[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-cuopt-numerical-optimization-api":3,"mdc--ewqrb1-key":37,"related-repo-nvidia-cuopt-numerical-optimization-api":920,"related-org-nvidia-cuopt-numerical-optimization-api":1026},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":26,"repoUrl":27,"updatedAt":28,"license":29,"forks":30,"topics":31,"repo":32,"sourceUrl":35,"mdContent":36},"cuopt-numerical-optimization-api","solve numerical optimization problems with cuOpt","LP, MILP, and QP (beta) with cuOpt — Python, C, and CLI. Use when the user is solving LP, MILP, or QP with any cuOpt interface.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,20,23],{"name":13,"slug":14,"type":15},"Mathematics","mathematics","tag",{"name":17,"slug":18,"type":15},"Optimization","optimization",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},"Python","python",{"name":24,"slug":25,"type":15},"API Development","api-development",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:30:16.593088","Apache-2.0",281,[],{"repoUrl":27,"stars":26,"forks":30,"topics":33,"description":34},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Fcuopt-numerical-optimization-api","---\nname: cuopt-numerical-optimization-api\nversion: \"26.08.00\"\ndescription: LP, MILP, and QP (beta) with cuOpt — Python, C, and CLI. Use when the user is solving LP, MILP, or QP with any cuOpt interface.\nlicense: Apache-2.0\nmetadata:\n  author: NVIDIA cuOpt Team\n  tags:\n    - cuopt\n    - linear-programming\n    - milp\n    - qp\n    - python\n    - c-api\n    - cli\n---\n\n\n# cuOpt Numerical Optimization API\n\nModel and solve LP, MILP, and QP problems using NVIDIA cuOpt's GPU-accelerated solver.\n\n## Interface Selection\n\nChoose the reference for the user's interface:\n\n| Interface | When to use | Reference |\n|-----------|-------------|-----------|\n| **Python** | User is writing Python code | [references\u002Fpython_api.md](references\u002Fpython_api.md) |\n| **C \u002F C++** | User is embedding in a C\u002FC++ application | [references\u002Fc_api.md](references\u002Fc_api.md) |\n| **CLI** | User is solving from MPS files on the command line | [references\u002Fcli_api.md](references\u002Fcli_api.md) |\n\nIf the interface is not yet clear, ask before writing any code.\n\n**Already using a modeling language?** cuOpt also works as a solver backend for third-party\nmodeling tools — **AMPL, GAMS \u002F GAMSPy, PuLP, JuMP, Pyomo, and CVXPY** — with near-zero code\nchanges (point the model's solver at cuOpt). CVXPY additionally covers convex QP and, in beta,\nQCQP \u002F SOCP. Prefer this when the user already has a model in one of these tools rather than porting\nit to the cuOpt API. See\n[Third-Party Modeling Languages](https:\u002F\u002Fdocs.nvidia.com\u002Fcuopt\u002Fuser-guide\u002Flatest\u002Fthirdparty_modeling_languages\u002Findex.html).\n\n## Choosing LP vs MILP vs QP\n\n**Decide from the objective and variables:**\n\n| If the objective is... | And variables are... | Use |\n|---|---|---|\n| Linear (sum of `c_i * x_i`) | All continuous | **LP** |\n| Linear | Some integer or binary | **MILP** |\n| Has squared (`x*x`) or cross (`x*y`) terms | Continuous (integer QP not supported) | **QP** (beta) |\n\n**Prefer LP when the problem allows it.** LP solves faster and has stronger optimality guarantees. Use MILP only when the problem logically requires whole numbers or yes\u002Fno decisions. Use QP only when the objective is genuinely quadratic (variance, squared error, kinetic energy).\n\n- **Use LP** when every quantity can meaningfully be fractional: flows, proportions, rates, dollars, hours, tonnes of material, etc.