vwemirror/TeslaRel410
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Correct §06: the emit map is the shared standard screen program

Browse Source 
Cross-checks against the vendor source settle the semantics: the 0x3xx words
are pixel-memory bit addresses / opcode fields (EOF.S emits them as hardcoded
constants; DIVPXMAP.H maps dvpx_* bit addresses 0-207 incl. eofr/g/b at
184/192/200), 0xec00 is a hardcoded control constant (EOF.S:2313), and the
whole 2,541-word payload set is 98.98% identical between trek and klngvid =
the standard screen program shared across captures. §06 now presents the
render as an emit-stream structural map with the demo-specific geometry still
to be located (beyond the 0x8014000-0x8018000 window; bin pages / far heap are
the candidates). The EOF.S<->EOF.C correlation method for the exact per-macro
decode is proven (send_em catalog: 30 emit sites with inline constants).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Cyd
2026-07-16 21:40:34 -05:00
co-authored by Claude Opus 4.8
parent c2db52e69b
commit e3897a8e28
1 changed files with 15 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
emulator/firmware-decomp/render-readout.html
+15 -10
View File
@@ -362,20 +362,25 @@
</section>
<section>
<h2><span class="idx">06</span> First frame from a live scene — the Star Trek demo</h2>
<h2><span class="idx">06</span> A live scene emits — the Star Trek demo</h2>
<p class="sub">The archive's <span class="mono">trek</span> capture — unreleased Star Trek
material — now <b>replays and draws</b>. After the dual-instruction-mode fix, its full
pipeline runs on the emulated i860: classify, LOD select, rasterise, and <b>emit compiled
micro-code</b> per 64×128 tile. Below is that frame, assembled by reading the emit stream
back out of DRAM: <b>21 payload blocks, 124 scanline spans</b>, each block a primitive of
the scene. This is a <b>first-approximation decode</b> (the span/edge field scaling is
still being pinned), drawn from the board's own compiled render program — the deepest
read-out of this hardware yet.</p>
micro-code</b> per 64×128 tile. Below is a <b>structural map of that emit stream</b>, read
back out of DRAM: <b>21 payload blocks</b>, each a compiled primitive program — colors
distinguish blocks, position reflects each program's instruction fields. It is <b>not yet
the screen image</b>: cross-checking against the board's own source
(<span class="mono">EOF.S</span> / <span class="mono">DIVPXMAP.H</span>) shows the fields are
pixel-memory bit addresses and coefficients, whose exact per-macro layout is the decode in
progress — and this block set turns out to be the <b>standard screen program</b> shared
across captures (98.98% identical between the trek and Klingon demos), with the demo-specific
geometry still to be located in memory. What it proves stands: <b>the full pipeline runs, the
emit stream is readable, and the compiled programs are being captured one by one.</b></p>
<div class="hero-render" style="margin-top:16px">
<span class="tag">decoded from the compiled emit stream · first approximation</span>
<canvas id="trekframe" width="832" height="512" aria-label="frame assembled from trek's compiled emit stream"></canvas>
<div class="cap"><b>trek, draw 1–300 (static showcase scene)</b> — every mark is a span
the firmware's rasteriser emitted; colors distinguish payload blocks (primitives).</div>
<span class="tag">emit-stream structure · exact frame decode in progress</span>
<canvas id="trekframe" width="832" height="512" aria-label="structural map of trek's compiled emit stream"></canvas>
<div class="cap"><b>trek, draws 1–300 (compile-once showcase scene)</b> — each colored
run is one compiled primitive program's field pattern in the emit stream.</div>
</div>
</section>