hyperloom-workload-optimizer
Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs and reports a validated gain, using the Hyperloom multi-agent optimizer.
Skill by TOMBOSS885
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/tomboss885-hyperloom-workload-optimizer
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: backend, ml, finance, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions: not checked yet.
What's inside
- Skill name
hyperloom-workload-optimizer- Description
- Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs and reports a validated gain, using the Hyperloom multi-agent optimizer. Given a model, framework, workload (TP/EP, concurrency, ISL/OSL, precision), an objective and a time budget, it explores per-workload which levers to pull (serving/config parameters and env, framework enablement and source patches, and hot GPU-kernel rewrites), benchmarks each candidate, and returns the optimization stack that produced the gain. Use when the user wants to make a model serve faster, raise tokens/sec or throughput, optimize or tune vLLM or SGLang on MI300X/MI325X/MI355X, run Hyperloom, run the kernel-agent, quantize-then-optimize with Quark, set up Hyperloom from scratch, or resume a Hyperloom session. Do not use to stand up a server for plain serving, diagnose a broken ROCm install, or run a one-off kernel/benchmark or trace analysis without the optimization loop.
- License
- unknown
- Instructions
- 126 lines in SKILL.md · view on GitHub
Files (5)
- .federated.json
- SKILL.md
- evals.json · evals/
- machine.yml · evals/
- skill-card.md
The license isn't open source, so the instructions aren't shown here. Read SKILL.md on GitHub
More from TOMBOSS885/fpga
- 0.556score
rtl-module-interface-contract
TOMBOSS885
Diagnose and repair Verilog modules that fail Vivado elaboration because the module name, port names, directions, or widths do not match the declared…
SkillAFreshness: AClaude Code +9 - 0.556score
rtl-golden-scoreboard-verifier
TOMBOSS885
Self-checking testbench architecture with behavioral golden models, race-condition free stimulus driving, and automated mismatch scoreboards.
SkillAFreshness: AClaude Code +9 - 0.556score
rtl-interface-contract
TOMBOSS885
Diagnose and repair Verilog modules that fail Vivado elaboration because the module name, port names, directions, or widths do not match the declared interface.
SkillAFreshness: AClaude Code +9 - 0.556score
rtl-self-testbench-generator
TOMBOSS885
Guidelines and patterns for synthesizing companion self-checking testbenches to verify RTL designs with Vivado xsim before official scoring.
SkillAFreshness: AClaude Code +9 - 0.556score
rtl-shift-register-pattern
TOMBOSS885
Idioms for sequence detection and Linear Feedback Shift Registers (LFSR), ensuring correct overlapping detection and non-zero initialization.
SkillAFreshness: AClaude Code +9 - 0.553score
rtl-control-priority-pattern
TOMBOSS885
Design patterns for strict control signal priority hierarchies (Reset > Clear > Load > Enable) in sequential RTL.
SkillAFreshness: AClaude Code +9 - 0.551score
rtl-arithmetic-truncation
TOMBOSS885
Bit-width sizing, signed arithmetic extension, and population count rules to prevent Vivado truncation and overflow warnings.
SkillAFreshness: AClaude Code +9 - 0.551score
rtl-fsm-idioms
TOMBOSS885
Idiomatic finite state machine patterns in Verilog/SystemVerilog, ensuring clean state transition, output registration, and zero latch inference.
SkillAFreshness: AClaude Code +9 - 0.551score
rtl-shift-register-pattern
TOMBOSS885
Implement specified sequence overlap, output timing, LFSR taps, shift direction and seed without changing required state behavior.
SkillAFreshness: AClaude Code +9 - 0.551score
rtl-clock-reset-conventions
TOMBOSS885
Rules and patterns for clocking and reset implementations in FPGA synthesis, preventing multi-driven nets, incorrect sensitivity lists, and…
SkillAFreshness: AClaude Code +9 - 0.551score
rtl-control-priority-pattern
TOMBOSS885
Implement specification-defined reset, clear, load and enable priority, and verify simultaneous assertions in sequential RTL.
SkillAFreshness: AClaude Code +9 - 0.548score
rtl-synthesis-latch-prevention
TOMBOSS885
Techniques to detect and eliminate unintended latch inference and non-synthesizable constructs in Vivado synthesis.
SkillAFreshness: AClaude Code +9
Signals
- Stars
- 0
- Forks
- 2
- Open issues
- 1
- Last commit (this folder)
- Oct 1, 2026 (27 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- License unknown
- Repository
- github.com/TOMBOSS885/fpga
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
Same skill elsewhere
Copies of this SKILL.md published by others; search shows only one of them.
More from TOMBOSS885
- 0.556score
rtl-module-interface-contract
TOMBOSS885
Diagnose and repair Verilog modules that fail Vivado elaboration because the module name, port names, directions, or widths do not match the declared…
SkillAFreshness: AClaude Code +9 - 0.556score
rtl-golden-scoreboard-verifier
TOMBOSS885
Self-checking testbench architecture with behavioral golden models, race-condition free stimulus driving, and automated mismatch scoreboards.
SkillAFreshness: AClaude Code +9 - 0.556score
rtl-interface-contract
TOMBOSS885
Diagnose and repair Verilog modules that fail Vivado elaboration because the module name, port names, directions, or widths do not match the declared interface.
SkillAFreshness: AClaude Code +9 - 0.556score
rtl-self-testbench-generator
TOMBOSS885
Guidelines and patterns for synthesizing companion self-checking testbenches to verify RTL designs with Vivado xsim before official scoring.
SkillAFreshness: AClaude Code +9 - 0.556score
rtl-shift-register-pattern
TOMBOSS885
Idioms for sequence detection and Linear Feedback Shift Registers (LFSR), ensuring correct overlapping detection and non-zero initialization.
SkillAFreshness: AClaude Code +9 - 0.553score
rtl-control-priority-pattern
TOMBOSS885
Design patterns for strict control signal priority hierarchies (Reset > Clear > Load > Enable) in sequential RTL.
SkillAFreshness: AClaude Code +9
Cite this page
hyperloom-workload-optimizer skill. Agent Mart. https://agentmart.dev/skills/tomboss885-hyperloom-workload-optimizer (last updated 2026-10-24)