CoreTechX On-Prem

Infrastructure Planner

Rough per-H100 sizing: users, workloads, and supporting CPU, memory, and storage.

Planning unit 1x NVIDIA H100

H100 GPU count 1

OCR pipeline Page image processing on local GPU and CPU Semantic indexing Embedding edited page content locally AI assistant Chat and assistant requests served by local LLM

Planning assumptions

Local LLM size Embedding size Peak load factor Storage retention

Estimated active users 0 Planning estimate

General application load

Document pipeline load

Local LLM load

Processed page capacity/day: 0
Edited page capacity/day: 0
AI request capacity/hour: 0
CPU: 0 vCPU
Memory: 0 GB
Storage: 0 TB