7978eaea14
- Root setup.sh orchestrator with per-phase dispatch (./setup.sh p1..p8 | all | status) - Makefile convenience targets (make infra, make llm, make status, etc.) - scripts/common.sh: shared bash library for OS detection, Docker helpers, service management (launchd/systemd), package install, env management - .env.example + .gitignore: shared config template and secret exclusions P1 (homeai-infra): full implementation - docker-compose.yml: Uptime Kuma, code-server, n8n - Note: Home Assistant, Portainer, Gitea are pre-existing instances - setup.sh: Docker install, homeai network, container health checks P2 (homeai-llm): full implementation - Ollama native install with CUDA/ROCm/Metal auto-detection - launchd plist (macOS) + systemd service (Linux) for auto-start - scripts/pull-models.sh: idempotent model puller from manifest - scripts/benchmark.sh: tokens/sec measurement per model - Open WebUI on port 3030 (avoids Gitea :3000 conflict) P3-P8: working stubs with prerequisite checks and TODO sections Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
89 lines
2.2 KiB
Bash
89 lines
2.2 KiB
Bash
#!/usr/bin/env bash
|
|
# scripts/benchmark.sh — Benchmark Ollama model inference speed
|
|
#
|
|
# Measures tokens/sec for each installed model.
|
|
# Results written to benchmark-results.md
|
|
#
|
|
# Usage:
|
|
# bash scripts/benchmark.sh
|
|
# bash scripts/benchmark.sh qwen2.5:7b # benchmark one model
|
|
|
|
set -euo pipefail
|
|
|
|
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
|
|
REPO_DIR="$(cd "${SCRIPT_DIR}/../.." && pwd)"
|
|
source "${REPO_DIR}/scripts/common.sh"
|
|
|
|
RESULTS_FILE="${SCRIPT_DIR}/../benchmark-results.md"
|
|
PROMPT="Tell me a short story about a robot who loves cooking. Keep it to exactly 200 words."
|
|
|
|
if ! command_exists ollama; then
|
|
die "Ollama not found."
|
|
fi
|
|
|
|
if ! curl -sf http://localhost:11434 -o /dev/null; then
|
|
die "Ollama is not running."
|
|
fi
|
|
|
|
benchmark_model() {
|
|
local model="$1"
|
|
log_step "Benchmarking $model..."
|
|
|
|
local start end elapsed
|
|
start=$(date +%s%3N)
|
|
|
|
local response
|
|
response=$(ollama run "$model" "$PROMPT" 2>&1) || {
|
|
log_error "Model $model failed to run."
|
|
echo "| $model | ERROR | — |"
|
|
return
|
|
}
|
|
|
|
end=$(date +%s%3N)
|
|
elapsed=$(( (end - start) ))
|
|
|
|
local word_count
|
|
word_count=$(echo "$response" | wc -w)
|
|
local tokens_est=$(( word_count * 4 / 3 )) # rough estimate: 1 token ≈ 0.75 words
|
|
local elapsed_sec
|
|
elapsed_sec=$(echo "scale=1; $elapsed / 1000" | bc)
|
|
local tps
|
|
tps=$(echo "scale=1; $tokens_est / ($elapsed / 1000)" | bc 2>/dev/null || echo "?")
|
|
|
|
printf " %-30s %6s tok/s (%ss)\n" "$model" "$tps" "$elapsed_sec"
|
|
echo "| \`$model\` | ${tps} tok/s | ${elapsed_sec}s |"
|
|
}
|
|
|
|
log_section "Ollama Benchmark"
|
|
log_info "Prompt: '$PROMPT'"
|
|
echo ""
|
|
|
|
if [[ -n "${1:-}" ]]; then
|
|
models=("$@")
|
|
else
|
|
# Get list of installed models
|
|
mapfile -t models < <(ollama list 2>/dev/null | tail -n +2 | awk '{print $1}')
|
|
fi
|
|
|
|
if [[ ${#models[@]} -eq 0 ]]; then
|
|
die "No models installed. Run: bash scripts/pull-models.sh"
|
|
fi
|
|
|
|
{
|
|
echo "# Ollama Benchmark Results"
|
|
echo "> Generated: $(date)"
|
|
echo ""
|
|
echo "| Model | Speed | Time for ~200 tok |"
|
|
echo "|---|---|---|"
|
|
} > "$RESULTS_FILE"
|
|
|
|
for model in "${models[@]}"; do
|
|
benchmark_model "$model" | tee -a "$RESULTS_FILE"
|
|
done
|
|
|
|
echo "" >> "$RESULTS_FILE"
|
|
|
|
log_success "Results written to $RESULTS_FILE"
|
|
echo ""
|
|
cat "$RESULTS_FILE"
|