[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-tilegym-converting-cutile-to-julia":3,"mdc-ecwbfr-key":34,"related-repo-nvidia-tilegym-converting-cutile-to-julia":1388,"related-org-nvidia-tilegym-converting-cutile-to-julia":1491},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":29,"sourceUrl":32,"mdContent":33},"tilegym-converting-cutile-to-julia","convert cuTile Python kernels to Julia","Converts cuTile Python GPU kernels (@ct.kernel) to cuTile.jl Julia equivalents. Handles kernel syntax translation, 0-indexed to 1-indexed conversion, broadcasting differences, memory layout (row-major to column-major), type system mapping, and launch API differences. Use when converting, porting, or translating cuTile Python kernels to Julia cuTile.jl, or debugging\u002Foptimizing existing Julia cuTile translations.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,17,20],{"name":13,"slug":14,"type":15},"Performance","performance","tag",{"name":9,"slug":8,"type":15},{"name":18,"slug":19,"type":15},"Engineering","engineering",{"name":21,"slug":22,"type":15},"Code Analysis","code-analysis",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:28:39.45458","CC-BY-4.0 AND Apache-2.0",281,[],{"repoUrl":24,"stars":23,"forks":27,"topics":30,"description":31},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Ftilegym-converting-cutile-to-julia","---\nname: tilegym-converting-cutile-to-julia\ndescription: Converts cuTile Python GPU kernels (@ct.kernel) to cuTile.jl Julia equivalents. Handles kernel syntax translation, 0-indexed to 1-indexed conversion, broadcasting differences, memory layout (row-major to column-major), type system mapping, and launch API differences. Use when converting, porting, or translating cuTile Python kernels to Julia cuTile.jl, or debugging\u002Foptimizing existing Julia cuTile translations.\nlicense: CC-BY-4.0 AND Apache-2.0\nmetadata:\n  author: \"TileGym Team \u003CTileGym@nvidia.com>\"\n  tags:\n    - cutile\n    - julia\n    - conversion\n    - gpu\n    - kernel\n---\n\n# cuTile Python → cuTile.jl (Julia) Conversion\n\nConvert `@ct.kernel` Python kernels to Julia `function ... end` cuTile.jl kernels.\n\n## Workflow Selection\n\n- **Standard conversion** → Full workflow: [`translations\u002Fworkflow.md`](translations\u002Fworkflow.md)\n- **Errors** (`MethodError`, `IRError`, numerical mismatch) → [`references\u002Fdebugging.md`](references\u002Fdebugging.md)\n- **Quick reference** → [`references\u002Fapi-mapping.md`](references\u002Fapi-mapping.md) + [`references\u002Fcritical-rules.md`](references\u002Fcritical-rules.md)\n- **Test patterns** → [`references\u002Ftesting.md`](references\u002Ftesting.md)\n\n## Architecture\n\nJulia kernels are **standalone** — no Python bridge, no pytest integration. The Julia sub-project\nlives in `julia\u002F` at the repo root with its own `Project.toml` for dependency management.\n\n```\njulia\u002F                          # Self-contained Julia sub-project\n├── Project.toml                # Dependencies: CUDA.jl, cuTile.jl, NNlib.jl, Test\n├── kernels\u002F                    # cuTile.jl kernel implementations\n│   ├── add.jl                  # ← Ground-truth: 1D element-wise with alpha scaling (tensor+tensor, tensor+scalar)\n│   ├── matmul.jl               # ← Ground-truth: 2D tiled MMA, standard Julia layout (M,K)×(K,N)→(M,N)\n│   └── softmax.jl              # ← Ground-truth: 3 strategies (TMA, online, chunked) using ct.load\u002Fct.store\n└── test\u002F                       # Julia-native tests (using Test stdlib)\n    ├── runtests.jl             # Test runner entry point\n    ├── test_add.jl\n    ├── test_matmul.jl\n    └── test_softmax.jl\n```\n\n**Ground-truth reference**: Always consult `julia\u002Fkernels\u002F*.jl` and `julia\u002Ftest\u002F*.jl` for patterns that compile and pass tests. These are the canonical examples of working cuTile.jl code.\n\n## Instructions\n\n1. **Analyze** the Python kernel: identify patterns, shapes, dtypes, operations\n2. **Write Julia kernel** — `julia\u002Fkernels\u002F\u003Cop>.jl` with cuTile.jl kernel + bridge function(s)\n3. **Convert** kernel signature (see `translations\u002Fworkflow.md` Phase 2)\n4. **Convert** kernel body (apply `references\u002Fapi-mapping.md` + `references\u002Fcritical-rules.md`)\n5. **Write Julia test** — `julia\u002Ftest\u002Ftest_\u003Cop>.jl` using `Test` stdlib + `NNlib.jl` for reference\n6. **Register test** — add `include(...)` in `julia\u002Ftest\u002Fruntests.jl`\n7. **Validate** — run the bundled validator: `python \u003Cskill-dir>\u002Fscripts\u002Fvalidate_cutile_jl.py \u003Cfile.jl>`\n8. **Test** — run `julia --project=julia\u002F julia\u002Ftest\u002Fruntests.jl`\n\nFull conversion checklist with post-conversion verification → [`translations\u002Fworkflow.md`](translations\u002Fworkflow.md)\n\n## ⚠️ Top Pitfalls\n\nThe most dangerous translation errors. Full rules (17 total) in [`references\u002Fcritical-rules.md`](references\u002Fcritical-rules.md).\n\n| # | Pitfall | One-line fix |\n|---|---------|-------------|\n| 1 | `ct.full()` doesn't exist in Julia | Use `fill(val, shape)`, `zeros(T, dims...)`, or `ones(T, dims...)` |\n| 2 | `max(a, b)` on tiles → `IRError` | Use `max.(a, b)` (broadcast dot) |\n| 3 | `IRError` \u002F `MethodError` mentioning `IRStructurizer` | Compiler bug — file upstream with minimal reproducer |\n| 4 | `ct.launch` arg order silently wrong | Args are positional — match kernel signature exactly |\n| 5 | `ct.load` with `order` — index positions wrong | `order` remaps BOTH shape AND index (Critical Rule 16) |\n\n## Worked Examples\n\nSide-by-side Python → Julia conversions matching the released Julia kernels in `julia\u002Fkernels\u002F`. Each directory contains `cutile_python.py` (before) and `cutile_julia.jl` (after).\n\n| # | Example | Key Patterns | When to Reference |\n|---|---------|-------------|-------------------|\n| 01 | [`add`](examples\u002F01_add\u002F) | 1D `ct.load`\u002F`ct.store`, alpha scaling, scalar broadcast, `fill`\u002F`zeros`, keyword load\u002Fstore | Starting point; basic TMA + element-wise patterns |\n| 02 | [`matmul`](examples\u002F02_matmul\u002F) | `muladd`, TF32 conversion, K-loop with `for`, 2D swizzle, standard Julia layout, `ct.