\n- **Use MILP** when the problem mentions **counts** of discrete entities, **yes\u002Fno** choices, or **either\u002For** decisions (e.g. open a facility or not, assign a person to a shift, number of trucks).\n- **Use QP** when the objective minimizes variance, squared error, or any expression with `x*x` or `x*y` terms (portfolio optimization, least squares, regularized regression).\n\n## Integer vs Continuous from Wording\n\n| Problem wording \u002F concept | Variable type | Examples |\n|---------------------------|---------------|----------|\n| **Discrete entities (counts)** | **INTEGER** | Workers, cars, trucks, machines, pilots, facilities, units to manufacture |\n| **Yes\u002Fno or on\u002Foff** | **INTEGER** (binary, lb=0 ub=1) | Open a facility, run a machine, assign a person to a shift |\n| **Amounts that can be fractional** | **CONTINUOUS** | Tonnes, litres, dollars, hours, kWh, proportion of capacity |\n| **Rates or fractions** | **CONTINUOUS** | Utilization, percentage, share of budget |\n\n**Rule of thumb:** \"How many *things*\" → INTEGER. \"How much\" → CONTINUOUS.\n\n## QP Rules (all interfaces)\n\n- **MINIMIZE only** — the solver rejects MAXIMIZE for quadratic objectives. To maximize `f(x)`, minimize `-f(x)` and negate the reported objective value.\n- **Continuous variables only** — integer QP is not supported.\n- **Q should be positive semi-definite** for a convex, well-posed problem.\n- **Beta** — API may evolve; treat as production-capable for typical convex QP.\n\n## Dual Values\n\nDuals and reduced costs are available for **LP and QP only**:\n- **MILP** — no duals (integer optima are not continuous).\n- **Quadratic constraints** — duals unavailable even for LP\u002FQP; all values return `NaN`.\n- **PDLP warmstart** — LP only; MILP solves do not accept a PDLP warmstart.\n\n## Common Issues (all interfaces)\n\n| Problem | Likely cause | Fix |\n|---------|-------------|-----|\n| Infeasible | Conflicting constraints | Check constraint logic and bounds |\n| Unbounded | Missing bounds | Add variable bounds |\n| Slow solve | Large problem | Set time limit; increase gap tolerance |\n| QP rejected with MAXIMIZE | QP only supports MINIMIZE | Negate the objective; negate the result |\n| QP returns non-optimal | Q not PSD or badly scaled | Check Q is PSD; rescale variables |\n\n## Solver Settings (concepts)\n\n| Setting | Purpose |\n|---------|---------|\n| `time_limit` | Stop after N seconds |\n| `mip_relative_gap` | Stop MILP when within X% of optimal |\n| `mip_absolute_tolerance` | Absolute MIP gap stop |\n| `log_to_console` | Enable solver logging |\n\nSyntax varies by interface — see the interface reference file.\n",{"data":38,"body":49},{"name":4,"version":39,"description":6,"license":29,"metadata":40},"26.08.00",{"author":41,"tags":42},"NVIDIA cuOpt Team",[43,44,45,46,22,47,48],"cuopt","linear-programming","milp","qp","c-api","cli",{"type":50,"children":51},"root",[52,60,66,73,78,184,189,216,222,230,347,357,427,433,556,574,580,639,645,657,696,702,819,825,915],{"type":53,"tag":54,"props":55,"children":56},"element","h1",{"id":4},[57],{"type":58,"value":59},"text","cuOpt Numerical Optimization API",{"type":53,"tag":61,"props":62,"children":63},"p",{},[64],{"type":58,"value":65},"Model and solve LP, MILP, and QP problems using NVIDIA cuOpt's GPU-accelerated solver.",{"type":53,"tag":67,"props":68,"children":70},"h2",{"id":69},"interface-selection",[71],{"type":58,"value":72},"Interface Selection",{"type":53,"tag":61,"props":74,"children":75},{},[76],{"type":58,"value":77},"Choose the reference for the user's interface:",{"type":53,"tag":79,"props":80,"children":81},"table",{},[82,106],{"type":53,"tag":83,"props":84,"children":85},"thead",{},[86],{"type":53,"tag":87,"props":88,"children":89},"tr",{},[90,96,101],{"type":53,"tag":91,"props":92,"children":93},"th",{},[94],{"type":58,"value":95},"Interface",{"type":53,"tag":91,"props":97,"children":98},{},[99],{"type":58,"value":100},"When