`launch()` asks for `--use-angle=vulkan` and gets the Radeon RX 6700 XT. On the
real film loop at 1440x900 that is 7.56 s a frame to 0.47, and about twelve
cores pegged down to less than one. A 180-frame film goes from twenty-two
minutes to ninety seconds; the three shipped reels go from over an hour to four
minutes.
The SwiftShader flag was inherited from the share-card script and its comment
said "a box with no display and no GPU". Half of that was never true. The card
has been in this machine the whole time, with the amdgpu driver and a working
RADV ICD, and Chrome reaches it with **no display server at all** — no DISPLAY,
no XDG_RUNTIME_DIR, nothing but read access to /dev/dri/renderD128, which group
`render` already grants. So the portability the flag was protecting is intact:
this still runs over ssh and under cron.
Naming the backend is the whole trick, and most of the plausible spellings are
traps. `--use-angle=gl`, `--use-gl=egl`, `--use-gl=desktop`,
`--enable-features=Vulkan` and passing no GL flags at all were each measured on
this box, and all five land on SwiftShader while reporting success. Only
`--use-angle=vulkan` and `--use-angle=gl-egl` reach the card.
So `launch()` verifies instead of assuming. `--use-angle=vulkan` is a demand,
not a preference: with the driver hidden it yields no WebGL context at all
rather than falling back, the boot curtain never lifts, and the caller sits in
`waitForFunction` until a three-minute timeout before failing with something
that looks nothing like "there is no GPU here". Reading
`UNMASKED_RENDERER_WEBGL` once per run costs about half a second and turns that
into a printed line and a slow, correct render. Both branches are tested; the
sabotaged one was verified with `VK_LOADER_DRIVERS_DISABLE`.
The pictures are the same pictures: mean absolute difference of half a level out
of 255, confined to MSAA edges and to cloud shadow that already drifts between
any two runs of the same renderer.
Also corrects every comment across the four scripts that asserted three frames a
second or twenty-two minutes a reel. Those were true when written and are not
now, and the `reducedMotion` rationale in two of them was resting on the number:
the real reason a chapter flight cannot be waited out is that it is ~40 frames
however fast they are drawn, which was twenty seconds under software GL and is a
shorter race on the GPU — still a race.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
`shots.mjs` renders the product imagery for lumbridgecorp.com/simulate: seven
frames of the running `dist/`, at chapters and times chosen rather than
defaulted, written as WebP into the sibling site checkout together with a
generated manifest that carries each picture's caption and alt text. The pages
there had described a renderer in prose for their whole life without ever
showing one, which is a strange way to sell a renderer.
The captions live next to the camera poses in this repo, not on the site, so
re-aiming a camera cannot leave a caption behind describing the old view. The
manifest emits a `ShotId` union, so a page asking for a picture that has been
renamed fails the site's typecheck instead of rendering a hole.
Along the way: the share cards were never reproducible. `scenekit.ts` eases a
chapter change over about two seconds of scene time, `stage.ts` clamps `dt` to
50 ms a frame, and SwiftShader here draws about three frames a second — so the
flight takes twenty seconds of wall clock and `capture.mjs` waited four. Every
run caught the camera at a different point over the bay, and none of them at
the chapter the key press asked for. Both scripts now open the page with
reduced motion, which is the app's own answer to "somebody clicked a name in a
list": `flyTo` sets the pose outright. `og-tera.png` is regenerated and is
Hayes Valley for the first time. The bytes still differ run to run, because
aircraft are crossing and cloud shadow is drifting; the framing no longer does.
`harness.mjs` is the static server, the SwiftShader flags, the two-hostnames-
one-dist trick and the clock shim, extracted because there are two consumers
now and two copies would have drifted apart while both claimed to photograph
the same app.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
karti