@compiler_options` | MMA \u002F tensor core operations |\n| 03 | [`softmax`](examples\u002F03_softmax\u002F) | Persistent scheduling, `for` loops, `gather`\u002F`scatter`, `padding_mode`, multi-pass | Large-tensor reduction patterns |\n\nThese match the released kernels in `julia\u002Fkernels\u002F` (`add.jl`, `matmul.jl`, `softmax.jl`). The examples are simplified teaching versions — always consult `julia\u002Fkernels\u002F*.jl` for the canonical, tested implementations.\n\n## Reference Documents\n\n| Category | Document | Content |\n|----------|----------|---------|\n| **Workflows** | [`translations\u002Fworkflow.md`](translations\u002Fworkflow.md) | Full conversion workflow with todo list, validation loop, checklist |\n| **Rules** | [`references\u002Fcritical-rules.md`](references\u002Fcritical-rules.md) | 17 Critical Rules for cuTile Python → Julia conversion |\n| **API** | [`references\u002Fapi-mapping.md`](references\u002Fapi-mapping.md) | Python↔Julia bidirectional API mapping + kernel patterns |\n| **Testing** | [`references\u002Ftesting.md`](references\u002Ftesting.md) | Julia-native test patterns, tolerances, failure diagnosis |\n| **Debugging** | [`references\u002Fdebugging.md`](references\u002Fdebugging.md) | Julia-specific error diagnosis + IR debug commands |\n| **Scripts** | [`scripts\u002Fvalidate_cutile_jl.py`](scripts\u002Fvalidate_cutile_jl.py) | Static validation for Julia anti-patterns (run it) |\n| **Ground Truth** | `julia\u002Fkernels\u002F*.jl` + `julia\u002Ftest\u002F*.jl` | Actual working implementations in the codebase |\n\n## Environment Setup\n\n**Prerequisite — Julia**: this skill requires the Julia version declared in `julia\u002FProject.toml` under `[compat] julia`. If `julia --version` is missing or older than that, install from the official Julia site at \u003Chttps:\u002F\u002Fjulialang.org\u002Finstall\u002F> following the verified installer instructions for your OS. Resume below once `julia --version` is compatible.\n\nThen, from the repo root:\n\n```bash\n# Install Julia dependencies declared in julia\u002FProject.toml\njulia --project=julia\u002F -e 'using Pkg; Pkg.instantiate()'\n\n# Run tests\njulia --project=julia\u002F julia\u002Ftest\u002Fruntests.jl\n```\n\nRequirements:\n- Julia (minimum version declared in `julia\u002FProject.toml` under `[compat] julia`)\n- CUDA 13.1+ driver\n- Blackwell GPU (compute capability 10+)\n- Dependencies managed via `julia\u002FProject.toml`: CUDA.jl, cuTile.jl, NNlib.jl, Test\n",{"data":35,"body":44},{"name":4,"description":6,"license":26,"metadata":36},{"author":37,"tags":38},"TileGym Team \u003CTileGym@nvidia.com>",[39,40,41,42,43],"cutile","julia","conversion","gpu","kernel",{"type":45,"children":46},"root",[47,56,79,86,195,201,229,241,267,273,432,445,451,466,677,683,712,915,955,961,1181,1187,1237,1242,1335,1340,1382],{"type":48,"tag":49,"props":50,"children":52},"element","h1",{"id":51},"cutile-python-cutilejl-julia-conversion",[53],{"type":54,"value":55},"text","cuTile Python → cuTile.jl (Julia) Conversion",{"type":48,"tag":57,"props":58,"children":59},"p",{},[60,62,69,71,77],{"type":54,"value":61},"Convert ",{"type":48,"tag":63,"props":64,"children":66},"code",{"className":65},[],[67],{"type":54,"value":68},"@ct.kernel",{"type":54,"value":70}," Python kernels to Julia ",{"type":48,"tag":63,"props":72,"children":74},{"className":73},[],[75],{"type":54,"value":76},"function ... end",{"type":54,"value":78}," cuTile.jl kernels.",{"type":48,"tag":80,"props":81,"children":83},"h2",{"id":82},"workflow-selection",[84],{"type":54,"value":85},"Workflow Selection",{"type":48,"tag":87,"props":88,"children":89},"ul",{},[90,112,147,177],{"type":48,"tag":91,"props":92,"children":93},"li",{},[94,100,102],{"type":48,"tag":95,"props":96,"children":97},"strong",{},[98],{"type":54,"value":99},"Standard conversion",{"type":54,"value":101}," → Full workflow: ",{"type":48,"tag":103,"props":104,"children":106},"a",{"href":105},"translations\u002Fworkflow.md",[107],{"type":48,"tag":63,"props":108,"children":110},{"className":109},[],[111],{"type":54,"value":105},{"type":48,"tag":91,"props":113,"children":114},{},[115,120,122,128,130,136,138],{"type":48,"tag":95,"props":116,"children":117},{},[118],{"type":54,"value":119},"Errors",{"type":54,"value":121}," (",{"type":48,"tag":63,"props":123,"children":125},{"className":124},[],[126],{"type":54,"value":127},"MethodError",{"type":54,"value":129},", ",{"type":48,"tag":63,"props":131,"children":133},{"className":132},[],[134],{"type":54,"value":135},"IRError",{"type":54,"value":137},", numerical mismatch) → ",{"type":48,"tag":103,"props":139,"children":141},{"href":140},"references\u002Fdebugging.md",[142],{"type":48,"tag":63,"props":143,"children":145},{"className":144},[],[146],{"type":54,"value":140},{"type":48,"tag":91,"props":148,"children":149},{},[150,155,157,166,168],{"type":48,"tag":95,"props":151,"children":152},{},[153],{"type":54,"value":154},"Quick reference",{"type":54,"value":156}," → ",{"type":48,"tag":103,"props":158,"children":160},{"href":159},"references\u002Fapi-mapping.md",[161],{"type":48,"tag":63,"props":162,"children":164},{"className":163},[],[165],{"type":54,"value":159},{"type":54,"value":167}," + ",{"type":48,"tag":103,"props":169,"children":171},{"href":170},"references\u002Fcritical-rules.md",[172],{"type":48,"tag":63,"props":173,"children":175},{"className":174},[],[176],{"type":54,"value":170},{"type":48,"tag":91,"props":178,"children":179},{},[180,185,186],{"type":48,"tag":95,"props":181,"children":182},{},[183],{"type":54,"value":184},"Test patterns",{"type":54,"value":156},{"type":48,"tag":103,"props":187,"children":189},{"href":188},"references\u002Ftesting.md",[190],{"type":48,"tag":63,"props":191,"children":193},{"className":192},[],[194],{"type":54,"value":188},{"type":48,"tag":80,"props":196,"children":198},{"id":197},"architecture",[199],{"type":54,"value":200},"Architecture",{"type":48,"tag":57,"props":202,"children":203},{},[204,206,211,213,219,221,227],{"type":54,"value":205},"Julia kernels are ",{"type":48,"tag":95,"props":207,"children":208},{},[209],{"type":54,"value":210},"standalone",{"type":54,"value":212}," — no