to use",{"type":53,"tag":91,"props":102,"children":103},{},[104],{"type":58,"value":105},"Reference",{"type":53,"tag":107,"props":108,"children":109},"tbody",{},[110,136,160],{"type":53,"tag":87,"props":111,"children":112},{},[113,122,127],{"type":53,"tag":114,"props":115,"children":116},"td",{},[117],{"type":53,"tag":118,"props":119,"children":120},"strong",{},[121],{"type":58,"value":21},{"type":53,"tag":114,"props":123,"children":124},{},[125],{"type":58,"value":126},"User is writing Python code",{"type":53,"tag":114,"props":128,"children":129},{},[130],{"type":53,"tag":131,"props":132,"children":134},"a",{"href":133},"references\u002Fpython_api.md",[135],{"type":58,"value":133},{"type":53,"tag":87,"props":137,"children":138},{},[139,147,152],{"type":53,"tag":114,"props":140,"children":141},{},[142],{"type":53,"tag":118,"props":143,"children":144},{},[145],{"type":58,"value":146},"C \u002F C++",{"type":53,"tag":114,"props":148,"children":149},{},[150],{"type":58,"value":151},"User is embedding in a C\u002FC++ application",{"type":53,"tag":114,"props":153,"children":154},{},[155],{"type":53,"tag":131,"props":156,"children":158},{"href":157},"references\u002Fc_api.md",[159],{"type":58,"value":157},{"type":53,"tag":87,"props":161,"children":162},{},[163,171,176],{"type":53,"tag":114,"props":164,"children":165},{},[166],{"type":53,"tag":118,"props":167,"children":168},{},[169],{"type":58,"value":170},"CLI",{"type":53,"tag":114,"props":172,"children":173},{},[174],{"type":58,"value":175},"User is solving from MPS files on the command line",{"type":53,"tag":114,"props":177,"children":178},{},[179],{"type":53,"tag":131,"props":180,"children":182},{"href":181},"references\u002Fcli_api.md",[183],{"type":58,"value":181},{"type":53,"tag":61,"props":185,"children":186},{},[187],{"type":58,"value":188},"If the interface is not yet clear, ask before writing any code.",{"type":53,"tag":61,"props":190,"children":191},{},[192,197,199,204,206,214],{"type":53,"tag":118,"props":193,"children":194},{},[195],{"type":58,"value":196},"Already using a modeling language?",{"type":58,"value":198}," cuOpt also works as a solver backend for third-party\nmodeling tools — ",{"type":53,"tag":118,"props":200,"children":201},{},[202],{"type":58,"value":203},"AMPL, GAMS \u002F GAMSPy, PuLP, JuMP, Pyomo, and CVXPY",{"type":58,"value":205}," — with near-zero code\nchanges (point the model's solver at cuOpt). CVXPY additionally covers convex QP and, in beta,\nQCQP \u002F SOCP. Prefer this when the user already has a model in one of these tools rather than porting\nit to the cuOpt API. See\n",{"type":53,"tag":131,"props":207,"children":211},{"href":208,"rel":209},"https:\u002F\u002Fdocs.nvidia.com\u002Fcuopt\u002Fuser-guide\u002Flatest\u002Fthirdparty_modeling_languages\u002Findex.html",[210],"nofollow",[212],{"type":58,"value":213},"Third-Party Modeling Languages",{"type":58,"value":215},".",{"type":53,"tag":67,"props":217,"children":219},{"id":218},"choosing-lp-vs-milp-vs-qp",[220],{"type":58,"value":221},"Choosing LP vs MILP vs QP",{"type":53,"tag":61,"props":223,"children":224},{},[225],{"type":53,"tag":118,"props":226,"children":227},{},[228],{"type":58,"value":229},"Decide from the objective and variables:",{"type":53,"tag":79,"props":231,"children":232},{},[233,254],{"type":53,"tag":83,"props":234,"children":235},{},[236],{"type":53,"tag":87,"props":237,"children":238},{},[239,244,249],{"type":53,"tag":91,"props":240,"children":241},{},[242],{"type":58,"value":243},"If the objective is...",{"type":53,"tag":91,"props":245,"children":246},{},[247],{"type":58,"value":248},"And