-Building and running on DOS
+Unnamed Mindlapse DOS demo for Pentium 133
+------------------------------------------
+The demo requires VESA Bios Extensions (VBE) 2.0. If your graphics card doesn't
+support VBE 2.0 or greater, then make sure to load the `univbe` TSR first.
+
+Source structure
+----------------
+ - src/ cross-platform demo framework and miscellaneous utility code
+ - src/scr/ demo screens (parts) and effects support code
+ - src/dos/ DOS platform code
+ - src/sdl/ SDL 1.x platform code (windows/UNIX version)
+ - libs/cgmath/ math library, header-file only
+ - libs/imago/ image loading library (includes libpng, zlib, libjpeg)
+ - libs/anim/ keyframe animation library
+
+Building on DOS with Watcom
---------------------------
+NOTE: Don't. Watcom produces significantly worse code than GCC, and at the
+moment watcom-compiled version of the demo crashes on 3D scenes for some reason
+which I need to investigate at some point. Suspect either inline assembly with
+missing "modify" part, or more likely some FPU optimization which fucks up the
+clipper.
+
Make sure you have Watcom or OpenWatcom installed, and the appropriate env-vars
set (the watcom installer automatically adds them to autoexec.bat by default).
Run wmake to build. Needs dos4gw.exe in current dir.
-The demo requires VESA Bios Extensions (VBE) 2.0. If your graphics card doesn't
-support VBE 2.0 or greater, then make sure to run the `univbe` TSR first, or
-the demo will fail to find a usable LFB video mode.
-
-
-Cross-compile on GNU/Linux
---------------------------
+OpenWatcom cross-compilation on UNIX
+------------------------------------
source owdev script with contents (change WATCOM var as necessary):
export WATCOM=$HOME/devel/ow
which will in turn start dosbox, which will execute the DOS binary! If the gods
are slumbering in valhalla, just typing `dosbox demo.exe` should do the trick.
+Building with DJGPP
+-------------------
+The DJGPP port of GCC is the recommended toolchain to use for building the demo,
+either natively or cross-compiled on UNIX.
+
+For native DOS builds add the DJGPP bin directory to the path (usually
+`c:\djgpp\bin`) and set the DJGPP environment variable to point to the
+`djgpp.env` file.
+
+For cross-compiling on UNIX simply source the `setenv` file which comes with
+DJGPP, which will set the `PATH` and `DJDIR` variables as necessary.
+
+In both cases, run `make -f Makefile.dj` to build. To run the resulting
+`demo.exe` you'll need to copy `cwsdpmi.exe` to the same directory. You can find
+it here: ftp://ftp.fu-berlin.de/pc/languages/djgpp/current/v2misc/csdpmi7b.zip
+
+When building natively on an old computer, and encounter a huge amount of disk
+swapping, and corresponding ridiculously long compile times, make sure to
+disable any memory managers and try again. `EMM386` may interfere with
+`CWSDPMI`'s ability to use more than 32mb of RAM, and modern versions of GCC
+need way more than that. Disable the memory manager with `emm386 off`, and
+verify the amount of usable RAM with `go32-v2`.
+
+Another problem is the MS-DOS version of `HIMEM.SYS` which only reports up to
+64mb. To use more than that, which is necessary for modern versions of GCC, you
+can either disable `HIMEM.SYS` completely, or use the `HIMEM.SYS` driver that
+comes with Windows 9x (DOS >= 7). Here's a copy in case you need it:
+http://mutantstargoat.com/~nuclear/tmp/d7himem.sys
+
Building and running with the SDL backend
-----------------------------------------
run the following in the demo root directory:
svn co svn://mutantstargoat.com/datadirs/dosdemo data
+
+Random optimization details about the Pentium1 (p54c)
+-----------------------------------------------------
+Use cround64 (util.h) for float -> integer conversions, instead of casts.
+
+Performance measurement with RDTSC:
+ perf_start();
+ /* code under test */
+ perf_end(); /* result in perf_interval_count */
+
+Cache organization (L1): 8kb data / 8kb instruction
+128 sets of 2 cache lines, 32 bytes per cache line.
+
+Addresses which are multiples of 4096 fall in the same set and can only have
+two of them in cache at any time.
+
+U/V pipe pairing rules:
+ - both instructions must be simple
+ - no read-after-write or write-after-write reg dependencies
+ - no displacement AND immediate in either instruction
+ - instr. with prefixes (except 0x0f) can only run on U pipe.
+ - prefixes are treated as separate 1-byte instructions (except 0x0f).
+ - branches can be paired if they are the second instr. of the pair only.
+
+Simple instructions are:
+ - mov reg, reg/mem/imm
+ - mov mem, reg/imm
+ - alu reg, reg/mem/imm (alu: add/sub/cmp/and/or/xor)
+ - alu mem, reg/imm
+ - inc reg/mem
+ - dec reg/mem
+ - push reg/mem
+ - pop reg
+ - lea reg,mem
+ - jmp/call/jcc near
+ - nop
+ - test reg,reg/mem
+ - test acc,imm
+
+U-only pairable instructions:
+ - adc, sbb
+ - shr, sar, shl, sal with immediate
+ - ror, rol, rcr, rcl with immediate=1
+
+Notes about DJGPP & CWSDPMI
+---------------------------
+Can't use the `hlt` instruction for waiting for interrupts, because we're
+running in ring3 by default. I surrounded all the `hlt` instructions with a
+`USE_HLT` conditional, which is undefined when building with DJGPP.
+
+It's possible to arrange for our code to run on ring0 by changing the DPMI
+provider from `cwsdpmi.exe` to `cwsdpr0.exe` by running:
+`stubedit demo.exe dpmi=cwsdpr0.exe`, but I haven't tested under win9x to see if
+it still works if we do that.
+
+Our fucking segments don't start at 0 ... to access arbitrary parts of physical
+memory we need to call `__djgpp_nearptr_enable()` and use the following macros I
+defined in `cdpmi.h`:
+
+ #define virt_to_phys(v) ((v) + __djgpp_base_address)
+ #define phys_to_virt(p) ((p) - __djgpp_base_address)
projects / dosdemo / blobdiff