Python bridge, no pytest integration. The Julia sub-project\nlives in ",{"type":48,"tag":63,"props":214,"children":216},{"className":215},[],[217],{"type":54,"value":218},"julia\u002F",{"type":54,"value":220}," at the repo root with its own ",{"type":48,"tag":63,"props":222,"children":224},{"className":223},[],[225],{"type":54,"value":226},"Project.toml",{"type":54,"value":228}," for dependency management.",{"type":48,"tag":230,"props":231,"children":235},"pre",{"className":232,"code":234,"language":54},[233],"language-text","julia\u002F                          # Self-contained Julia sub-project\n├── Project.toml                # Dependencies: CUDA.jl, cuTile.jl, NNlib.jl, Test\n├── kernels\u002F                    # cuTile.jl kernel implementations\n│   ├── add.jl                  # ← Ground-truth: 1D element-wise with alpha scaling (tensor+tensor, tensor+scalar)\n│   ├── matmul.jl               # ← Ground-truth: 2D tiled MMA, standard Julia layout (M,K)×(K,N)→(M,N)\n│   └── softmax.jl              # ← Ground-truth: 3 strategies (TMA, online, chunked) using ct.load\u002Fct.store\n└── test\u002F                       # Julia-native tests (using Test stdlib)\n    ├── runtests.jl             # Test runner entry point\n    ├── test_add.jl\n    ├── test_matmul.jl\n    └── test_softmax.jl\n",[236],{"type":48,"tag":63,"props":237,"children":239},{"__ignoreMap":238},"",[240],{"type":54,"value":234},{"type":48,"tag":57,"props":242,"children":243},{},[244,249,251,257,259,265],{"type":48,"tag":95,"props":245,"children":246},{},[247],{"type":54,"value":248},"Ground-truth reference",{"type":54,"value":250},": Always consult ",{"type":48,"tag":63,"props":252,"children":254},{"className":253},[],[255],{"type":54,"value":256},"julia\u002Fkernels\u002F*.jl",{"type":54,"value":258}," and ",{"type":48,"tag":63,"props":260,"children":262},{"className":261},[],[263],{"type":54,"value":264},"julia\u002Ftest\u002F*.jl",{"type":54,"value":266}," for patterns that compile and pass tests. These are the canonical examples of working cuTile.jl code.",{"type":48,"tag":80,"props":268,"children":270},{"id":269},"instructions",[271],{"type":54,"value":272},"Instructions",{"type":48,"tag":274,"props":275,"children":276},"ol",{},[277,287,305,322,344,377,401,417],{"type":48,"tag":91,"props":278,"children":279},{},[280,285],{"type":48,"tag":95,"props":281,"children":282},{},[283],{"type":54,"value":284},"Analyze",{"type":54,"value":286}," the Python kernel: identify patterns, shapes, dtypes, operations",{"type":48,"tag":91,"props":288,"children":289},{},[290,295,297,303],{"type":48,"tag":95,"props":291,"children":292},{},[293],{"type":54,"value":294},"Write Julia kernel",{"type":54,"value":296}," — ",{"type":48,"tag":63,"props":298,"children":300},{"className":299},[],[301],{"type":54,"value":302},"julia\u002Fkernels\u002F\u003Cop>.jl",{"type":54,"value":304}," with cuTile.jl kernel + bridge function(s)",{"type":48,"tag":91,"props":306,"children":307},{},[308,313,315,320],{"type":48,"tag":95,"props":309,"children":310},{},[311],{"type":54,"value":312},"Convert",{"type":54,"value":314}," kernel signature (see ",{"type":48,"tag":63,"props":316,"children":318},{"className":317},[],[319],{"type":54,"value":105},{"type":54,"value":321}," Phase 2)",{"type":48,"tag":91,"props":323,"children":324},{},[325,329,331,336,337,342],{"type":48,"tag":95,"props":326,"children":327},{},[328],{"type":54,"value":312},{"type":54,"value":330}," kernel body (apply ",{"type":48,"tag":63,"props":332,"children":334},{"className":333},[],[335],{"type":54,"value":159},{"type":54,"value":167},{"type":48,"tag":63,"props":338,"children":340},{"className":339},[],[341],{"type":54,"value":170},{"type":54,"value":343},")",{"type":48,"tag":91,"props":345,"children":346},{},[347,352,353,359,361,367,369,375],{"type":48,"tag":95,"props":348,"children":349},{},[350],{"type":54,"value":351},"Write Julia test",{"type":54,"value":296},{"type":48,"tag":63,"props":354,"children":356},{"className":355},[],[357],{"type":54,"value":358},"julia\u002Ftest\u002Ftest_\u003Cop>.jl",{"type":54,"value":360}," using ",{"type":48,"tag":63,"props":362,"children":364},{"className":363},[],[365],{"type":54,"value":366},"Test",{"type":54,"value":368}," stdlib + ",{"type":48,"tag":63,"props":370,"children":372},{"className":371},[],[373],{"type":54,"value":374},"NNlib.jl",{"type":54,"value":376}," for reference",{"type":48,"tag":91,"props":378,"children":379},{},[380,385,387,393,395],{"type":48,"tag":95,"props":381,"children":382},{},[383],{"type":54,"value":384},"Register test",{"type":54,"value":386}," — add ",{"type":48,"tag":63,"props":388,"children":390},{"className":389},[],[391],{"type":54,"value":392},"include(...)",{"type":54,"value":394}," in ",{"type":48,"tag":63,"props":396,"children":398},{"className":397},[],[399],{"type":54,"value":400},"julia\u002Ftest\u002Fruntests.jl",{"type":48,"tag":91,"props":402,"children":403},{},[404,409,411],{"type":48,"tag":95,"props":405,"children":406},{},[407],{"type":54,"value":408},"Validate",{"type":54,"value":410}," — run the bundled validator: ",{"type":48,"tag":63,"props":412,"children":414},{"className":413},[],[415],{"type":54,"value":416},"python \u003Cskill-dir>\u002Fscripts\u002Fvalidate_cutile_jl.py \u003Cfile.jl>",{"type":48,"tag":91,"props":418,"children":419},{},[420,424,426],{"type":48,"tag":95,"props":421,"children":422},{},[423],{"type":54,"value":366},{"type":54,"value":425}," — run ",{"type":48,"tag":63,"props":427,"children":429},{"className":428},[],[430],{"type":54,"value":431},"julia --project=julia\u002F julia\u002Ftest\u002Fruntests.jl",{"type":48,"tag":57,"props":433,"children":434},{},[435,437],{"type":54,"value":436},"Full conversion checklist with post-conversion verification → ",{"type":48,"tag":103,"props":438,"children":439},{"href":105},[440],{"type":48,"tag":63,"props":441,"children":443},{"className":442},[],[444],{"type":54,"value":105},{"type":48,"tag":80,"props":446,"children":448},{"id":447},"️-top-pitfalls",[449],{"type":54,"value":450},"⚠️ Top Pitfalls",{"type":48,"tag":57,"props":452,"children":453},{},[454,456,464],{"type":54,"value":455},"The most