bytelyst/learning_ai_common_plat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

docs(windows): flesh out HP Z240 spec + update README with all machines

Browse Source 
- HP Z240 spec: expanded from 7-line raw info to full spec document
  with hardware details, capabilities assessment, OpenClaw server
  use case, home lab services table, GPU upgrade path, machine
  comparison, and setup recommendations
- README: added Machines table (Razer/HP Z240/Dell P16s), expanded
  files table, added Related: OpenClaw section with links to OPEN_CLAW/
This commit is contained in:
saravanakumardb1 2026-02-22 15:41:20 -08:00
parent 09d3ffe20d
commit b3da5dd66e
2 changed files with 271 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
__LOCAL_LLMs/windows_specific/README.md
View File

@ -58,15 +58,37 @@ Dashboard should show all green. Done.
---
## Machines
| Machine | Hostname | Role | Spec File |
| ----------------------------- | ----------- | ---------------------------------- | ------------------------------------------------------ |
| **Razer Blade 18** (RTX 5090) | — | ML powerhouse, GPU inference | [razer-blade-18-spec.md](razer-blade-18-spec.md) |
| **HP Z240 Tower** (i7-7700K) | bl1box | Always-on server, OpenClaw Gateway | [hp-z240-windows-spec.md](hp-z240-windows-spec.md) |
| **Dell P16s** (Ryzen 7 PRO) | WIN-6TAK... | Portable workstation | [dell-P16s-windows-spec.md](dell-P16s-windows-spec.md) |
## Files in This Directory
| File | Purpose |
| -------------------------- | ------------------------------------------ |
| **README.md** | This file — quick start guide |
| **setup-windows.ps1** | PowerShell script — Windows-side setup |
| **setup-wsl.sh** | Bash script — WSL2-side setup |
| **setup-guide.md** | Detailed manual guide with troubleshooting |
| **razer-blade-18-spec.md** | Full hardware specs for the Razer Blade 18 |
| File | Purpose |
| -------------------------------- | --------------------------------------------- |
| **README.md** | This file — quick start guide |
| **setup-windows.ps1** | PowerShell script — Windows-side setup |
| **setup-wsl.sh** | Bash script — WSL2-side setup |
| **setup-guide.md** | Detailed manual guide with troubleshooting |
| **razer-blade-18-spec.md** | Full hardware specs for the Razer Blade 18 |
| **hp-z240-windows-spec.md** | HP Z240 spec + OpenClaw server use case guide |
| **dell-P16s-windows-spec.md** | Dell P16s system info |
| **mac-vs-windows-comparison.md** | Side-by-side Mac vs Windows comparison |
| **capabilities/** | 7 deep-dive GPU capability guides |
## Related: OpenClaw AI Assistant
The HP Z240 is the recommended always-on host for [OpenClaw](https://github.com/openclaw/openclaw) — a self-hosted AI assistant that connects to WhatsApp, Telegram, Slack, Discord, and more.
| File | Purpose |
| -------------------------------------------------------------------------------------------------- | --------------------------------------------- |
| [`../OPEN_CLAW/SETUP_GUIDE.md`](../OPEN_CLAW/SETUP_GUIDE.md) | Step-by-step install + secure setup guide |
| [`../OPEN_CLAW/openclaw-personal-ai-assistant.md`](../OPEN_CLAW/openclaw-personal-ai-assistant.md) | Reference doc — features, security, tips |
| [`../OPEN_CLAW/validate-security.sh`](../OPEN_CLAW/validate-security.sh) | Security validation script (run post-install) |
---

