67280 wl tr ef
65176 q
56680 rm
56680 cr_n2_sp_d2

i think any tr game never had game in background, in any version

This is very very useful dude! Thanks! I think the mystery of where the firmware goes is now solved.

robertmo3 wrote on 2024-05-08, 18:44:

i think any tr game never had game in background, in any version

I think @WhiteMagicRaven vids on YouTube had the game in the background while in game :

https://www.youtube.com/live/kbgDBQ5-UAY?si=t3kjATtTMxPJn0Uk

sharangad wrote on 2024-05-08, 18:45:

This is very very useful dude! Thanks! I think the mystery of where the firmware goes is now solved.

My bad. The mystery of where the framebuffer starts is now solved.

Just goes to how that I know next nothing about the card.

here are all dos4gw.exe we found
i don't think any of them has some secret verite code hiden inside 😉

they are here too
DOS Memory Extenders - by Snover and DosFreak

robertmo3 wrote on 2024-05-08, 18:52:

here are all dos4gw.exe we found
i don't think any of them has some secret verite code hiden inside 😉

If not dos4gw, the actual game binary. The uc file is for the verite board not for the exe. The exe reads the file and loads it onto the gpu and runs it.

This might be of interest to you

https://github.com/sarah-walker-pcem/veritas

Disassembler for Rendition Verite V1000/V2000 RISC processor

veritas <microcode.uc>

Disassembly is output to stdout

Does the map render with whiplash with the cr uc file?

I think a minor modification should allow everything to work properly irrespective of firmware version.

Thanks @robertmo3. I might not believe the uc file contains the game ovl/dll files but the fact it drastically affects performance and the framebuffer address gives me something to work work with. You could be right, even if I don't think so at the moment.

no map

robertmo3 wrote on 2024-05-08, 20:30:

no map

The view in the top left isn't the mirror, rather a mis-rendered frame.

All the buffer addresses are wrong, so double buffering doesn't work. I was very worried about the UC having such a massive effect. It's not. The addreses of textures/framebuffers are all wrong.

For instance with the CR UC with Whiplash, the front and back buffers are at:
V_FIFO_DISPLAY 0000E000
and
V_FIFO_DISPLAY 0008B000

With the original firmware:
V_FIFO_DISPLAY 00011000
and
V_FIFO_DISPLAY 0008E000

So the game literally POKEs the uc's program segment into VRAM and sets the frame buffer address to come after that.

V_FIFO_DISPLAY <address in hex> tells the Verite chip to set the visible part of the screen to start at address.

The reason you're getting better performance isn't because the UC file is working magic. What's happening is addresses are being miscalculated and the wrapper's doing less (or wrong) work, leading to a lower CPU load.

Ideally the wrapper should handle any uc firmware. It's not good enough to say it'll only work with the original firmware.

that is camera on opponent HAL 9TH in the tunnel
in the camera you see BOB 7TH and BISHOP 8TH too
you can see HAL 9TH on the main picture in the tunnel too
as well as GORT 10TH and VOYAGER 11TH
you are 12TH before the tunnel

press f4 or f3 more to get to the mirror

robertmo3 wrote on 2024-05-09, 04:30:

that is camera on opponent HAL 9TH in the tunnel in the camera you see BOB 7TH and BISHOP 8TH too you can see HAL 9TH on the mai […]
Show full quote

that is camera on opponent HAL 9TH in the tunnel
in the camera you see BOB 7TH and BISHOP 8TH too
you can see HAL 9TH on the main picture in the tunnel too
as well as GORT 10TH and VOYAGER 11TH
you are 12TH before the tunnel

press f4 or f3 more to get to the mirror

The point is if I adjust the framebuffer address in the wrapper to handler that framebuffer, the rendering will go back to what it was with the original firmware. Give me 30mins and I'll get you two binaries to do just that.

Try this version of dosclient (place it in your rlauncher folder and run dosbox from from rlauncher):

https://1drv.ms/u/c/5277ef8d4f2a9dcf/ERUZeXxs … aMOPnw?e=bByg0R
MD5: d948930d33a251b31e73560093afd002

Try that with Whiplash with your CR UC file. See if the mirrors render. Please don't try it with the original firmware, I have no idea what'll happen with it.

with cr uc it is the same as it was before with original (same poblems)
with original uc intro problems are gone

robertmo3 wrote on 2024-05-09, 05:32:

with cr uc it is the same as it was before with original (same poblems)
with original uc intro problems are gone

I'm working on fixing them right now. The UC file runs on the actual hardware and with RReady/DR only controls the frame buffer address and nothing else and even that's by virtue of occupying space in virtual VRAM.

mirror/camera is gone too 😉

robertmo3 wrote on 2024-05-09, 05:32:

with original uc intro problems are gone

less visible/heard