dangerous translation errors. Full rules (17 total) in ",{"type":48,"tag":103,"props":457,"children":458},{"href":170},[459],{"type":48,"tag":63,"props":460,"children":462},{"className":461},[],[463],{"type":54,"value":170},{"type":54,"value":465},".",{"type":48,"tag":467,"props":468,"children":469},"table",{},[470,494],{"type":48,"tag":471,"props":472,"children":473},"thead",{},[474],{"type":48,"tag":475,"props":476,"children":477},"tr",{},[478,484,489],{"type":48,"tag":479,"props":480,"children":481},"th",{},[482],{"type":54,"value":483},"#",{"type":48,"tag":479,"props":485,"children":486},{},[487],{"type":54,"value":488},"Pitfall",{"type":48,"tag":479,"props":490,"children":491},{},[492],{"type":54,"value":493},"One-line fix",{"type":48,"tag":495,"props":496,"children":497},"tbody",{},[498,544,580,616,640],{"type":48,"tag":475,"props":499,"children":500},{},[501,507,518],{"type":48,"tag":502,"props":503,"children":504},"td",{},[505],{"type":54,"value":506},"1",{"type":48,"tag":502,"props":508,"children":509},{},[510,516],{"type":48,"tag":63,"props":511,"children":513},{"className":512},[],[514],{"type":54,"value":515},"ct.full()",{"type":54,"value":517}," doesn't exist in Julia",{"type":48,"tag":502,"props":519,"children":520},{},[521,523,529,530,536,538],{"type":54,"value":522},"Use ",{"type":48,"tag":63,"props":524,"children":526},{"className":525},[],[527],{"type":54,"value":528},"fill(val, shape)",{"type":54,"value":129},{"type":48,"tag":63,"props":531,"children":533},{"className":532},[],[534],{"type":54,"value":535},"zeros(T, dims...)",{"type":54,"value":537},", or ",{"type":48,"tag":63,"props":539,"children":541},{"className":540},[],[542],{"type":54,"value":543},"ones(T, dims...)",{"type":48,"tag":475,"props":545,"children":546},{},[547,552,568],{"type":48,"tag":502,"props":548,"children":549},{},[550],{"type":54,"value":551},"2",{"type":48,"tag":502,"props":553,"children":554},{},[555,561,563],{"type":48,"tag":63,"props":556,"children":558},{"className":557},[],[559],{"type":54,"value":560},"max(a, b)",{"type":54,"value":562}," on tiles → ",{"type":48,"tag":63,"props":564,"children":566},{"className":565},[],[567],{"type":54,"value":135},{"type":48,"tag":502,"props":569,"children":570},{},[571,572,578],{"type":54,"value":522},{"type":48,"tag":63,"props":573,"children":575},{"className":574},[],[576],{"type":54,"value":577},"max.(a, b)",{"type":54,"value":579}," (broadcast dot)",{"type":48,"tag":475,"props":581,"children":582},{},[583,588,611],{"type":48,"tag":502,"props":584,"children":585},{},[586],{"type":54,"value":587},"3",{"type":48,"tag":502,"props":589,"children":590},{},[591,596,598,603,605],{"type":48,"tag":63,"props":592,"children":594},{"className":593},[],[595],{"type":54,"value":135},{"type":54,"value":597}," \u002F ",{"type":48,"tag":63,"props":599,"children":601},{"className":600},[],[602],{"type":54,"value":127},{"type":54,"value":604}," mentioning ",{"type":48,"tag":63,"props":606,"children":608},{"className":607},[],[609],{"type":54,"value":610},"IRStructurizer",{"type":48,"tag":502,"props":612,"children":613},{},[614],{"type":54,"value":615},"Compiler bug — file upstream with minimal reproducer",{"type":48,"tag":475,"props":617,"children":618},{},[619,624,635],{"type":48,"tag":502,"props":620,"children":621},{},[622],{"type":54,"value":623},"4",{"type":48,"tag":502,"props":625,"children":626},{},[627,633],{"type":48,"tag":63,"props":628,"children":630},{"className":629},[],[631],{"type":54,"value":632},"ct.launch",{"type":54,"value":634}," arg order silently wrong",{"type":48,"tag":502,"props":636,"children":637},{},[638],{"type":54,"value":639},"Args are positional — match kernel signature exactly",{"type":48,"tag":475,"props":641,"children":642},{},[643,648,667],{"type":48,"tag":502,"props":644,"children":645},{},[646],{"type":54,"value":647},"5",{"type":48,"tag":502,"props":649,"children":650},{},[651,657,659,665],{"type":48,"tag":63,"props":652,"children":654},{"className":653},[],[655],{"type":54,"value":656},"ct.load",{"type":54,"value":658}," with ",{"type":48,"tag":63,"props":660,"children":662},{"className":661},[],[663],{"type":54,"value":664},"order",{"type":54,"value":666}," — index positions wrong",{"type":48,"tag":502,"props":668,"children":669},{},[670,675],{"type":48,"tag":63,"props":671,"children":673},{"className":672},[],[674],{"type":54,"value":664},{"type":54,"value":676}," remaps BOTH shape AND index (Critical Rule 16)",{"type":48,"tag":80,"props":678,"children":680},{"id":679},"worked-examples",[681],{"type":54,"value":682},"Worked Examples",{"type":48,"tag":57,"props":684,"children":685},{},[686,688,694,696,702,704,710],{"type":54,"value":687},"Side-by-side Python → Julia conversions matching the released Julia kernels in ",{"type":48,"tag":63,"props":689,"children":691},{"className":690},[],[692],{"type":54,"value":693},"julia\u002Fkernels\u002F",{"type":54,"value":695},". Each directory contains ",{"type":48,"tag":63,"props":697,"children":699},{"className":698},[],[700],{"type":54,"value":701},"cutile_python.py",{"type":54,"value":703}," (before) and ",{"type":48,"tag":63,"props":705,"children":707},{"className":706},[],[708],{"type":54,"value":709},"cutile_julia.jl",{"type":54,"value":711}," (after).",{"type":48,"tag":467,"props":713,"children":714},{},[715,740],{"type":48,"tag":471,"props":716,"children":717},{},[718],{"type":48,"tag":475,"props":719,"children":720},{},[721,725,730,735],{"type":48,"tag":479,"props":722,"children":723},{},[724],{"type":54,"value":483},{"type":48,"tag":479,"props":726,"children":727},{},[728],{"type":54,"value":729},"Example",{"type":48,"tag":479,"props":731,"children":732},{},[733],{"type":54,"value":734},"Key Patterns",{"type":48,"tag":479,"props":736,"children":737},{},[738],{"type":54,"value":739},"When to Reference",{"type":48,"tag":495,"props":741,"children":742},{},[743,804,855],{"type":48,"tag":475,"props":744,"children":745},{},[746,751,764,799],{"type":48,"tag":502,"props":747,"children":748},{},[749],{"type":54,"value":750},"01",{"type":48,"tag":502,"props":752,"children":753},{},[754],{"type":48,"tag":103,"props":755,"children":757},{"href":756},"examples\u002F01_add\u002F",[758],{"type":48,"tag":63,"props":759,"children":761},{"className":760},[],[762],{"type":54,"value":763},"add",{"type":48,"tag":502,"props":765,"children":766},{},[767,769,774,776,782,784,790,791,797],{"type":54,"value":768},"1D ",{"type":48,"tag":63,"props":770,"children":772},{"className":771},[],[773],{"type":54,"value":656},{"type":54,"value":775},"\u002F",{"type":48,"tag":63,"props":777,"children":779},{"className":778},[],[780],{"type":54,"value":781},"ct.store",{"type":54,"value":783},", alpha scaling, scalar broadcast, ",{"type":48,"tag":63,"props":785,"children":787},{"className":786},[],[788],{"type":54,"value":789},"fill",{"type":54,"value":775},{"type":48,"tag":63,"props":792,"children":794},{"className":793},[],[795],{"type":54,"value":796},"zeros",{"type":54,"value":798},", keyword load\u002Fstore",{"type":48,"tag":502,"props":800,"children":801},{},[802],{"type":54,"value":803},"Starting point; basic TMA + element-wise patterns",{"type":48,"tag":475,"props":805,"children":806},{},[807,812,825,850],{"type":48,"tag":502,"props":808,"children":809},{},[810],{"type":54,"value":811},"02",{"type":48,"tag":502,"props":813,"children":814},{},[815],{"type":48,"tag":103,"props":816,"children":818},{"href":817},"examples\u002F02_matmul\u002F",[819],{"type":48,"tag":63,"props":820,"children":822},{"className":821},[],[823],{"type":54,"value":824},"matmul",{"type":48,"tag":502,"props":826,"children":827},{},[828,834,836,842,844],{"type":48,"tag":63,"props":829,"children":831},{"className":830},[],[832],{"type":54,"value":833},"muladd",{"type":54,"value":835},", TF32 conversion, K-loop with ",{"type":48,"tag":63,"props":837,"children":839},{"className":838},[],[840],{"type":54,"value":841},"for",{"type":54,"value":843},", 2D swizzle, standard Julia layout, ",{"type":48,"tag":63,"props":845,"children":847},{"className":846},[],[848],{"type":54,"value":849},"ct.@compiler_options",{"type":48,"tag":502,"props":851,"children":852},{},[853],{"type":54,"value":854},"MMA \u002F tensor core operations",{"type":48,"tag":475,"props":856,"children":857},{},[858,863,876,910],{"type":48,"tag":502,"props":859,"children":860},{},[861],{"type":54,"value":862},"03",{"type":48,"tag":502,"props":864,"children":865},{},[866],{"type":48,"tag":103,"props":867,"children":869},{"href":868},"examples\u002F03_softmax\u002F",[870],{"type":48,"tag":63,"props":871,"children":873},{"className":872},[],[874],{"type":54,"value":875},"softmax",{"type":48,"tag":502,"props":877,"children":878},{},[879,881,886,888,894,895,901,902,908],{"type":54,"value":880},"Persistent scheduling, ",{"type":48,"tag":63,"props":882,"children":884},{"className":883},[],[885],{"type":54,"value":841},{"type":54,"value":887}," loops, ",{"type":48,"tag":63,"props":889,"children":891},{"className":890},[],[892],{"type":54,"value":893},"gather",{"type":54,"value":775},{"type":48,"tag":63,"props":896,"children":898},{"className":897},[],[899],{"type":54,"value":900},"scatter",{"type":54,"value":129},{"type":48,"tag":63,"props":903,"children":905},{"className":904},[],[906],{"type":54,"value":907},"padding_mode",{"type":54,"value":909},", multi-pass",{"type":48,"tag":502,"props":911,"children":912},{},[913],{"type":54,"value":914},"Large-tensor reduction patterns",{"type":48,"tag":57,"props":916,"children":917},{},[918,920,925,926,932,933,939,940,946,948,953],{"type":54,"value":919},"These match the released kernels in ",{"type":48,"tag":63,"props":921,"children":923},{"className":922},[],[924],{"type":54,"value":693},{"type":54,"value":121},{"type":48,"tag":63,"props":927,"children":929},{"className":928},[],[930],{"type":54,"value":931},"add.jl",{"type":54,"value":129},{"type":48,"tag":63,"props":934,"children":936},{"className":935},[],[937],{"type":54,"value":938},"matmul.jl",{"type":54,"value":129},{"type":48,"tag":63,"props":941,"children":943},{"className":942},[],[944],{"type":54,"value":945},"softmax.jl",{"type":54,"value":947},"). The examples are simplified teaching versions — always consult ",{"type":48,"tag":63,"props":949,"children":951},{"className":950},[],[952],{"type":54,"value":256},{"type":54,"value":954}," for the canonical, tested implementations.",{"type":48,"tag":80,"props":956,"children":958},{"id":957},"reference-documents",[959],{"type":54,"value":960},"Reference Documents",{"type":48,"tag":467,"props":962,"children":963},{},[964,985],{"type":48,"tag":471,"props":965,"children":966},{},[967],{"type":48,"tag":475,"props":968,"children":969},{},[970,975,980],{"type":48,"tag":479,"props":971,"children":972},{},[973],{"type":54,"value":974},"Category",{"type":48,"tag":479,"props":976,"children":977},{},[978],{"type":54,"value":979},"Document",{"type":48,"tag":479,"props":981,"children":982},{},[983],{"type":54,"value":984},"Content",{"type":48,"tag":495,"props":986,"children":987},{},[988,1015,1042,1069,1096,1123,1151],{"type":48,"tag":475,"props":989,"children":990},{},[991,999,1010],{"type":48,"tag":502,"props":992,"children":993},{},[994],{"type":48,"tag":95,"props":995,"children":996},{},[997],{"type":54,"value":998},"Workflows",{"type":48,"tag":502,"props":1000,"children":1001},{},[1002],{"type":48,"tag":103,"props":1003,"children":1004},{"href":105},[1005],{"type":48,"tag":63,"props":1006,"children":1008},{"className":1007},[],[1009],{"type":54,"value":105},{"type":48,"tag":502,"props":1011,"children":1012},{},[1013],{"type":54,"value":1014},"Full conversion workflow with todo list, validation loop, checklist",{"type":48,"tag":475,"props":1016,"children":1017},{},[1018,1026,1037],{"type":48,"tag":502,"props":1019,"children":1020},{},[1021],{"type":48,"tag":95,"props":1022,"children":1023},{},[1024],{"type":54,"value":1025},"Rules",{"type":48,"tag":502,"props":1027,"children":1028},{},[1029],{"type":48,"tag":103,"props":1030,"children":1031},{"href":170},[1032],{"type":48,"tag":63,"props":1033,"children":1035},{"className":1034},[],[1036],{"type":54,"value":170},{"type":48,"tag":502,"props":1038,"children":1039},{},[1040],{"type":54,"value":1041},"17 Critical Rules for cuTile Python → Julia