summaryrefslogtreecommitdiffstats
path: root/wiki/Execution-Frontier.md
diff options
context:
space:
mode:
Diffstat (limited to 'wiki/Execution-Frontier.md')
-rw-r--r--wiki/Execution-Frontier.md77
1 files changed, 45 insertions, 32 deletions
diff --git a/wiki/Execution-Frontier.md b/wiki/Execution-Frontier.md
index bfecad9..950902a 100644
--- a/wiki/Execution-Frontier.md
+++ b/wiki/Execution-Frontier.md
@@ -14,7 +14,7 @@
> [`../docs/04-implementation/03-execution-frontier.md`](../docs/04-implementation/03-execution-frontier.md);
> the docs tree wins on disagreement.
-Status: 2026-11-16, tree `master @ a06d4b13`.
+Status: 2026-11-17, tree `master @ 1db3831e`.
Tags: `[live]` observed; `[wip]` partial; `[wall]` current blocker; theories
are labelled as such.
@@ -144,19 +144,28 @@ chunks, a 128-slot candidate ring), the beef draw lane latches a **clean DS**
capture-less boots get a **golden 2799-word MME pack**, and
`MME-CALL-HOLD -> MME-CARRY-SAVE -> MME-CARRY-FIRE -> MME-PACK-REARM ->
MME-SA-REARM` fires with the prelude latch kept across the channel recycle.
-The **decisive finding is guest-side**: across 36 boots (c990-c1019) the
-client never submitted a non-null `CALL(34)` draw macro — every
-`0x3910/0x3914` hit is an interior method write with `data=0x0`, no CALL
-header exists in the admitted windows, and those cycles logged guest
-EGL/dri2 failures and OOM kills. The next work is guest-first (repair and
-verify the guest GL/EGL path and memory, then capture one real frame
-submission), followed by the vertex/store and the window/screen present
-path; the draw still does not store on the vGPU path (`wrote=0`, `0/1025`,
-`px0=0`), glxgears' window is black, a synchronous window draw can wedge X,
-and the per-boot X first-generation start lottery is the dominant throughput
-limiter. A GZDoom video was separately delivered through the guest's
-**software GL** path (Mesa llvmpipe) as a guaranteed-quality deliverable; it
-is not a vGPU rendering claim. The guest Vulkan
+The **late-11-16 wave crossed the guest submission gap**: an offscreen SDL
+EGL/GBM shim gives GZDoom a real NVIDIA GL context without Xorg/DDX
+(`GL_RENDERER: NVIDIA GeForce RTX 2080 SUPER/PCIe/SSE2`, `GL_VERSION: 4.6.0
+NVIDIA 570.144`; no llvmpipe), and the real draw window surfaced — 12068 B
+at push `0x402f9d8`, an op=4 IMMD `CALL(0)` to method `0x3800` (`imm=0xd48`)
+plus slots 44/57/58 (`0x3960/0x39c8/0x39d0`), with no `0x3910/0x3914` and no
+colour-target bind of its own. Recognition covers the real marker, admission
+happens on the marker even without a bind, the CALL-bearing chunk is forced
+first and every chunk replays verbatim. `CALL(0)` no longer wedges: the
+immediate is resolved as an IR address through client SA[0] (then golden)
+and a not-resident call is elided to a NOP; c1026/c1027 entered the captured
+macro without trapping and c1027 produced the first DS with real stage data
+(`DS-FAST2 ds_after=83044206 stage=015e0900,015e0a00,015c0400`). The
+remaining gate is the client's **per-draw MME macro body**, never captured
+(the 59580-byte prelude carries only the 12 setup macros; the draw push has
+no `0x114/0x11c`), with the synthetic golden body the surviving wedge
+(`badf1201 pc=513 mac=0xf1`). The draw still does not store on the vGPU path
+(`wrote=0`, `0/1025`, `px0=0`), glxgears' window is black, a synchronous
+window draw can wedge X, and the per-boot X first-generation start lottery
+is the dominant throughput limiter. A GZDoom video was separately delivered
+through the guest's **software GL** path (Mesa llvmpipe) as a
+guaranteed-quality deliverable; it is not a vGPU rendering claim. The guest Vulkan
pipeline's inline-to-memory upload arm is its gap. The full picture lives on
[`Graphics-Execution.md`](Graphics-Execution.md).
