Well, the best advice I can give you by this point is to double check your wiring and make sure they are all correct.

You're missing a few handshake lines. I think you need all handshake lines for it to work properly as some games likely need it.

Here's a reference to wiring a null modem cable with full handshaking: https://www.lammertbies.nl/comm/info/RS ... .html#full

No, don't connect the 3.3v line as that would likely fry something on the console.

Have you followed the standard null modem cable wiring when making the serial cable? (ie. tx -> rx, rx ->tx)

I never had any issues working with code compiled using GCC 7.2.0 on the PS1 so I don't exactly know what you're talking about. The compiler's register usage adheres to the R3000 specification and I've had no issue interacting with hardware registers (though it might be because I do such operations ...

These are intended to be used with open source libraries. I haven't bothered compiling any of the existing ones with these builds of the GNU toolchain since I write my own libraries instead (yeah, I'm that hardcore) as the existing ones don't really fit my standards very well but I occasionally use ...

Just added downloads of the toolchain compiled for various hosts including linux-arm (Raspberry Pi). I also added a download of the libgcc library since it is apparently mandatory for things like int64 emulation and for whatever reason it will never compile under Windows.

You can find the pinout right here: http://hitmen.c02.at/html/psx_siocable.html I built my own link cable by using the same wiring as a standard null modem cable but adjusted to go in line with the PlayStation's serial pinout. Handshake lines must be wired accordingly in case of games that use hands...

Ah, you need to Z sort your primitives using the OTz result. DivideF4( &vp[0], &vp[1], &vp[3], &vp[2], color[i], obj, (u_long *)myOT[ActiveBuffer].org+OTc, &div); and... AddPrim((u_long *)myOT[ActiveBuffer].org+OTc, obj); That should fix your Z sorting issues. Also, you may want ...

Ah, looks like things aren't z-sorting right. Nothing to do with matrices at least.

How do you sort TMDs into the ordering table? Also, I see my old lightbulb model .

That should do it for the first person camera. I'm not sure how libgs would treat it though as I haven't used that library in a long time as I usually use plain libgte+libgpu for graphics.

I recently figured out the right parameters to compile the toolchain with absolutely no errors. I replaced the download with a new version of the archive containing an error free build of the toolchain along with being configured with --with-gnu-as and --with-gnu-ld options which fixes include paths...

That's just my preferred naming convention. You can name it whatever you want really.

I usually just have a pair of vectors for both the camera position and angle and then applying it to a matrix using the first person view matrix logic from my examples featuring a first person style camera. Putting the relevant code here for easy reference: VECTOR cam_pos; // Camera position SVECTOR...

I did try to build GCC to output ECOFF (using mipsel-unknown-ecoff) a few days ago but sadly, the toolchain no longer supports it anymore. To those who want to try to build a toolchain yourself, make sure you have MinGW and msys installed and referenced in your PATH environment variable. Create a bu...

You may want to take a look at this example I made a long time ago that demonstrates how to do 3D transformations with the GTE directly using macros. It is much faster than RotTransPersp3() and gives you more control on how the GTE is being utilized. You'll need to use dmpsx to convert the weird wor...

Here's a download of the GNU GCC 7.2.0 Toolchain and Binutils 2.29.1 compiled using MinGW32 for 32-bit Windows hosts. I decided to put this up when I managed to compile the toolchain under Windows some time ago which took me quite awhile to figure out even with msys and cygwin64 environments for con...

Yeah, TMD is not exactly suitable for rendering large 3D environments so a serious developer would often make their own rendering system. More sophisticated renderers have stuff like mipmaps and geometry LOD as seen in the Spyro games. The TMD rendering functions are also pretty bloat and have some ...

MKPSXISO should be outputting error messages if the license file specified is either not found or does not match a specific file size. Mednafen's likely doing some checks in some really obscure parts of the ISO file system header which my tool isn't replicating. Have you tested the ISO on ePSXe, pSX...

That's sadly where tech seems to be heading... Bloat is the future.

I could've used malloc and realloc instead of std::vector but I made this as a quick little tool mostly.

I wrote this little tool when I experimented with creating my own SDK (called PSn00bSDK) using the latest GCC toolchain to produce PlayStation executables as PSXSDK's elf2exe never seemed to work for me. Plus it didn't really look like a good converter to me judging by its source code which simply r...