conversion",{"type":48,"tag":475,"props":1043,"children":1044},{},[1045,1053,1064],{"type":48,"tag":502,"props":1046,"children":1047},{},[1048],{"type":48,"tag":95,"props":1049,"children":1050},{},[1051],{"type":54,"value":1052},"API",{"type":48,"tag":502,"props":1054,"children":1055},{},[1056],{"type":48,"tag":103,"props":1057,"children":1058},{"href":159},[1059],{"type":48,"tag":63,"props":1060,"children":1062},{"className":1061},[],[1063],{"type":54,"value":159},{"type":48,"tag":502,"props":1065,"children":1066},{},[1067],{"type":54,"value":1068},"Python↔Julia bidirectional API mapping + kernel patterns",{"type":48,"tag":475,"props":1070,"children":1071},{},[1072,1080,1091],{"type":48,"tag":502,"props":1073,"children":1074},{},[1075],{"type":48,"tag":95,"props":1076,"children":1077},{},[1078],{"type":54,"value":1079},"Testing",{"type":48,"tag":502,"props":1081,"children":1082},{},[1083],{"type":48,"tag":103,"props":1084,"children":1085},{"href":188},[1086],{"type":48,"tag":63,"props":1087,"children":1089},{"className":1088},[],[1090],{"type":54,"value":188},{"type":48,"tag":502,"props":1092,"children":1093},{},[1094],{"type":54,"value":1095},"Julia-native test patterns, tolerances, failure diagnosis",{"type":48,"tag":475,"props":1097,"children":1098},{},[1099,1107,1118],{"type":48,"tag":502,"props":1100,"children":1101},{},[1102],{"type":48,"tag":95,"props":1103,"children":1104},{},[1105],{"type":54,"value":1106},"Debugging",{"type":48,"tag":502,"props":1108,"children":1109},{},[1110],{"type":48,"tag":103,"props":1111,"children":1112},{"href":140},[1113],{"type":48,"tag":63,"props":1114,"children":1116},{"className":1115},[],[1117],{"type":54,"value":140},{"type":48,"tag":502,"props":1119,"children":1120},{},[1121],{"type":54,"value":1122},"Julia-specific error diagnosis + IR debug commands",{"type":48,"tag":475,"props":1124,"children":1125},{},[1126,1134,1146],{"type":48,"tag":502,"props":1127,"children":1128},{},[1129],{"type":48,"tag":95,"props":1130,"children":1131},{},[1132],{"type":54,"value":1133},"Scripts",{"type":48,"tag":502,"props":1135,"children":1136},{},[1137],{"type":48,"tag":103,"props":1138,"children":1140},{"href":1139},"scripts\u002Fvalidate_cutile_jl.py",[1141],{"type":48,"tag":63,"props":1142,"children":1144},{"className":1143},[],[1145],{"type":54,"value":1139},{"type":48,"tag":502,"props":1147,"children":1148},{},[1149],{"type":54,"value":1150},"Static validation for Julia anti-patterns (run it)",{"type":48,"tag":475,"props":1152,"children":1153},{},[1154,1162,1176],{"type":48,"tag":502,"props":1155,"children":1156},{},[1157],{"type":48,"tag":95,"props":1158,"children":1159},{},[1160],{"type":54,"value":1161},"Ground Truth",{"type":48,"tag":502,"props":1163,"children":1164},{},[1165,1170,1171],{"type":48,"tag":63,"props":1166,"children":1168},{"className":1167},[],[1169],{"type":54,"value":256},{"type":54,"value":167},{"type":48,"tag":63,"props":1172,"children":1174},{"className":1173},[],[1175],{"type":54,"value":264},{"type":48,"tag":502,"props":1177,"children":1178},{},[1179],{"type":54,"value":1180},"Actual working implementations in the codebase",{"type":48,"tag":80,"props":1182,"children":1184},{"id":1183},"environment-setup",[1185],{"type":54,"value":1186},"Environment Setup",{"type":48,"tag":57,"props":1188,"children":1189},{},[1190,1195,1197,1203,1205,1211,1213,1219,1221,1228,1230,1235],{"type":48,"tag":95,"props":1191,"children":1192},{},[1193],{"type":54,"value":1194},"Prerequisite — Julia",{"type":54,"value":1196},": this skill requires the Julia version declared in ",{"type":48,"tag":63,"props":1198,"children":1200},{"className":1199},[],[1201],{"type":54,"value":1202},"julia\u002FProject.toml",{"type":54,"value":1204}," under ",{"type":48,"tag":63,"props":1206,"children":1208},{"className":1207},[],[1209],{"type":54,"value":1210},"[compat] julia",{"type":54,"value":1212},". If ",{"type":48,"tag":63,"props":1214,"children":1216},{"className":1215},[],[1217],{"type":54,"value":1218},"julia --version",{"type":54,"value":1220}," is missing or older than that, install from the official Julia site at ",{"type":48,"tag":103,"props":1222,"children":1226},{"href":1223,"rel":1224},"https:\u002F\u002Fjulialang.org\u002Finstall\u002F",[1225],"nofollow",[1227],{"type":54,"value":1223},{"type":54,"value":1229}," following the verified installer instructions for your OS. Resume below once ",{"type":48,"tag":63,"props":1231,"children":1233},{"className":1232},[],[1234],{"type":54,"value":1218},{"type":54,"value":1236}," is compatible.",{"type":48,"tag":57,"props":1238,"children":1239},{},[1240],{"type":54,"value":1241},"Then, from the repo root:",{"type":48,"tag":230,"props":1243,"children":1247},{"className":1244,"code":1245,"language":1246,"meta":238,"style":238},"language-bash shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","# Install Julia dependencies declared in julia\u002FProject.toml\njulia --project=julia\u002F -e 'using Pkg; Pkg.instantiate()'\n\n# Run tests\njulia --project=julia\u002F julia\u002Ftest\u002Fruntests.jl\n","bash",[1248],{"type":48,"tag":63,"props":1249,"children":1250},{"__ignoreMap":238},[1251,1263,1299,1309,1318],{"type":48,"tag":1252,"props":1253,"children":1256},"span",{"class":1254,"line":1255},"line",1,[1257],{"type":48,"tag":1252,"props":1258,"children":1260},{"style":1259},"--shiki-light:#90A4AE;--shiki-light-font-style:italic;--shiki-default:#546E7A;--shiki-default-font-style:italic;--shiki-dark:#676E95;--shiki-dark-font-style:italic",[1261],{"type":54,"value":1262},"# Install Julia dependencies declared in julia\u002FProject.toml\n",{"type":48,"tag":1252,"props":1264,"children":1266},{"class":1254,"line":1265},2,[1267,1272,1278,1283,1289,1294],{"type":48,"tag":1252,"props":1268,"children":1270},{"style":1269},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[1271],{"type":54,"value":40},{"type":48,"tag":1252,"props":1273,"children":1275},{"style":1274},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[1276],{"type":54,"value":1277}," --project=julia\u002F",{"type":48,"tag":1252,"props":1279,"children":1280},{"style":1274},[1281],{"type":54,"value":1282}," -e",{"type":48,"tag":1252,"props":1284,"children":1286},{"style":1285},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[1287],{"type":54,"value":1288}," '",{"type":48,"tag":1252,"props":1290,"children":1291},{"style":1274},[1292],{"type":54,"value":1293},"using