v0.8.2 — multi-threaded inference + verified rooflines

Hologram v0.8.2

Installation

Via Git Dependency

Add to your Cargo.toml:

[dependencies]
# Depend on the specific crates you need (all share one version):
hologram-compiler = { git = "https://github.com/Hologram-Technologies/hologram", tag = "v0.8.2" }
hologram-exec     = { git = "https://github.com/Hologram-Technologies/hologram", tag = "v0.8.2" }
hologram-backend  = { git = "https://github.com/Hologram-Technologies/hologram", tag = "v0.8.2" }

CLI Binary

Download the appropriate binary for your platform from the assets below and add it to your PATH.

Changes

Commits since v0.8.1

• fix(backend): pool admission — work admits, width declines; all-tier native prefill (f031e8b)
• feat(backend): pool the m>1 GEMM in the wasm worker pool; verify the rooflines (0ac84a0)

What's Changed

• feat: pool the m>1 prefill GEMM; verify kernels at their rooflines (v0.8.2) by @afflom in #40

Full Changelog: v0.8.1...v0.8.2