@@ -231,16 +240,20 @@ the beef draw lane latches a **clean DS** (`DS-FAST2 ... ds_after=80000000`,
c963) at the live-derived GS attribute map, capture-less boots get a
**golden 2799-word MME pack**, and
`MME-CALL-HOLD -> MME-CARRY-FIRE -> MME-PACK-REARM -> MME-SA-REARM` fires
-with the prelude latch kept across the channel recycle. The **decisive
-finding is guest-side**: across 36 boots (c990-c1019) the client never
-submitted a non-null `CALL(34)` draw macro — every `0x3910/0x3914` hit is an
-interior method write with `data=0x0`, no CALL header exists in the admitted
-windows, and the same cycles logged guest EGL/dri2 failures and OOM kills.
-A GZDoom video was separately delivered through the guest's **software GL**
-path (Mesa llvmpipe) as a guaranteed-quality deliverable; it is not a vGPU
-rendering claim. The next work is guest-first (a real draw submission), then
-the vertex/store, the window/screen present, and the per-boot X
-first-generation start lottery
+with the prelude latch kept across the channel recycle. The **late-11-16
+wave crossed the guest submission gap**: an offscreen EGL/GBM shim gives
+GZDoom a real NVIDIA GL context without Xorg/DDX, the real op=4 `CALL(0)`
+`0x3800` draw is admitted and replayed (12068 B at push `0x402f9d8`), and
+`CALL(0)` executes its captured macro without trapping (eng89t resolves the
+immediate through client SA[0]/golden and elides not-resident calls to
+NOPs); c1027 reached the first DS with real stage data. The remaining gate
+is the client's per-draw MME macro body, never captured (the prelude carries
+only the 12 setup macros; the draw push has no `0x114/0x11c`), with the
+synthetic golden body the surviving wedge (`badf1201`). A GZDoom video was
+separately delivered through the guest's **software GL** path (Mesa
+llvmpipe) as a guaranteed-quality deliverable; it is not a vGPU rendering
+claim. The next work is the per-draw macro body, then the vertex/store, the
+window/screen present, and the per-boot X first-generation start lottery
([`Graphics-Execution.md`](Graphics-Execution.md)).
## 6. Lines of investigation (live)
@@ -255,14 +268,14 @@ first-generation start lottery
4. **Hold the acceptance battery** — vector-add, matrix-multiply and the
transformer kernel; each entry is accepted only on its own verification
output (`bad=0` + `STATUS OK`).
-5. **The graphics pixel frontier** — the released replay now executes the
- draw macro (`DRAW_IAB`, `mac=0x22`, TCS-off DS clean, synthetic viewport,
- chained release); next identify the draw's true vertex buffer (no captured
- client push binds it; the eng71 scan candidate is a heuristic), make the
- render-target store/alias visible (`GUEST-WIT` coverage of the big-surface
- alias pages) so the draw stores (`wrote>0`) at the bound render target,
- then make the window/screen present land, and reduce the per-boot
- binding/X start lottery so the proven off-screen frames become
+5. **The graphics pixel frontier** — the real guest GL draw is now
+ submitted, admitted and replayed (`CALL(0)` `0x3800`, no wedge, first DS
+ with real stage data); next capture/verify the client's per-draw MME
+ macro body (the prelude carries only the setup macros) so the real body
+ executes instead of the synthetic golden one, then make the
+ render-target store visible (`GUEST-WIT` surface pages) so the draw
+ stores (`wrote>0`), then make the window/screen present land, and reduce
+ the per-boot X start lottery so the proven off-screen frames become
window pixels; the lanes, the walls and the milestone ladder are on
[`Graphics-Execution.md`](Graphics-Execution.md).