variables are...",{"type":53,"tag":91,"props":250,"children":251},{},[252],{"type":58,"value":253},"Use",{"type":53,"tag":107,"props":255,"children":256},{},[257,287,308],{"type":53,"tag":87,"props":258,"children":259},{},[260,274,279],{"type":53,"tag":114,"props":261,"children":262},{},[263,265,272],{"type":58,"value":264},"Linear (sum of ",{"type":53,"tag":266,"props":267,"children":269},"code",{"className":268},[],[270],{"type":58,"value":271},"c_i * x_i",{"type":58,"value":273},")",{"type":53,"tag":114,"props":275,"children":276},{},[277],{"type":58,"value":278},"All continuous",{"type":53,"tag":114,"props":280,"children":281},{},[282],{"type":53,"tag":118,"props":283,"children":284},{},[285],{"type":58,"value":286},"LP",{"type":53,"tag":87,"props":288,"children":289},{},[290,295,300],{"type":53,"tag":114,"props":291,"children":292},{},[293],{"type":58,"value":294},"Linear",{"type":53,"tag":114,"props":296,"children":297},{},[298],{"type":58,"value":299},"Some integer or binary",{"type":53,"tag":114,"props":301,"children":302},{},[303],{"type":53,"tag":118,"props":304,"children":305},{},[306],{"type":58,"value":307},"MILP",{"type":53,"tag":87,"props":309,"children":310},{},[311,332,337],{"type":53,"tag":114,"props":312,"children":313},{},[314,316,322,324,330],{"type":58,"value":315},"Has squared (",{"type":53,"tag":266,"props":317,"children":319},{"className":318},[],[320],{"type":58,"value":321},"x*x",{"type":58,"value":323},") or cross (",{"type":53,"tag":266,"props":325,"children":327},{"className":326},[],[328],{"type":58,"value":329},"x*y",{"type":58,"value":331},") terms",{"type":53,"tag":114,"props":333,"children":334},{},[335],{"type":58,"value":336},"Continuous (integer QP not supported)",{"type":53,"tag":114,"props":338,"children":339},{},[340,345],{"type":53,"tag":118,"props":341,"children":342},{},[343],{"type":58,"value":344},"QP",{"type":58,"value":346}," (beta)",{"type":53,"tag":61,"props":348,"children":349},{},[350,355],{"type":53,"tag":118,"props":351,"children":352},{},[353],{"type":58,"value":354},"Prefer LP when the problem allows it.",{"type":58,"value":356}," LP solves faster and has stronger optimality guarantees. Use MILP only when the problem logically requires whole numbers or yes\u002Fno decisions. Use QP only when the objective is genuinely quadratic (variance, squared error, kinetic energy).",{"type":53,"tag":358,"props":359,"children":360},"ul",{},[361,372,403],{"type":53,"tag":362,"props":363,"children":364},"li",{},[365,370],{"type":53,"tag":118,"props":366,"children":367},{},[368],{"type":58,"value":369},"Use LP",{"type":58,"value":371}," when every quantity can meaningfully be fractional: flows, proportions, rates, dollars, hours, tonnes of material, etc.",{"type":53,"tag":362,"props":373,"children":374},{},[375,380,382,387,389,394,396,401],{"type":53,"tag":118,"props":376,"children":377},{},[378],{"type":58,"value":379},"Use MILP",{"type":58,"value":381}," when the problem mentions ",{"type":53,"tag":118,"props":383,"children":384},{},[385],{"type":58,"value":386},"counts",{"type":58,"value":388}," of discrete entities, ",{"type":53,"tag":118,"props":390,"children":391},{},[392],{"type":58,"value":393},"yes\u002Fno",{"type":58,"value":395}," choices, or ",{"type":53,"tag":118,"props":397,"children":398},{},[399],{"type":58,"value":400},"either\u002For",{"type":58,"value":402}," decisions (e.g. open a facility or not, assign a person to a shift, number of trucks).",{"type":53,"tag":362,"props":404,"children":405},{},[406,411,413,418,420,425],{"type":53,"tag":118,"props":407,"children":408},{},[409],{"type":58,"value":410},"Use QP",{"type":58,"value":412}," when the objective minimizes variance, squared