Pkg; Pkg.instantiate()",{"type":48,"tag":1252,"props":1295,"children":1296},{"style":1285},[1297],{"type":54,"value":1298},"'\n",{"type":48,"tag":1252,"props":1300,"children":1302},{"class":1254,"line":1301},3,[1303],{"type":48,"tag":1252,"props":1304,"children":1306},{"emptyLinePlaceholder":1305},true,[1307],{"type":54,"value":1308},"\n",{"type":48,"tag":1252,"props":1310,"children":1312},{"class":1254,"line":1311},4,[1313],{"type":48,"tag":1252,"props":1314,"children":1315},{"style":1259},[1316],{"type":54,"value":1317},"# Run tests\n",{"type":48,"tag":1252,"props":1319,"children":1321},{"class":1254,"line":1320},5,[1322,1326,1330],{"type":48,"tag":1252,"props":1323,"children":1324},{"style":1269},[1325],{"type":54,"value":40},{"type":48,"tag":1252,"props":1327,"children":1328},{"style":1274},[1329],{"type":54,"value":1277},{"type":48,"tag":1252,"props":1331,"children":1332},{"style":1274},[1333],{"type":54,"value":1334}," julia\u002Ftest\u002Fruntests.jl\n",{"type":48,"tag":57,"props":1336,"children":1337},{},[1338],{"type":54,"value":1339},"Requirements:",{"type":48,"tag":87,"props":1341,"children":1342},{},[1343,1360,1365,1370],{"type":48,"tag":91,"props":1344,"children":1345},{},[1346,1348,1353,1354,1359],{"type":54,"value":1347},"Julia (minimum version declared in ",{"type":48,"tag":63,"props":1349,"children":1351},{"className":1350},[],[1352],{"type":54,"value":1202},{"type":54,"value":1204},{"type":48,"tag":63,"props":1355,"children":1357},{"className":1356},[],[1358],{"type":54,"value":1210},{"type":54,"value":343},{"type":48,"tag":91,"props":1361,"children":1362},{},[1363],{"type":54,"value":1364},"CUDA 13.1+ driver",{"type":48,"tag":91,"props":1366,"children":1367},{},[1368],{"type":54,"value":1369},"Blackwell GPU (compute capability 10+)",{"type":48,"tag":91,"props":1371,"children":1372},{},[1373,1375,1380],{"type":54,"value":1374},"Dependencies managed via ",{"type":48,"tag":63,"props":1376,"children":1378},{"className":1377},[],[1379],{"type":54,"value":1202},{"type":54,"value":1381},": CUDA.jl, cuTile.jl, NNlib.jl, Test",{"type":48,"tag":1383,"props":1384,"children":1385},"style",{},[1386],{"type":54,"value":1387},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":1389,"total":1490},[1390,1405,1419,1433,1444,1461,1476],{"slug":1391,"name":1391,"fn":1392,"description":1393,"org":1394,"tags":1395,"stars":23,"repoUrl":24,"updatedAt":1404},"accelerated-computing-cudf","accelerate data processing with cuDF","Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV\u002FParquet I\u002FO, nullable semantics, and multi-GPU DataFrame workloads.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1396,1399,1402,1403],{"name":1397,"slug":1398,"type":15},"Data Analysis","data-analysis",{"name":1400,"slug":1401,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":13,"slug":14,"type":15},"2026-07-14T05:28:43.176466",{"slug":1406,"name":1406,"fn":1407,"description":1408,"org":1409,"tags":1410,"stars":23,"repoUrl":24,"updatedAt":1418},"aiq-deploy","deploy and manage NVIDIA AI-Q infrastructure","Use when asked to install, deploy, run, validate, troubleshoot, or stop NVIDIA AI-Q Blueprint infrastructure.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1411,1414,1417],{"name":1412,"slug":1413,"type":15},"Deployment","deployment",{"name":1415,"slug":1416,"type":15},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":1420,"name":1420,"fn":1421,"description":1422,"org":1423,"tags":1424,"stars":23,"repoUrl":24,"updatedAt":1432},"aiq-research","conduct deep research with AI-Q","Use when asked to run deep research or AI-Q research through a reachable NVIDIA AI-Q Blueprint backend.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1425,1428,1429],{"name":1426,"slug":1427,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":1430,"slug":1431,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":1434,"name":1434,"fn":1435,"description":1436,"org":1437,"tags":1438,"stars":23,"repoUrl":24,"updatedAt":1443},"amc-run-sample-calibration","run AMC sample dataset calibration","Run end-to-end calibration on the shipped sample dataset (sdg_08_2_sample_data_010926.zip) against a running AMC microservice. Use when user says 'test sample dataset', 'run sample calibration', 'verify AMC install', or 'launch and test'.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1439,1440,1441],{"name":1397,"slug":1398,"type":15},{"name":9,"slug":8,"type":15},{"name":1079,"slug":1442,"type":15},"testing","2026-07-17T05:29:03.913266",{"slug":1445,"name":1445,"fn":1446,"description":1447,"org":1448,"tags":1449,"stars":23,"repoUrl":24,"updatedAt":1460},"amc-run-video-calibration","calibrate video datasets with AutoMagicCalib","Calibrate a new dataset from pre-recorded video files via the AutoMagicCalib REST API. Use when user has local MP4s and says 'calibrate my videos', 'run AMC on these videos', or similar. For RTSP\u002Flive streams, use amc-run-rtsp-calibration instead.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1450,1453,1456,1457],{"name":1451,"slug":1452,"type":15},"Automation","automation",{"name":1454,"slug":1455,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":1458,"slug":1459,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":1462,"name":1462,"fn":1463,"description":1464,"org":1465,"tags":1466,"stars":23,"repoUrl":24,"updatedAt":1475},"amc-setup-calibration-stack","deploy AutoMagicCalib microservice with Docker","Launch AutoMagicCalib microservice and web UI from NGC release images via Docker Compose. Use when user says 'deploy auto calibration', 'launch auto calibration', 'launch AMC', 'start MS+UI', or 'set up auto-magic-calib'. Requires NGC API key.