Home

Grok16 @ 3.0.0

Stable release — self-hosted g16 @ 16.2.0, versioned speed bench, uncompiled doctrine, Queen-themed manual.

Web manual	https://zacharygeurts.github.io/Grok16/
Speed bench	https://zacharygeurts.github.io/Grok16/speed-bench.html
Repo	https://github.com/ZacharyGeurts/Grok16
Tag	v3.0.0
Compiler	g16 (Grok16-16.2.0) 16.2.0

Speed bench (report v3.0.0)

Category	Winner	Compile	Execution
Fastest execution	C++ g16 belt_2_0	2,494 ms	85.3M ops/s
Fastest compile	C g16 belt_2_0	318 ms	79.5M ops/s
Best Python	host CPython 3	—	778K ops/s

Report: SPEED-BENCH-REPORT.md · Wiki: Speed-Bench

Single fabric (2.0 technology)

Knowing is fixed-size. Parallel I/O may fan in; truth collapses to one belt amplitude at depth 0.

• belt_2_0 — chunked redata (8192), wave-massive, single-location reads
• Depth-field creation forbidden at integrated consumers
• Time is linear — sovereign linear_ns (ironclad:time:1)

Start here

git clone https://github.com/ZacharyGeurts/Grok16.git
cd Grok16 && git checkout v3.0.0
export G16_PREFIX="$(pwd)"
export G16_BELT_PROFILE=belt_2_0
G16_RELEASE_PROFILE=1 ./scripts/grok16-toolchain.sh rebuild
./scripts/grok16-toolchain.sh test-battery-release
./scripts/grok16-toolchain.sh test-battery-belt
SPEED_DEMO_TARGET_SEC=3 ./scripts/grok16-toolchain.sh exec-full-bench
./scripts/grok16-toolchain.sh integrate

Manual map

Wiki	Web	Topic
Speed Bench	speed-bench.html	Versioned compile + exec bench
Uncompiled Execution	uncompiled.html	Python interpreter, chamber compile-ahead
CMake and Linking	cmake-linking.html	Toolchain, g16-ld, speed_demo CMake
Release 3.0	release.html	3.0 changelog
Single Fabric	single-fabric.html	Belt knowing, one amplitude
Safety	safety.html	Depth impossible, Ironclad
Getting Started	getting-started.html	Bootstrap, rebuild, verify
Performance	performance.html	Belt triad + speed bench
Integration	integration.html	Queen, World_Redata, integrate