[jcurtis2]

Reorganizes and expands batch.rst, and corrects GPU information based on current cluster state.

Reorganization:

• Added "Available compute resources" section at the top consolidating partition and node tables
• Expanded "Batch computing" section with an explanation of how batch jobs work, an sbatch usage example, and directive reference tables
• Added "Accessing GPU nodes" subsection covering both GPU partitions
• Promoted subsections and genericized example scripts (removed hardcoded job names/emails)

Corrections:

• gpu partition: updated GPU type from K40m to Titan RTX (K40m GPUs are no longer supported in AL9.x)
• gpu partition: corrected wall clock limit from 7 days to 2 days
• Added GPU VRAM to node table: Titan RTX (24 GB), L40S (46 GB)
• Updated --gres flag in examples to match Slurm identifiers (gpu:RTX:1, gpu:L40S:1)

---

📚 Documentation preview 📚: https://computing-docs--31.org.readthedocs.build/en/31/