Makaron Test 11.1 - 模擬器

http://dknute.livejournal.com/

This time the package includes NAOMI emulator as well. And I've modified the
code a bit, let's hope this will make the Data Execution Prevention exception
go away.
Both executables can now reside in the same directory, but because of that
there are some minor changes:
* Dreamcast BIOS/FLASH filenames must now start with Dreamcast_ prefix.
* NAOMI BIOS filenames must start with NAOMI_ prefix.
* The main configuration file for NAOMI emulator is now called NAOMI.ini

And yes, you still need to extract and decrypt any NAOMI games you wish to
run from the GDI image. It's also possible to extract them from CHD files.

Remember to install the runtime libraries for MSVC 2008 SP1 if you get those
silly "not installed correctly" errors.
http://s-url.jp?23597

UPDATE: It seems I forgot to put the newly compiled Makaron executable into
the package. Sorry :)
Redownload, I've updated the link. Also, if you add "aniso = 0" to the
Settings section it should disable anisotropic filtering. Not really tested


17th-Nov-2008

Still busy. Good news is I′ve been forced to add FEG emulation to AICA
recently and now I dare say it′s finally complete. Filtering isn′t really
used much but Shenmue needed it for the in-game music.
Since I′m still in the woods with other changes I suppose I could release
another test version of Makaron with what I have now. Not much of that I′m
afraid:

- Fixes to TA I′ve already wrote about (ie Excelica)
- Revised VMU plugin, it is now possible to save in WinCE games
- FEG/digital lowpass filter emulation in AICA

Stay tuned.

Also, some random pictures. If you know what it is you can start drooling now
This will take time mind you, but maybe I′ll have something working before
2009. Still waiting for the Altera board to arrive...


UPDATE: Makaron T11/1 is out.
You can grab it from RapidShare. It seems they now have a download limit
system - so in case it kicks in you can try SendSpace as an alternative.

Now, pay attention:
* It′s only the Dreamcast version for now.
* It has been compiled with Microsoft Visual C++ so you might need to install
the runtime libraries to get it to work. It′s only 4MB and you can download
it from Microsoft.
* Hit F8 to exit the emulator at any time
* You can now use Pause/Break key to pause/resume the emulation

Minimum hardware requirements have also changed:
* Processor must have SSE support
* Graphics card must be at least Shader Model 2.0 compatible

By not having to support ancient Pentium 2s and Durons, and Radeon 8500, I
can make the code simpler and maybe a bit faster too. That old hardware wasn′
t going to run Makaron at any acceptable speeds anyway :)

14th-Oct-2008

I′ve moved Makaron project to Microsoft Visual C++ 2008. Express edition for
now as it′s free and has about all the tools I need (resource editor would
be nice though). Too bad I won′t be using a profiler anytime soon - it′s
only present in the Team suite and that′s a bit out of my price range Maybe
I could use CodeAnalyst, though on Intel hardware it can only do timer-based
analysis. But hey, this is still about as good as gprof and it′s also free,
whereas VTune prices start at 700$ - and that′s just a licence for the basic
module.
You know, it′s funny but I also liked GPU ShaderAnalyzer better than NVidia
tools. It′s not as sophisticated but also easier to install and use. Get a
hint, Intel and NVidia :)

There weren′t too many problems with moving to MSVC. Mostly I had to switch
to different alignment and structure packing semantics. That and some inline
assembly had to give way to more generic C code - it was only worthwhile when
using AT&T syntax with clobbered register hints. The YUV converter is now
using MMX/SSE2 intrinsics (as it was meant to but I had no choice but to use
assembly under MinGW 3.4.5).
All in all, it compiles and works. I′m still using controller and keyboard
plugins done with MinGW but that will soon change as well.

I′ve also moved to Unicode - that was done even before I switched compilers
though. The INI files can now be encoded as ANSI (that′s ASCII plus current
Windows code page character set), UTF-8 or (preffered) UTF-16. Little endian
of course, as all x86 architecture is. The format will be autodetected based
on the BOM and so Unicode files must start with one. Notepad will take care
of that for you.

So far there are only two downsides to MSVC: it comes with standard STL
headers only (so no slist for example) and running the executable requires
one to download and install a redistributable package from Microsoft. The
latter is only 4MB so it shouldn′t be a problem.
As for STL, I′ve been experimenting with it for some time now. My own
classes are not that universal and not really faster either. I did port some
of that code back to plain C to use on Dreamcast, but that doesn′t mean I
need to keep Makaron STL-free forever. There wasn′t much point in switching
to STL under MinGW, but I did anyway, and I got rewared for that in MSVC. It′
s debug build helped me find some deeply buried bugs in the way I handled
texture lists. Nothing major but still.

What else is new? The YUV->RGB texture conversion performance on NVidia cards
should be a bit higher now. It also turns out using PUREDEVICE is not the
best idea considering how many state changes I do per frame (even if I do try
to limit the most costly ones).

Oh, and thanks to new compiler Makaron is now a tad bit faster. Not much, not
always, but I′m getting up to 120 MIPS out of SH4 in full MMU mode where I
was getting 90 before. Unfortunately the slower your CPU is the less boost
you will see - but still, that′s an improvement.

--