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1467,1468,1471,1472],{"name":1412,"slug":1413,"type":15},{"name":1469,"slug":1470,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":1473,"slug":1474,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":1477,"name":1477,"fn":1478,"description":1479,"org":1480,"tags":1481,"stars":23,"repoUrl":24,"updatedAt":1489},"cudaq-guide","develop quantum applications with CUDA-Q","CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1482,1483,1486],{"name":9,"slug":8,"type":15},{"name":1484,"slug":1485,"type":15},"Quantum Computing","quantum-computing",{"name":1487,"slug":1488,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305,{"items":1492,"total":1640},[1493,1511,1527,1538,1550,1563,1574,1588,1599,1608,1622,1631],{"slug":1494,"name":1494,"fn":1495,"description":1496,"org":1497,"tags":1498,"stars":1508,"repoUrl":1509,"updatedAt":1510},"nemoclaw-user-guide","retrieve NemoClaw documentation and configuration","Guides human users' AI agents to the NemoClaw docs MCP server and canonical Fern documentation in Markdown form. Use when users ask how to install, configure, operate, troubleshoot, secure, or learn NemoClaw with an AI coding assistant. Trigger keywords - nemoclaw docs, use nemoclaw with ai agent, nemoclaw mcp docs, nemoclaw install help, nemoclaw quickstart, nemoclaw markdown docs, llms.txt, agent skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1499,1502,1505],{"name":1500,"slug":1501,"type":15},"Documentation","documentation",{"name":1503,"slug":1504,"type":15},"MCP","mcp",{"name":1506,"slug":1507,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1512,"name":1512,"fn":1513,"description":1514,"org":1515,"tags":1516,"stars":1524,"repoUrl":1525,"updatedAt":1526},"mcore-build-and-dependency","manage Megatron-LM development environments","Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1517,1520,1521],{"name":1518,"slug":1519,"type":15},"Containers","containers",{"name":1412,"slug":1413,"type":15},{"name":1522,"slug":1523,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1528,"name":1528,"fn":1529,"description":1530,"org":1531,"tags":1532,"stars":1524,"repoUrl":1525,"updatedAt":1537},"mcore-bump-base-image","update NVIDIA PyTorch base images","Bump the NVIDIA PyTorch base image (`nvcr.io\u002Fnvidia\u002Fpytorch:YY.MM-py3`) used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker\u002F.ngc_version.dev` and GitLab CI in `.gitlab\u002Fstages\u002F01.build.yml`), the post-bump CI loop (re-run functional tests, refresh golden values, mark broken tests), and the gotchas that bit PRs",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1533,1536],{"name":1534,"slug":1535,"type":15},"CI\u002FCD","ci-cd",{"name":1412,"slug":1413,"type":15},"2026-07-14T05:25:59.97109",{"slug":1539,"name":1539,"fn":1540,"description":1541,"org":1542,"tags":1543,"stars":1524,"repoUrl":1525,"updatedAt":1549},"mcore-cicd","manage CI\u002FCD pipelines for Megatron-LM","CI\u002FCD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI (which force-pushes the current branch to a pull-request\u002FBRANCH ref — always dry-run and verify the destination first; never run against shared or protected branches), and CI failure investigation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1544,1545,1546],{"name":1534,"slug":1535,"type":15},{"name":1412,"slug":1413,"type":15},{"name":1547,"slug":1548,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1551,"name":1551,"fn":1552,"description":1553,"org":1554,"tags":1555,"stars":1524,"repoUrl":1525,"updatedAt":1562},"mcore-create-issue","investigate CI failures and create issues","Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1556,1558,1559],{"name":1106,"slug":1557,"type":15},"debugging",{"name":1547,"slug":1548,"type":15},{"name":1560,"slug":1561,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1564,"name":1564,"fn":1565,"description":1566,"org":1567,"tags":1568,"stars":1524,"repoUrl":1525,"updatedAt":1573},"mcore-linting-and-formatting","lint and format Megatron-LM code","Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1569,1572],{"name":1570,"slug":1571,"type":15},"Best Practices","best-practices",{"name":21,"slug":22,"type":15},"2026-07-14T05:25:56.18433",{"slug":1575,"name":1575,"fn":1576,"description":1577,"org":1578,"tags":1579,"stars":1524,"repoUrl":1525,"updatedAt":1587},"mcore-migrate-gpt-to-hybrid","migrate Megatron-LM models to HybridModel","Migration guide for moving Megatron Core GPTModel checkpoints, model providers, training commands, and layer mappings to HybridModel.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1580,1583,1586],{"name":1581,"slug":1582,"type":15},"Machine Learning","machine-learning",{"name":1584,"slug":1585,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1589,"name":1589,"fn":1590,"description":1591,"org":1592,"tags":1593,"stars":1524,"repoUrl":1525,"updatedAt":1598},"mcore-onboard-gb200-1node-tests","onboard functional tests for GB200","Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1594,1597],{"name":1595,"slug":1596,"type":15},"QA","qa",{"name":1079,"slug":1442,"type":15},"2026-07-14T05:25:53.673039",{"slug":1600,"name":1600,"fn":1601,"description":1602,"org":1603,"tags":1604,"stars":1524,"repoUrl":1525,"updatedAt":1607},"mcore-run-on-slurm","launch distributed training jobs on SLURM","How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1605,1606],{"name":1412,"slug":1413,"type":15},{"name":1415,"slug":1416,"type":15},"2026-07-14T05:25:49.362534",{"slug":1609,"name":1609,"fn":1610,"description":1611,"org":1612,"tags":1613,"stars":1524,"repoUrl":1525,"updatedAt":1621},"mcore-split-pr","split pull requests to reduce review load","Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1614,1617,1618],{"name":1615,"slug":1616,"type":15},"Code Review","code-review",{"name":1547,"slug":1548,"type":15},{"name":1619,"slug":1620,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1623,"name":1623,"fn":1624,"description":1625,"org":1626,"tags":1627,"stars":1524,"repoUrl":1525,"updatedAt":1630},"mcore-testing","run and manage Megatron-LM tests","Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1628,1629],{"name":1595,"slug":1596,"type":15},{"name":1079,"slug":1442,"type":15},"2026-07-14T05:25:54.928983",{"slug":1632,"name":1632,"fn":1633,"description":1634,"org":1635,"tags":1636,"stars":1524,"repoUrl":1525,"updatedAt":1639},"nightly-sync","manage nightly main-to-dev sync workflows","Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1637,1638],{"name":1451,"slug":1452,"type":15},{"name":1534,"slug":1535,"type":15},"2026-07-30T05:29:03.275638",496]