Homelab
Given that fluid simulation is typically resource intensive, I always had dreams of building my own cluster/supercomputer. After being exposed to more IT-related work in my role with Ansys, I found an excuse to spend too much money on computer hardware.
Hardware
Expand tags for more info
battlestation
My primary desktop workstation for productivity and gaming, dual boots Debian 12 and Windows 10
- AMD Ryzen 5800X3D 8c/16t
- ASRock X570 Phantom Gaming 4
- 4x 16 GB DDR4-3200
- XFX 7900XTX 24 GB
- 2x 2 TB NVMe SSDs
- Lian Li O11 Air Mini ATX chassis
foam
Rack-mounted dual-socket system spec’d as much as my wallet allows, runs Debian 12
- 2x AMD EPYC 7K62 48c/48c
- ASRock Rack ROME2D16-2T
- 16x 64 GB ECC DDR4-3200
- 1 TB NVMe SSD
- Supermicro CSE-826 2U chassis
pve
Rack-mounted dual-socket system for virtualization, runs Proxmox
- 2x AMD EPYC 7532 32c/64t
- Supermicro H12DSi-NT6
- 8x 64 GB ECC DDR4-3200
- 512 GB NVMe SSD
- Supermicro CSE-826 2U chassis
reservoir
TrueNAS Scale box for network storage
- AMD Ryzen 5 3600 6c/12t
- ASRock Rack X570D4U-2L2T
- 4x 16 GB DDR4-3200
- 2x 58 GB Optane P1600X mirrored boot drives
- 7x 14 TB HDD RAID-Z2, 1x spare
- Fractal Node 804 mATX chassis
minis
Low-powered mini PCs currently used in a Proxmox cluster for most home server applications
- 3x Beelink EQ12
- Intel N100 4c/4t
- 16 GB DDR5-4800
- 500GB NVMe SSD
compute
Underutilized PC that has been rock solid, runs DietPi
- AMD EPYC 7F32 8c/16t
- AsRock Rack ROMED8-2T
- 256 GB DDR4
- 500GB NVMe Proxmox boot drive
- Fractal Define 7 ATX chassis
Applications
- Blog: Hugo
- AI stack: Open WebUI + Ollama
- Media server stack: Jellyfin + *arr apps
- File server: NFS/SMB shares
- Git repo: Gitea, GitLab
- Password vault: Vaultwarden
- Collaborative Markdown editor: CodiMD
Future Goals
- Configure Kubernetes cluster