#1873 perf(vk): a start reads the pipeline notes file once, not twice (about 0.28 s off every start)
Summary
The change passed on Linux NVIDIA, Linux RADV and Windows NVIDIA; these tests had no performance thresholds. On the Linux GTX 1070, field frame time p99 worsened from 23.66 to 24.88 ms (+5.2%); on RADV, p99 rose from 21.00 to 21.18 ms (+0.9%), while maximum field frame time rose from 28.26 to 67.63 ms and one field hitch was recorded. On the Windows RTX 3090, p99 was essentially unchanged at 13.63 → 13.65 ms, while load-at-create fell from 287 to 21 ms. A separate Windows RTX 4080 report failed during the baseline build before the game ran, indicating a PC-specific build problem rather than a candidate gameplay failure.
Commit
Plan (gpt-6.1-terra)
Compare startup and Hyrule Field on main and the PR across Linux NVIDIA, Linux RADV and Windows NVIDIA
In plain words
Compare two alternating starts per side on Linux NVIDIA, Linux RADV and Windows NVIDIA PCs, then teleport from Kokiri Forest to Hyrule Field. Each tester takes about 22 minutes; startup, shader preparation, load-at-create time and gameplay frame metrics are reported without performance thresholds. A crash or failure to reach Hyrule Field within the timeout fails the test. Matching hardware must be available for each target; otherwise that target waits.
Asked for
Start-up time, this PR against main: the PR reads the pipeline notes file once a start instead of twice, about 0.28 s less on an RX 9070 XT with AMD's Windows driver. Nothing changes in play. What this PC can't show: Linux and NVIDIA. - Targets: Linux with NVIDIA, Linux with AMD (RADV), Windows with NVIDIA (`vram_gb = ">= 6"` on Windows targets), as many as are free. - One scenario from `kokiri-start`, uncapped and windowed: `frame_rate_cap = 6`, `vsync = 2`, `frame_generation = 0`, `fullscreen = 0`. `wait_s = 10` before the teleport to Hyrule Field, `timeout_s = 120` on its map wait, then measure 20 s. - 2 starts a side (main and this PR, alternating). - Probes: `boot_time` and `shader_compile` (the start's `vk: precompile:` seconds) are what this PR changes. Also a `log_value` of the load-at-create line's milliseconds (`load at create \(RX_VK_LOAD=create\).*; in (\d+) ms`), which holds main's second walk. - Pass rules: no crash or hang. `boot_time_s`, `shader_compile_s`, `fps_mean`, `frametime_p99_ms`, `run_hitch_count` and `run_frametime_max_ms` are report-only (one start's time is noisy, and play should be unchanged).
Raw plan Download plan.json
{
"pass": {
"crash": false
},
"probe": [
{
"kind": "crash"
},
{
"kind": "boot_time"
},
{
"kind": "shader_compile"
},
{
"kind": "frametime"
},
{
"kind": "draws"
},
{
"kind": "log_value",
"value": [
{
"name": "load_create_ms",
"take": "max",
"pattern": "load at create \\(RX_VK_LOAD=create\\).*; in (\\d+) ms"
}
]
}
],
"matrix": {},
"schema": 7,
"target": [
{
"name": "linux+nvidia",
"tags": [
"linux",
"nvidia"
],
"max_reports": 2,
"min_reports": 1
},
{
"name": "linux+radv",
"tags": [
"linux",
"radv"
],
"max_reports": 2,
"min_reports": 1
},
{
"name": "windows+nvidia+vram-ge6gb",
"tags": [
"windows",
"nvidia"
],
"vram_gb": ">= 6",
"max_reports": 2,
"min_reports": 1
}
],
"summary": "Compare startup and Hyrule Field on main and the PR across Linux NVIDIA, Linux RADV and Windows NVIDIA",
"baseline": {
"ref": "main"
},
"scenario": [
{
"env": {},
"name": "startup-field",
"runs": 2,
"step": [
{
"do": "wait",
"event": "in_play",
"timeout_s": 900
},
{
"do": "wait_s",
"seconds": 10
},
{
"do": "teleport",
"to": "HyruleField_Child"
},
{
"do": "wait",
"map": "HyruleField_Child",
"event": "map",
"timeout_s": 120
},
{
"do": "measure",
"name": "field",
"warmup_s": 10,
"duration_s": 20
}
],
"start": "save:kokiri-start",
"matrix": true,
"settings": {
"vsync": 2,
"fullscreen": 0,
"frame_rate_cap": 6,
"frame_generation": 0
}
}
]
}
Targets
Reports
| Report | Commit | Target | Tester | Result | When |
|---|---|---|---|---|---|
| rp_2c0bd8afb270d731 | cefaf33 | linux+nvidia | jabbaxipil/Blaine | pass outlier | |
| rp_e9712ff617aa30e4 | cefaf33 | windows+nvidia+vram-ge6gb | Flashman4/agent-de36c5 | pass | |
| rp_a723f9e4a05d5bea | cefaf33 | linux+nvidia | Nerdtality/agent-0452e3 | pass outlier | |
| rp_8df7ec4cbef3f6da | cefaf33 | linux+radv | icyblue8/agent-368f00 | pass | |
| rp_a4df5a2c5215f9e9 | cefaf33 | windows+nvidia+vram-ge6gb | CubanLegend/agent-341924 | error |
Jobs
| Job | Commit | Target | State | Tester | Started | Attempts | Estimate | Note |
|---|---|---|---|---|---|---|---|---|
jb_ca048cd72903397a log | cefaf33 | linux+radv | done report | icyblue8/agent-368f00 | 1 | |||
jb_d07fe2bcef62bde0 log | cefaf33 | windows+nvidia+vram-ge6gb | done report | Flashman4/agent-de36c5 | 1 | |||
jb_b2efb1e0ee7c9234 log | cefaf33 | linux+nvidia | done run | jabbaxipil/Blaine | 1 | |||
jb_6ec1aff1fb39e44a log | cefaf33 | linux+nvidia | done run | Nerdtality/agent-0452e3 | 1 | |||
jb_11719fab4c3c84bb log | cefaf33 | windows+nvidia+vram-ge6gb | done build-base | CubanLegend/agent-341924 | 1 | |||
jb_5f3a73dd52526eb9 log | cefaf33 | linux+radv | cancelled | — | 6 | the PR was closed |