series 6 graphics -- a whole generational leap above Series 5 in Vita.

This shoudl give a good idea of how far things have come since late 2011:

charts.001-980x720.png


iPod Touch 5 = ARM Cortex-A9 @800MHz, 2011.
iPhone 5s = Apple Cyclone, 2013
iPod Touch 6 = Apple Cyclone Plus @1Ghz, 2014

charts.0031-980x720.png


Source: http://arstechnica.co.uk/apple/2015...e-the-sixth-generation-ipod-touch-reviewed/2/
 

c0de

Member
Mm, it was cutting edge (on par with what Apple was shipping in the iPad 3) on release, but by the time the iPhone 5, iPad 4 and similar devices with Snapdragon S4 Plus/Pro were released in 2012, it was surpassed and then some.

The Swift cores in the Apple A6 SoC and Krait in S4 Pro/Plus were a whole generation ahead of the Cortex-A9 in Vita/iPhone 4s/iPad 3.

And then Apple shipped the first SoC with ARM v8 in 2013 with the iPhone 5s' Apple A7, which also had PowerVR Series 6 graphics -- a whole generational leap above Series 5 in Vita.

This shoudl give a good idea of how far things have come since late 2011:

charts.001-980x720.png


iPod Touch 5 = ARM Cortex-A9 @800MHz, 2011.
iPhone 5s = Apple Cyclone, 2013
iPod Touch 6 = Apple Cyclone Plus @1Ghz, 2014

charts.0031-980x720.png


Source: http://arstechnica.co.uk/apple/2015...e-the-sixth-generation-ipod-touch-reviewed/2/

The core itself is definitely beefy but the clock rate cripples the CPU in a way I didn't expect.
 
Huh? So it runs dosbox in dosbox? Wow, hackers nowadays! I think they should try to run Windows 95 in the dosbox that runs dosbox in the Windows 95. And then they could try to run dosbox! Endless possibilities.

Haha, sorry, I worded that a bit bad. It runs Windows 95 through Dosbox (through Retroarch), which now has added mouse support (I guess to the c-stick).
 

Syriel

Member
I figured there was a homebrew of Doom on 3DS. I see Doom as a means test to any form of technological screwing about. I want Doom inside Doom inside windows 95 inside a 3DS emulator on Windows 10 inside OSX via Wine inside a Linux virtual machine.

You haven't run Doom until you've run psDoom on a Linux box.

Where else can you kill processes by shooting them?
 

2+2=5

The Amiga Brotherhood
How did you reach that conclusion? Here's the page in question:

Code:
Performance of Unity for PSM
• PSM === PS Vita, except
    • Just-In-Time compilation of scripts to native code, “on demand”
        • When a method is accessed
    • Runtime compiled shaders
        • When loading a level (currently)
            • But cached - [B]subsequent runs will be (much) faster[/B]
• Any optimizations done for PS Vita will benefit PSM automatically



The discussion started with the premise that PSM somehow did not expose full CPU functionality or performance to Vita-PSM compared to native vita apps. I brought up Unity-PSM as a counter-argument.

The one disadvantage of Unity-for-PSM was in the mandatory managed C#, shaders and scripts, which got JIT'd to native only upon execution. Yes, that could pose a performance hit to some types of code. No, that was not because the OS somehow throttled the CPU or GPU for PSM apps. Your original assumption was wrong.

There never was a "true native (tm)" vita mode where apps were given access to a fabled 2GHz CPU. The quoted 444MHz is the most a vita app could get out of a single core, and that was only under certain concessions; 333MHz was the nominal CPU speed.

I said page 23, not 21(where there's clearly written that shaders run much faster after the first run) ;)
At page 23 you can see a comparison between times needed to do various operations and you can clearly see that most of them are lower in unity for PSM, sorry but at this point i don't trust what's written.

As for performances a game written for an engine that run on multiple hardwares can never push the hardware and be as optimized as a game written specifically for the hardware, that was what i intended when i was talking about native software, it wasn't related to vita's max cpu speed at all.

In any case i doubt that unity for psm gave full access to vita hardware because in that case hackers would have already found a way to fully hack vita imo.
 

blu

Wants the largest console games publisher to avoid Nintendo's platforms.
I said page 23
Ah, let's see..

.. not 21(where there's clearly written that shaders run much faster after the first run) ;)
Of course they do, that's how JIT with caching works. What did you expect?

At page 23 you can see a comparison between times needed to do various operations and you can clearly see that most of them are lower in unity for PSM, sorry but at this point i don't trust what's written.
Erm, that's two separate profiling sessions - no two profiling sessions can show the exact same results. We're talking fractions of milliseconds here - without knowing the profiled tasks in intimate details, that very well could be standard variance for this code. What the picture shows is not that Unity-PSM is faster than Unity-vita (and that was never claimed by the authors), but that the two have virtually identical performance, where it comes to the engine. BTW, JIT'd code can be faster than statically-built code, it's the JITting itself that is problematic (if one does not account for it, which is not always possible).

As for performances a game written for an engine that run on multiple hardwares can never push the hardware and be as optimized as a game written specifically for the hardware that was what i intended when i was talking about native software, it wasn't related to vita's max cpu speed at all.
The bold is a blanket statement. Depends entirely on the engine and the application.

In any case i doubt that unity for psm gave full access to vita hardware because in that case hackers would have already found a way to fully hack vita imo.
Obtaining full access to the performance resources of the machine does not equate obtaining full privileges. Apparently you cannot run pre-built binaries via Unity-PSM, as PSM apps are essentially sanboxed. Unity engineers claim the performance delta for their PSM-based engine is essentially nil. Of course, you can believe whatever you like.
 

Peltz

Member
Cool feat. This got me thinking....

I feel like companies are missing out by not porting old PC titles to handhelds. We really SHOULD be able to play games like StarCraft, Diablo, and Doom on the go with 3DS and/or Vita.

It's kind of a missed opportunity. Or maybe I'm just crazy.
 

LordOfChaos

Member
Cool feat. This got me thinking....

I feel like companies are missing out by not porting old PC titles to handhelds. We really SHOULD be able to play games like StarCraft, Diablo, and Doom on the go with 3DS and/or Vita.

It's kind of a missed opportunity. Or maybe I'm just crazy.

Then again a $99 dollar Windows 8/10 tablet does that with aplomb. See the Stream 7 etc.
 
Top Bottom