249
__LOCAL_LLMs/windows_specific/hp-z240-windows-spec.md
View File

@ -1,7 +1,242 @@
Device name bl1box
Processor Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz (4.20 GHz)
Installed RAM 32.0 GB
Device ID 4DA67C13-70D5-4D44-AFF2-311D1F42FD1E
Product ID 00330-51031-03023-AAOEM
System type 64-bit operating system, x64-based processor
Pen and touch No pen or touch input is available for this display
# HP Z240 Tower Workstation — Specification & Use Case Guide
> **Hostname:** `bl1box` · **Form Factor:** Tower Workstation · **Era:** 2017 (Kaby Lake)
> **Primary Role:** Always-on home server — OpenClaw Gateway, Docker, file server
---
## 1. System Overview
The HP Z240 Tower is an enterprise-class workstation from ~2017. While no longer competitive for AI/ML inference, it's an excellent low-power, always-on home server for services that don't need a GPU.
### Raw System Info (from Windows)
| Field | Value |
| ------------- | --------------------------------------------------- |
| Device name | bl1box |
| Processor | Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz (4.20 GHz) |
| Installed RAM | 32.0 GB |
| Device ID | 4DA67C13-70D5-4D44-AFF2-311D1F42FD1E |
| Product ID | 00330-51031-03023-AAOEM |
| System type | 64-bit operating system, x64-based processor |
---
## 2. Hardware Specifications
### CPU
| Attribute | Specification |
| -------------------- | ------------------------- |
| **Model** | Intel Core i7-7700K |
| **Architecture** | Kaby Lake (7th Gen) |
| **Base Clock** | 4.20 GHz |
| **Boost Clock** | 4.50 GHz |
| **Cores / Threads** | 4 / 8 |
| **TDP** | 91W |
| **Fabrication** | 14nm |
| **Instruction Sets** | SSE4.2, AVX2 (no AVX-512) |
| **Integrated GPU** | Intel HD Graphics 630 |
### Memory
| Attribute | Specification |
| ----------------- | ---------------------------------------- |
| **Installed** | 32 GB |
| **Type** | DDR4 (likely 2400 MHz, HP Z240 max) |
| **Max Supported** | 64 GB (4 DIMM slots) |
| **ECC** | Supported (Z240 Tower supports ECC DDR4) |
### Storage
| Attribute | Specification |
| ------------------ | -------------------------------------------- |
| **Drive Bays** | 2x 3.5" + 1x 2.5" internal |
| **M.2 Slot** | 1x M.2 2280 (PCIe 3.0 x4 NVMe or SATA) |
| **SATA Ports** | 4x SATA III (6 Gbps) |
| **Current Config** | TBD — check with `wmic diskdrive list brief` |
### Expansion
| Attribute | Specification |
| --------------- | --------------------------------- |
| **PCIe x16** | 1 slot (PCIe 3.0) — for GPU |
| **PCIe x4** | 1 slot (PCIe 3.0) |
| **PCIe x1** | 2 slots |
| **PSU** | 400W (80+ Platinum, standard ATX) |
| **Form Factor** | Tower (full ATX) |
### GPU
| Attribute | Status |
| ---------------- | ------------------------------------------------------------ |
| **Current GPU** | Unknown — likely Intel HD 630 (integrated) |
| **GPU Slot** | PCIe 3.0 x16 (available for upgrade) |
| **PSU Headroom** | ~250W available for GPU (400W PSU - ~150W system) |
| **Max GPU** | RTX 3060 12GB (~170W) fits easily; RTX 3070 (~220W) possible |
> **Check GPU:** Run `wmic path win32_videocontroller get name,adapterram` on the machine.
### Network
| Attribute | Specification |
| ------------ | ------------------------------- |
| **Ethernet** | Intel I219-LM Gigabit (onboard) |
| **WiFi** | None (add via PCIe or USB) |
### Power
| State | Estimated Draw |
| ----------------------------- | ------------------ |
| **Idle** | ~45–65W |
| **Light Load** (server tasks) | ~70–90W |
| **Heavy Load** (CPU stress) | ~130–150W |
| **Monthly Cost** (24/7 idle) | ~$5–7 at $0.12/kWh |
---
## 3. Capabilities Assessment
### What It CAN Do Well
| Use Case | Performance | Notes |
| --------------------------------- | ----------- | -------------------------------------------- |
| **OpenClaw Gateway** | Excellent | CPU-only, needs < 500 MB RAM |
| **Docker containers** | Excellent | 32 GB RAM, 4c/8t is plenty |
| **File server (SMB/NFS)** | Excellent | Gigabit Ethernet, multiple drive bays |
| **Git server (Gitea)** | Excellent | Lightweight, runs on anything |
| **Reverse proxy (Traefik/Nginx)** | Excellent | Minimal resources needed |
| **Database (PostgreSQL/Redis)** | Good | 32 GB RAM is generous |
| **Ollama (CPU-only)** | Slow | 5–10 tok/s on 7B models — usable for testing |
| **CI runner (GitHub Actions)** | Good | 4c/8t handles builds fine |
| **Tailscale exit node** | Excellent | Always-on VPN gateway |
| **Home Assistant** | Excellent | Very lightweight |
| **Pi-hole / DNS** | Excellent | Trivial workload |
### What It CANNOT Do (Without GPU Upgrade)
| Use Case | Why Not |
| --------------------------------------- | ------------------ |
| **GPU inference (Ollama CUDA)** | No discrete GPU |
