hpc-agentic-sdk

PJM (pjsub/pjstat/pjdel) job management, filesystem operations, and documentation search for the RIKEN Fugaku supercomputer

7 skills · 2 MCP servers

Slurm job management, filesystem operations, and documentation search for the RIKEN HOKUSAI BigWaterfall2 (HBW2) supercomputer

6 skills · 2 MCP servers

Slurm job management, filesystem operations, and documentation search for the RIKEN R-CCS Cloud (heterogeneous: A64FX, x86_64, NVIDIA/AMD/Intel GPUs)

6 skills · 2 MCP servers

Slurm job management, filesystem operations, and documentation search for the RIKEN AI4S supercomputer (NVIDIA GB200, aarch64)

6 skills · 2 MCP servers

Expose Makefile targets as MCP tools for building and running code on local or remote HPC systems via SLURM

1 skill · 1 MCP server

Check whether loop transformations (interchange, tiling, fusion) are mathematically safe using polyhedral analysis

1 skill · 1 MCP server

Use and contribute to BenchKit, a shell-first framework for building and running HPC benchmarks across multiple systems

1 skill

Use and contribute to Benchpark, a Spack+Ramble-based framework for specifying reproducible HPC benchmarks across multiple systems

1 skill

Set up a new HPC project by determining goals, installing required skills, and creating a CLAUDE.md with mandatory workflows

1 skill

Match application kernels to representative HeCBench workloads by computational motif.

1 skill

Match application kernels to representative HeCBench workloads by computational motif

1 skill

Measure per-kernel hot working sets, memory traffic, and FLOPs in MPI C/C++/Fortran codes for GPU memory planning.

1 skill

Measure per-kernel hot working sets, memory traffic, and FLOPs in MPI C/C++/Fortran codes for GPU memory planning

1 skill

Profile GPU and MPI applications with NVIDIA Nsight Systems/Compute (nsys/ncu) on isolated HPC clusters: per-rank wrapper injection, replay control, Grid-Size fingerprinting, and latency-bound detection from the CLI.

1 skill

Profile GPU and MPI applications with NVIDIA Nsight Systems/Compute (nsys/ncu) on isolated HPC clusters: per-rank wrapper injection, replay control, Grid-Size fingerprinting, and latency-bound detection from the CLI

1 skill

Automatically insert OpenACC directives into Fortran code using a verified template library for reliable GPU parallelization

1 skill

Profile and alter floating-point precision in C, C++, and Fortran using RAPTOR

1 skill