Matmul Kernel -- MCP Tool Specification
Version: v1.1.9 | Status: Active | Last Updated: March 2026
Overview
This document defines the MCP tools exposed by the matmul_kernel module.
These tools are auto-discovered by the PAI MCP bridge via the @mcp_tool decorator
in mcp_tools.py and surfaced as part of the ~303 dynamic tools available to Claude.
Auto-Discovery
| Property | Value |
|---|---|
| Discovery method | @mcp_tool decorator scan |
| Namespace | matmul_kernel |
| Trust default | Safe |
| PAI bridge | src/codomyrmex/agents/pai/mcp/ |
Tool Reference
matmul_compute
Description: Multiply two matrices using tiled algorithm. Trust Level: Safe Category: analysis
Parameters:
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
a | list[list[float]] | Yes | -- | 2D list representing matrix A (MxK) |
b | list[list[float]] | Yes | -- | 2D list representing matrix B (KxN) |
tile_size | int | No | 32 | Cache tile size |
Returns: dict -- Dictionary with result (2D list), shape, flops, max_error_vs_numpy, and correct (bool).
Example:
from codomyrmex.matmul_kernel.mcp_tools import matmul_compute
result = matmul_compute(
a=[[1.0, 2.0], [3.0, 4.0]],
b=[[5.0, 6.0], [7.0, 8.0]],
tile_size=16,
)
matmul_benchmark
Description: Benchmark tiled matmul against numpy for square matrices. Trust Level: Safe Category: analysis
Parameters:
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
max_size | int | No | 128 | Largest matrix size to test (max 512 to keep fast) |
Returns: dict -- Performance comparison results per matrix size with status.
Example:
from codomyrmex.matmul_kernel.mcp_tools import matmul_benchmark
result = matmul_benchmark(max_size=256)
Integration Notes
- Auto-discovered: Yes (via
@mcp_toolinmcp_tools.py) - Trust Gateway: All tools are safe -- no destructive operations
- PAI Phases: VERIFY (performance benchmarking), BUILD (matrix computation)
- Dependencies: Requires
numpyand internalkernelmodule (tiled_matmul, benchmark_matmul, matmul_flops)