| **Whisper transcription (CUDA)** | No discrete GPU |
| **TTS generation (CUDA/MPS)** | No discrete GPU |
| **Fine-tuning / training** | No GPU, no AVX-512 |
| **Image generation (Stable Diffusion)** | No GPU |
---
## 4. Recommended Use Cases
### Primary: OpenClaw Always-On Gateway
The HP Z240 is ideal for running OpenClaw 24/7:
```
┌──────────────────────────────────────────────────────────────────┐
│ HP Z240 "bl1box" — Always-On Server │
│ │
│ ┌────────────────────────────────────────────────────────┐ │
│ │ WSL2 Ubuntu 24.04 │ │
│ │ │ │
│ │ ┌──────────────────┐ ┌──────────────────────────┐ │ │
│ │ │ OpenClaw Gateway │ │ Docker Containers │ │ │
│ │ │ ws://127.0.0.1: │ │ • Traefik (reverse proxy)│ │ │
│ │ │ 18789 │ │ • Gitea (git server) │ │ │
│ │ │ WhatsApp ✓ │ │ • PostgreSQL │ │ │
│ │ │ Telegram ✓ │ │ • Redis │ │ │
│ │ │ Discord ✓ │ │ • Uptime Kuma │ │ │
│ │ └──────────────────┘ └──────────────────────────┘ │ │
│ │ │ │
│ │ Tailscale → secure remote access from Mac / phone │ │
│ └────────────────────────────────────────────────────────┘ │
│ │
│ Power: ~65W idle → ~$5/month │
│ Noise: Quiet (tower workstation fans) │
│ Uptime: 24/7/365 │
│ │
└──────────────────────────────────────────────────────────────────┘
```
See: [`../OPEN_CLAW/SETUP_GUIDE.md`](../OPEN_CLAW/SETUP_GUIDE.md) for full install & security guide.
### Secondary: Home Lab Services
| Service | Port | Purpose | RAM Usage |
| ---------------- | -------- | ------------------------------------- | ------------------- |
| OpenClaw Gateway | 18789 | AI assistant (WhatsApp/Telegram/etc.) | ~200 MB |
| Traefik | 80, 443 | Reverse proxy + auto SSL | ~50 MB |
| Gitea | 3000 | Self-hosted Git repos | ~100 MB |
| Uptime Kuma | 3001 | Service monitoring | ~80 MB |
| Pi-hole | 53, 8080 | DNS ad blocking | ~50 MB |
| PostgreSQL | 5432 | Database server | ~200 MB |
| Redis | 6379 | Cache / message broker | ~50 MB |
| Tailscale | — | VPN mesh (always-on) | ~30 MB |
| **Total** | | | **~760 MB / 32 GB** |
Plenty of headroom — 32 GB RAM leaves ~31 GB free for additional services.
---
## 5. Optional GPU Upgrade Path
If you want to add GPU inference capability:
| GPU | VRAM | Price (Used) | Power | Fits Z240? | Performance |
| ----------------- | ----- | ------------ | ----- | ---------------------- | ---------------- |
| **RTX 3060 12GB** | 12 GB | ~$180 | 170W | Yes (400W PSU OK) | 30–50 tok/s (7B) |
| **RTX 3060 Ti** | 8 GB | ~$200 | 200W | Yes | 40–60 tok/s (7B) |
| **RTX 3070** | 8 GB | ~$250 | 220W | Tight (may need PSU) | 50–70 tok/s (7B) |
| **RTX 3090** | 24 GB | ~$700 | 350W | No (needs PSU upgrade) | 60–80 tok/s (7B) |
**Best value:** RTX 3060 12GB — 12 GB VRAM fits larger models (13B quantized), and 170W works within the 400W PSU.
---
## 6. Comparison with Other Machines
| Capability | HP Z240 (bl1box) | Mac M4 Pro 48GB | Razer RTX 5090 |
| ----------------- | -------------------- | ----------------- | --------------------- |
| **Role** | Always-on server | Daily driver | ML powerhouse |
| **CPU** | i7-7700K (4c/8t) | M4 Pro (14c) | Ultra 9 275HX (24c) |
| **RAM** | 32 GB DDR4 | 48 GB unified | 64 GB DDR5 |
| **GPU** | None (integrated) | M4 Pro (MPS) | RTX 5090 24GB |
| **LLM Inference** | CPU-only (~5 tok/s) | Fast (MPS) | Fastest (CUDA) |
| **OpenClaw** | Perfect | Good (daily use) | Overkill |
| **Docker** | Excellent | Good | Excellent |
| **Power (idle)** | ~65W | ~10W | ~30W (sleep) |
| **Always-on?** | Yes — primary server | No — daily laptop | No — gaming/ML laptop |
| **Cost** | ~$100 used | ~$2,500 | ~$4,500 |
---
## 7. Setup Recommendations
### For OpenClaw + Home Lab
1. Install Windows 11 (or keep existing) + WSL2 Ubuntu 24.04
2. Install Tailscale (Windows-native + WSL2)
3. Install OpenClaw in WSL2 → follow [`../OPEN_CLAW/SETUP_GUIDE.md`](../OPEN_CLAW/SETUP_GUIDE.md)
4. Run `validate-security.sh` → fix all issues
5. Set up Docker for additional services
6. Enable Windows auto-login + WSL2 auto-start on boot
7. Connect wired Ethernet, place somewhere quiet
### Auto-Start on Boot (Windows)
```powershell
# Create a scheduled task to start WSL2 on boot (PowerShell Admin)
$Action = New-ScheduledTaskAction -Execute "wsl" -Argument "-d Ubuntu"
$Trigger = New-ScheduledTaskTrigger -AtLogon
$Settings = New-ScheduledTaskSettingsSet -AllowStartIfOnBatteries
Register-ScheduledTask -TaskName "Start WSL2" -Action $Action `
-Trigger $Trigger -Settings $Settings -RunLevel Highest
```
### Auto-Login (Windows)
```
Settings → Accounts → Sign-in options →
"If you've been away, when should Windows require you to sign in again?" → Never
```
Or use `netplwiz` to disable the login screen entirely (single-user machine only).