forums.ps2dev.org

One challenge here is to properly support HDMA effects while getting optimal performance. Basically, most backgrounds and sprites can be blitted "straight up", ie. the whole tile or sprite in one fast blit. However, as soon as you use HDMA to dynamically alter PPU registers during the scre...

We already use the GU for blitting and scaling (not internally though, I don't see how that would help) By freeing the CPU from quite a lot of tedious pixelpushing (for sprites, backgrounds, overlays etc), speeding up the whole engine a couple of notches? Granted, I haven't looked at how ScummVM ha...

TommyBear: Some minutes into Fate of Atlantis I encountered a blue screen of death (!). Any other reports of that? It would seem it was displayed by the PSP kernel rather than the ScummVM engine, since it wasn't upscaled like all the ScummVM output. I also wonder if we can hope for a nice speed-boos...

Very nice. I suppose you use hardware blitting for all sprite and bg-plane drawing (not just the optional upscaling) to get it this fast?

It seems just a couple of cycles away from being able to run consistent with vsync enabled. Any further optimizations in mind? :)

If you really want to watch TV on the PSP it might be a better idea to look out for this.

That's cool, just equip Tarnished Amulet of Drundix (+50% damage vs undead) and you're sorted.

UAE requires a fairly hi spec PC to run decently. It would be interesting to see how it performs on PS2. At least you have all the videomodes (and a proper TV hook-up) needed to achieve perfect image quality.

Too true. :) But neither did the example in $PS2SDK/samples/hello ! I'm curious how printf() got declared there. By the way, after some further tinkering I link the following libraries at the end of the linker command: -lkernel -lstdc++ -lc -lgcc -lkernel I saw your comment that syscalls wasn't used...

After some tinkering I can build some C++ code, using both templates, new, etc. Printf() won't work though, with the exact same settings as when building the sample code (except the use of ee-g++ for compiling instead of ee-gcc). When building the example, gcc throwns a warning: hello.c:27: warning:...

OK. Using the makefile included with ps2sdk (Makefile.eeglobal_cpp) the linker command expands to this: ee-gcc -T/usr/local/ps2dev/ps2sdk/ee/startup/linkfile L/usr/local/ps2dev/ps2sdk/ee/lib -o test.elf test.o -lc -lkernel -lstdc++ -lc -lkernel -lsyscall Which won't resolve the reference to strlen. ...

I just tried to compile some basic C++ code and ran into the same problem: undefined reference to `strlen'.

Is rolling up and link an own strlen.o the best solution, or is there a cleaner way?

(btw, I just cooked up a fresh ps2dev environment using ooPo's script)

Looking forward to it!

Incidentally, I once had a pile of 48x Memorex discs that absolutely refused to work on my PS2.

Allright, I have the packages here now. I suppose your *bigger* bandwidth settles the question of where to put them? :) Some more testing needs to be done, or the purpose of the packages -- minimizing the amount of cygwin build questions on the forum -- is defeated.. I really don't have any code to ...

I've been flicking through the VU-compo entries from the PS2Linux community (which incidentally got me interested in writing some 16k of VU code myself). But! I also was a bit disappointed to see that some of the more impressive offerings ("Universe" was particularly groovy) was optimized ...

There is no "situation" - licensed developers that have legal access to Sony's GCC are free to distribute those changes under the GPL. No one has done it yet. So, in effect, any SCEx-developed (or by any other party) patch for the GNU tools are okay to spread -- provided you have the mean...

I believe the first patches originated from Cygnus (Red Hat), who wrote them for SN Systems' original PS2 SDK's in 1999. It would seem that Dylan Cuthbert's VU0 patches also filtered down to the PS2dev community, according to this: http://ps2dev.sourceforge.net/hypermail-past/0009.html So I reckon t...

Yes, but there are also modifications done to GCC that adds inline support for VU0-code (referred to as "Dylan's VU0 support patches"). These are mentioned in the following project at PS2linux: http://playstation2-linux.com/project/download_single_file.php/5_125 Some details here: http://p...

Fresh! Will test these under Cygwin during the next couple of days. On a related note, what is "the situation" with SCE-developed binutils/gcc patches? For example, the support for VU0 in their EE toolchain is rather nifty, and as far as I know we don't have it in our compilers (correct me...

OK, I'll roll some packages during the weekend and send them over. What do you think about having only one EE package (binutils 2.13.2.1, gcc 3.2.2 with c++ support including newlib + the old dvp binutils), one IOP package (binutils 2.9.1, gcc 2.8.1) and a ps2lib 2.1 package?

ooPo: Worked like a charm, thanks! I'll look into making a ps2dev installation package for Cygwin during the weekend (no such exist, right?).

I also ran into the same problem trying to build the IOP toolchain. I tried to touch configure after patching, like this: tar xfvz binutils-2.9.1.tar.gz cd binutils-2.9.1 cat ../ps2dev-iop-binutils-20020804.patch | patch -p1 touch configure ./configure --target=iop --prefix=$PS2DEV/iop make ...make ...