error, or any expression with ",{"type":53,"tag":266,"props":414,"children":416},{"className":415},[],[417],{"type":58,"value":321},{"type":58,"value":419}," or ",{"type":53,"tag":266,"props":421,"children":423},{"className":422},[],[424],{"type":58,"value":329},{"type":58,"value":426}," terms (portfolio optimization, least squares, regularized regression).",{"type":53,"tag":67,"props":428,"children":430},{"id":429},"integer-vs-continuous-from-wording",[431],{"type":58,"value":432},"Integer vs Continuous from Wording",{"type":53,"tag":79,"props":434,"children":435},{},[436,457],{"type":53,"tag":83,"props":437,"children":438},{},[439],{"type":53,"tag":87,"props":440,"children":441},{},[442,447,452],{"type":53,"tag":91,"props":443,"children":444},{},[445],{"type":58,"value":446},"Problem wording \u002F concept",{"type":53,"tag":91,"props":448,"children":449},{},[450],{"type":58,"value":451},"Variable type",{"type":53,"tag":91,"props":453,"children":454},{},[455],{"type":58,"value":456},"Examples",{"type":53,"tag":107,"props":458,"children":459},{},[460,484,509,533],{"type":53,"tag":87,"props":461,"children":462},{},[463,471,479],{"type":53,"tag":114,"props":464,"children":465},{},[466],{"type":53,"tag":118,"props":467,"children":468},{},[469],{"type":58,"value":470},"Discrete entities (counts)",{"type":53,"tag":114,"props":472,"children":473},{},[474],{"type":53,"tag":118,"props":475,"children":476},{},[477],{"type":58,"value":478},"INTEGER",{"type":53,"tag":114,"props":480,"children":481},{},[482],{"type":58,"value":483},"Workers, cars, trucks, machines, pilots, facilities, units to manufacture",{"type":53,"tag":87,"props":485,"children":486},{},[487,495,504],{"type":53,"tag":114,"props":488,"children":489},{},[490],{"type":53,"tag":118,"props":491,"children":492},{},[493],{"type":58,"value":494},"Yes\u002Fno or on\u002Foff",{"type":53,"tag":114,"props":496,"children":497},{},[498,502],{"type":53,"tag":118,"props":499,"children":500},{},[501],{"type":58,"value":478},{"type":58,"value":503}," (binary, lb=0 ub=1)",{"type":53,"tag":114,"props":505,"children":506},{},[507],{"type":58,"value":508},"Open a facility, run a machine, assign a person to a shift",{"type":53,"tag":87,"props":510,"children":511},{},[512,520,528],{"type":53,"tag":114,"props":513,"children":514},{},[515],{"type":53,"tag":118,"props":516,"children":517},{},[518],{"type":58,"value":519},"Amounts that can be fractional",{"type":53,"tag":114,"props":521,"children":522},{},[523],{"type":53,"tag":118,"props":524,"children":525},{},[526],{"type":58,"value":527},"CONTINUOUS",{"type":53,"tag":114,"props":529,"children":530},{},[531],{"type":58,"value":532},"Tonnes, litres, dollars, hours, kWh, proportion of capacity",{"type":53,"tag":87,"props":534,"children":535},{},[536,544,551],{"type":53,"tag":114,"props":537,"children":538},{},[539],{"type":53,"tag":118,"props":540,"children":541},{},[542],{"type":58,"value":543},"Rates or fractions",{"type":53,"tag":114,"props":545,"children":546},{},[547],{"type":53,"tag":118,"props":548,"children":549},{},[550],{"type":58,"value":527},{"type":53,"tag":114,"props":552,"children":553},{},[554],{"type":58,"value":555},"Utilization, percentage, share of budget",{"type":53,"tag":61,"props":557,"children":558},{},[559,564,566,572],{"type":53,"tag":118,"props":560,"children":561},{},[562],{"type":58,"value":563},"Rule of thumb:",{"type":58,"value":565}," \"How many ",{"type":53,"tag":567,"props":568,"children":569},"em",{},[570],{"type":58,"value":571},"things",{"type":58,"value":573},"\" → INTEGER. \"How much\" → CONTINUOUS.",{"type":53,"tag":67,"props":575,"children":577},{"id":576},"qp-rules-all-interfaces",[578],{"type":58,"value":579},"QP Rules (all