Tenstorrent AI
Official@tenstorrent · United States of America
Offers specialized hardware-software integration for high-performance tensor processing, silicon emulation, and custom kernel development on proprietary mesh architectures.
Agent Skills by Tenstorrent AI
Showing 25 vetted skills indexed across 4 GitHub repositories.
tenstorrentmemory-debug
Diagnose data-corruption and memory-related failures in the tt-emule software emulator.
tenstorrentparallel-mock-implementation
Coordinates parallel worker agents to generate emulator mock files for the repository.
tenstorrentshepherd-emule-pr
Review and locally rebase tt-emule pull requests with JIT compile-probes.
tenstorrentcompute-llk-bringup
Automate compute-kernel LLK shim implementation and integration in tt-emule.
tenstorrentarch-lookup
Cross-reference Tenstorrent silicon specifications with tt-emule mock-API implementations.
tenstorrentuplift
Bisect tt-metal and tt-umd dependency updates to isolate C++ and JIT regressions.
tenstorrentindex-based-ops
Debug index-based compute operations like TopK, Sort, and Argmax in tt-emule.
tenstorrentworkarounds
Track deliberate code workarounds and their removal requirements in the tt-emule repository.
tenstorrentimplement-mock
Standardize silicon API mocks in the tt-emule emulator with strategy selection and verification.
tenstorrentverify-mock
Validate silicon API mocks in tt-emule against verification checklists.
tenstorrenttt-enable-tracing
Enable TTNN trace capture and replay to eliminate host dispatch overhead.
tenstorrenttt-bug-report
Creates GitHub issues for bug reports with minimized reproducers across TT repos.
tenstorrenttt-lang
Write TT hardware kernels using a Python DSL with data movement primitives.
tenstorrenttt-connect-remote-device
Establish remote connections to Tenstorrent hardware and run kernels, copy files, and read logs.
tenstorrenttt-lang-profile-optimize
Profile and optimize TT-Lang kernels using TT-Metal tooling.
tenstorrentttnn
Convert PyTorch tensors and execute tensor operations on Tenstorrent MeshDevice topologies.
tenstorrentadd-ttir-builder-op
...
tenstorrentvalidate-tt-mlir-against-tt-xla
Automate cross-repo validation of tt-mlir pull requests against tt-xla CI.
tenstorrentadd-op
Guide engineers through adding a new operation to the tt-mlir stack.
tenstorrentttir-decomposition-for-ttmetal
Register TTIR composite op decomposition patterns for the TTMetal backend.
tenstorrentadd-ttir-d2m-lowering
Automate TTIR elementwise lowering to D2M for TTMetal in MLIR.
tenstorrentexcalidraw-diagram
Generate Excalidraw diagram JSON encoding visual arguments for workflows and concepts.
tenstorrentanalyze-nightly
Summarize GitHub Actions nightly CI failures grouped by ownership area.
tenstorrentcode-reviewer
Automate structured code-review checks for the tt-xla project.
Frequently Asked Questions About Tenstorrent AI
FAQPage SchemaWhat specific tasks can engineers perform using Tenstorrent software stacks?▼
Engineers can execute tensor operations on mesh hardware, develop custom kernels with data movement primitives, profile performance bottlenecks, and validate silicon-level operations through specialized emulation environments.
Which technical personas benefit from these capabilities?▼
These capabilities are designed for hardware-software co-design engineers, compiler developers, and systems researchers focused on optimizing tensor-based workloads for custom silicon architectures.
What are the primary dependencies for running Tenstorrent kernel development?▼
Development requires access to the Tenstorrent software stack, including the TT-Metal environment, MLIR-based compiler infrastructure, and compatible Tenstorrent silicon or the provided software emulator for verification.