dd403a907c
Add `make serve` and `make serve-ngram` for launching llama-server with baked-in optimal settings (Vulkan RADV, q4_0 KV cache, flash attention, no-mmap, full GPU offload). N-gram speculative decoding gives 1.1-1.4x tg speedup on repetitive content without upstream PR dependencies. Update Phase 5 status: MTP is months away (4 unmerged PRs, no MoE support), draft-model speculation stalled on ROCm buffer crashes.
89 lines
2.8 KiB
Makefile
89 lines
2.8 KiB
Makefile
.PHONY: help audit audit-full monitor monitor-simple benchmark benchmark-baseline benchmark-compare optimize verify test agentic-setup agentic-quick agentic-full
|
|
|
|
help: ## Show available commands
|
|
@echo "Strix Halo Optimization Toolkit"
|
|
@echo ""
|
|
@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | awk 'BEGIN {FS = ":.*?## "}; {printf " \033[36m%-22s\033[0m %s\n", $$1, $$2}'
|
|
|
|
# --- Audit ---
|
|
audit: ## Quick system status (single screen)
|
|
@bash bin/audit --quick
|
|
|
|
audit-full: ## Full system report (saved to data/audits/)
|
|
@bash bin/audit --full
|
|
|
|
# --- Monitor ---
|
|
monitor: ## Launch tmux monitoring dashboard
|
|
@bash bin/monitor --dashboard
|
|
|
|
monitor-simple: ## Launch amdgpu_top only
|
|
@bash bin/monitor --simple
|
|
|
|
monitor-install: ## Install monitoring tools (amdgpu_top, btop)
|
|
@bash scripts/monitor/install-tools.sh
|
|
|
|
monitor-log: ## Start background metric logger
|
|
@bash bin/monitor --log
|
|
|
|
# --- Benchmark ---
|
|
benchmark-setup: ## Ensure toolboxes and test models are ready
|
|
@bash scripts/benchmark/setup.sh
|
|
|
|
benchmark-baseline: ## Capture pre-optimization baseline (supports ARGS="--max-size 20 --skip-longctx")
|
|
@bash bin/benchmark baseline $(ARGS)
|
|
|
|
benchmark: ## Run full benchmark suite (supports ARGS="--tag NAME --max-size 20")
|
|
@bash bin/benchmark run $(ARGS)
|
|
|
|
benchmark-compare: ## Compare two benchmark runs (usage: make benchmark-compare BEFORE=dir AFTER=dir)
|
|
@bash bin/benchmark compare $(BEFORE) $(AFTER)
|
|
|
|
# --- Serve ---
|
|
serve: ## Launch llama-server with optimized settings (ARGS="-m MODEL.gguf")
|
|
@bash bin/serve $(ARGS)
|
|
|
|
serve-ngram: ## Launch with n-gram speculative decoding (ARGS="-m MODEL.gguf")
|
|
@bash bin/serve --ngram $(ARGS)
|
|
|
|
# --- Optimize ---
|
|
optimize: ## Interactive optimization walkthrough
|
|
@bash bin/optimize --all
|
|
|
|
optimize-kernel: ## Configure kernel boot parameters
|
|
@bash scripts/optimize/kernel-params.sh
|
|
|
|
optimize-tuned: ## Switch to accelerator-performance profile
|
|
@bash scripts/optimize/tuned-profile.sh
|
|
|
|
optimize-power: ## Apply Phase 2 tuning (ryzenadj, sysctl, THP, RADV)
|
|
@bash scripts/optimize/power-profile.sh
|
|
|
|
optimize-vram: ## BIOS VRAM guidance + GTT verification
|
|
@bash scripts/optimize/vram-gtt.sh
|
|
|
|
verify: ## Post-optimization verification checklist
|
|
@bash scripts/optimize/verify.sh
|
|
|
|
rollback: ## Rollback optimizations
|
|
@bash scripts/optimize/rollback.sh
|
|
|
|
# --- Agentic Evaluation ---
|
|
agentic-setup: ## Install agentic evaluation frameworks (inspect-ai, evalplus)
|
|
@bash bin/agentic setup
|
|
|
|
agentic-quick: ## EvalPlus + IFEval quick eval (needs --model, ~1h)
|
|
@bash bin/agentic quick $(ARGS)
|
|
|
|
agentic-code: ## Code generation eval: EvalPlus + BigCodeBench (~2-3h)
|
|
@bash bin/agentic code $(ARGS)
|
|
|
|
agentic-tooluse: ## Tool/function calling eval: BFCL (~1-2h)
|
|
@bash bin/agentic tooluse $(ARGS)
|
|
|
|
agentic-full: ## All agentic evaluations (~5-6h)
|
|
@bash bin/agentic full $(ARGS)
|
|
|
|
# --- Tests ---
|
|
test: ## Run BATS test suite
|
|
@bats tests/
|