interfaces)",{"type":53,"tag":358,"props":581,"children":582},{},[583,609,619,629],{"type":53,"tag":362,"props":584,"children":585},{},[586,591,593,599,601,607],{"type":53,"tag":118,"props":587,"children":588},{},[589],{"type":58,"value":590},"MINIMIZE only",{"type":58,"value":592}," — the solver rejects MAXIMIZE for quadratic objectives. To maximize ",{"type":53,"tag":266,"props":594,"children":596},{"className":595},[],[597],{"type":58,"value":598},"f(x)",{"type":58,"value":600},", minimize ",{"type":53,"tag":266,"props":602,"children":604},{"className":603},[],[605],{"type":58,"value":606},"-f(x)",{"type":58,"value":608}," and negate the reported objective value.",{"type":53,"tag":362,"props":610,"children":611},{},[612,617],{"type":53,"tag":118,"props":613,"children":614},{},[615],{"type":58,"value":616},"Continuous variables only",{"type":58,"value":618}," — integer QP is not supported.",{"type":53,"tag":362,"props":620,"children":621},{},[622,627],{"type":53,"tag":118,"props":623,"children":624},{},[625],{"type":58,"value":626},"Q should be positive semi-definite",{"type":58,"value":628}," for a convex, well-posed problem.",{"type":53,"tag":362,"props":630,"children":631},{},[632,637],{"type":53,"tag":118,"props":633,"children":634},{},[635],{"type":58,"value":636},"Beta",{"type":58,"value":638}," — API may evolve; treat as production-capable for typical convex QP.",{"type":53,"tag":67,"props":640,"children":642},{"id":641},"dual-values",[643],{"type":58,"value":644},"Dual Values",{"type":53,"tag":61,"props":646,"children":647},{},[648,650,655],{"type":58,"value":649},"Duals and reduced costs are available for ",{"type":53,"tag":118,"props":651,"children":652},{},[653],{"type":58,"value":654},"LP and QP only",{"type":58,"value":656},":",{"type":53,"tag":358,"props":658,"children":659},{},[660,669,686],{"type":53,"tag":362,"props":661,"children":662},{},[663,667],{"type":53,"tag":118,"props":664,"children":665},{},[666],{"type":58,"value":307},{"type":58,"value":668}," — no duals (integer optima are not continuous).",{"type":53,"tag":362,"props":670,"children":671},{},[672,677,679,685],{"type":53,"tag":118,"props":673,"children":674},{},[675],{"type":58,"value":676},"Quadratic constraints",{"type":58,"value":678}," — duals unavailable even for LP\u002FQP; all values return ",{"type":53,"tag":266,"props":680,"children":682},{"className":681},[],[683],{"type":58,"value":684},"NaN",{"type":58,"value":215},{"type":53,"tag":362,"props":687,"children":688},{},[689,694],{"type":53,"tag":118,"props":690,"children":691},{},[692],{"type":58,"value":693},"PDLP warmstart",{"type":58,"value":695}," — LP only; MILP solves do not accept a PDLP warmstart.",{"type":53,"tag":67,"props":697,"children":699},{"id":698},"common-issues-all-interfaces",[700],{"type":58,"value":701},"Common Issues (all interfaces)",{"type":53,"tag":79,"props":703,"children":704},{},[705,726],{"type":53,"tag":83,"props":706,"children":707},{},[708],{"type":53,"tag":87,"props":709,"children":710},{},[711,716,721],{"type":53,"tag":91,"props":712,"children":713},{},[714],{"type":58,"value":715},"Problem",{"type":53,"tag":91,"props":717,"children":718},{},[719],{"type":58,"value":720},"Likely cause",{"type":53,"tag":91,"props":722,"children":723},{},[724],{"type":58,"value":725},"Fix",{"type":53,"tag":107,"props":727,"children":728},{},[729,747,765,783,801],{"type":53,"tag":87,"props":730,"children":731},{},[732,737,742],{"type":53,"tag":114,"props":733,"children":734},{},[735],{"type":58,"value":736},"Infeasible",{"type":53,"tag":114,"props":738,"children":739},{},[740],{"type":58,"value":741},"Conflicting constraints",{"type":53,"tag":114,"props":743,"children":744},{},[745],{"type":58,"value":746},"Check constraint logic and