KV-Cache Receipts β 3Γ the 128K users per A100 (vLLM)
Collection
Three-arm, needle-gated serving receipts (fp16 / fp8-KV / compressed) β reproducible for β$12 with one command. β’ 3 items β’ Updated
Prebuilt binary wheel of the fraqtl vLLM runtime, published so anyone can independently reproduce the fraQtl KV-cache receipts on a rented A100 for β$15 β without the kernel or calibration source being distributed.
fraqtl/_native/libggml-cuda.so). No .cu/.c/header source ships; the build hard-fails if any leaks in (see wheel_build_receipt.json).torch.compile-traceable or the engine provisions a smaller KV pool than the receipts. It is orchestration glue: how eigenbases are computed is not in the wheel in any form (no calibration or sidecar-factory code; remaining runtime modules are Cython-compiled binaries).vllm.general_plugins entry point β pip install the wheel and vLLM's plugin loader finds the backend in every worker process.Don't drive this by hand β use the one-command repro script (fraqtl_repro_receipts.py, in this repo and in the sidecar repos):
pip install modal && modal setup
modal secret create huggingface HF_TOKEN=hf_...
modal run fraqtl_repro_receipts.py
It installs this wheel, pulls the calibrated sidecars from
fraQtl/mistral-7b-instruct-v0.3-kv-sidecars,
runs the three-arm receipt (fraQtl D2 vs fp16 vs fp8-KV) at 8K and 32K with a
needle-in-a-haystack gate per arm, and prints the table.
| File | What |
|---|---|
fraqtl_runtime-*.whl |
The prebuilt wheel (binary-only) |
wheel_build_receipt.json |
Build provenance: sha256, source-leak audit result, module list |
fraqtl_repro_receipts.py |
The one-command repro script |
Proprietary; free to install and run for verification and evaluation. No reverse engineering. Sidecars and receipts in the companion repos are Apache 2.0.