bounds",{"type":53,"tag":87,"props":748,"children":749},{},[750,755,760],{"type":53,"tag":114,"props":751,"children":752},{},[753],{"type":58,"value":754},"Unbounded",{"type":53,"tag":114,"props":756,"children":757},{},[758],{"type":58,"value":759},"Missing bounds",{"type":53,"tag":114,"props":761,"children":762},{},[763],{"type":58,"value":764},"Add variable bounds",{"type":53,"tag":87,"props":766,"children":767},{},[768,773,778],{"type":53,"tag":114,"props":769,"children":770},{},[771],{"type":58,"value":772},"Slow solve",{"type":53,"tag":114,"props":774,"children":775},{},[776],{"type":58,"value":777},"Large problem",{"type":53,"tag":114,"props":779,"children":780},{},[781],{"type":58,"value":782},"Set time limit; increase gap tolerance",{"type":53,"tag":87,"props":784,"children":785},{},[786,791,796],{"type":53,"tag":114,"props":787,"children":788},{},[789],{"type":58,"value":790},"QP rejected with MAXIMIZE",{"type":53,"tag":114,"props":792,"children":793},{},[794],{"type":58,"value":795},"QP only supports MINIMIZE",{"type":53,"tag":114,"props":797,"children":798},{},[799],{"type":58,"value":800},"Negate the objective; negate the result",{"type":53,"tag":87,"props":802,"children":803},{},[804,809,814],{"type":53,"tag":114,"props":805,"children":806},{},[807],{"type":58,"value":808},"QP returns non-optimal",{"type":53,"tag":114,"props":810,"children":811},{},[812],{"type":58,"value":813},"Q not PSD or badly scaled",{"type":53,"tag":114,"props":815,"children":816},{},[817],{"type":58,"value":818},"Check Q is PSD; rescale variables",{"type":53,"tag":67,"props":820,"children":822},{"id":821},"solver-settings-concepts",[823],{"type":58,"value":824},"Solver Settings (concepts)",{"type":53,"tag":79,"props":826,"children":827},{},[828,844],{"type":53,"tag":83,"props":829,"children":830},{},[831],{"type":53,"tag":87,"props":832,"children":833},{},[834,839],{"type":53,"tag":91,"props":835,"children":836},{},[837],{"type":58,"value":838},"Setting",{"type":53,"tag":91,"props":840,"children":841},{},[842],{"type":58,"value":843},"Purpose",{"type":53,"tag":107,"props":845,"children":846},{},[847,864,881,898],{"type":53,"tag":87,"props":848,"children":849},{},[850,859],{"type":53,"tag":114,"props":851,"children":852},{},[853],{"type":53,"tag":266,"props":854,"children":856},{"className":855},[],[857],{"type":58,"value":858},"time_limit",{"type":53,"tag":114,"props":860,"children":861},{},[862],{"type":58,"value":863},"Stop after N seconds",{"type":53,"tag":87,"props":865,"children":866},{},[867,876],{"type":53,"tag":114,"props":868,"children":869},{},[870],{"type":53,"tag":266,"props":871,"children":873},{"className":872},[],[874],{"type":58,"value":875},"mip_relative_gap",{"type":53,"tag":114,"props":877,"children":878},{},[879],{"type":58,"value":880},"Stop MILP when within X% of optimal",{"type":53,"tag":87,"props":882,"children":883},{},[884,893],{"type":53,"tag":114,"props":885,"children":886},{},[887],{"type":53,"tag":266,"props":888,"children":890},{"className":889},[],[891],{"type":58,"value":892},"mip_absolute_tolerance",{"type":53,"tag":114,"props":894,"children":895},{},[896],{"type":58,"value":897},"Absolute MIP gap stop",{"type":53,"tag":87,"props":899,"children":900},{},[901,910],{"type":53,"tag":114,"props":902,"children":903},{},[904],{"type":53,"tag":266,"props":905,"children":907},{"className":906},[],[908],{"type":58,"value":909},"log_to_console",{"type":53,"tag":114,"props":911,"children":912},{},[913],{"type":58,"value":914},"Enable solver logging",{"type":53,"tag":61,"props":916,"children":917},{},[918],{"type":58,"value":919},"Syntax varies by interface — see the interface reference file.",{"items":921,"total":1025},[922,939,953,967,979,996,1011],{"slug":923,"name":923,"fn":924,"description":925,"org":926,"tags":927,"stars":26,"repoUrl":27,"updatedAt":938},"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},[928,931,934,935],{"name":929,"slug":930,"type":15},"Data Analysis","data-analysis",{"name":932,"slug":933,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":936,"slug":937,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":940,"name":940,"fn":941,"description":942,"org":943,"tags":944,"stars":26,"repoUrl":27,"updatedAt":952},"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},[945,948,951],{"name":946,"slug":947,"type":15},"Deployment","deployment",{"name":949,"slug":950,"type":15},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":954,"name":954,"fn":955,"description":956,"org":957,"tags":958,"stars":26,"repoUrl":27,"updatedAt":966},"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},[959,962,963],{"name":960,"slug":961,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":964,"slug":965,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":968,"name":968,"fn":969,"description":970,"org":971,"tags":972,"stars":26,"repoUrl":27,"updatedAt":978},"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},[973,974,975],{"name":929,"slug":930,"type":15},{"name":9,"slug":8,"type":15},{"name":976,"slug":977,"type":15},"Testing","testing","2026-07-17T05:29:03.913266",{"slug":980,"name":980,"fn":981,"description":982,"org":983,"tags":984,"stars":26,"repoUrl":27,"updatedAt":995},"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},[985,988,991,992],{"name":986,"slug":987,"type":15},"Automation","automation",{"name":989,"slug":990,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":993,"slug":994,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":997,"name":997,"fn":998,"description":999,"org":1000,"tags":1001,"stars":26,"repoUrl":27,"updatedAt":1010},"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},[1002,1003,1006,1007],{"name":946,"slug":947,"type":15},{"name":1004,"slug":1005,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":1008,"slug":1009,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":1012,"name":1012,"fn":1013,"description":1014,"org":1015,"tags":1016,"stars":26,"repoUrl":27,"updatedAt":1024},"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},[1017,1018,1021],{"name":9,"slug":8,"type":15},{"name":1019,"slug":1020,"type":15},"Quantum Computing","quantum-computing",{"name":1022,"slug":1023,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305,{"items":1027,"total":1176},[1028,1046,1060,1071,1083,1097,1110,1124,1135,1144,1158,1167],{"slug":1029,"name":1029,"fn":1030,"description":1031,"org":1032,"tags":1033,"stars":1043,"repoUrl":1044,"updatedAt":1045},"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},[1034,1037,1040],{"name":1035,"slug":1036,"type":15},"Documentation","documentation",{"name":1038,"slug":1039,"type":15},"MCP","mcp",{"name":1041,"slug":1042,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1047,"name":1047,"fn":1048,"description":1049,"org":1050,"tags":1051,"stars":1057,"repoUrl":1058,"updatedAt":1059},"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},[1052,1055,1056],{"name":1053,"slug":1054,"type":15},"Containers","containers",{"name":946,"slug":947,"type":15},{"name":21,"slug":22,"type":15},17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1061,"name":1061,"fn":1062,"description":1063,"org":1064,"tags":1065,"stars":1057,"repoUrl":1058,"updatedAt":1070},"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},[1066,1069],{"name":1067,"slug":1068,"type":15},"CI\u002FCD","ci-cd",{"name":946,"slug":947,"type":15},"2026-07-14T05:25:59.97109",{"slug":1072,"name":1072,"fn":1073,"description":1074,"org":1075,"tags":1076,"stars":1057,"repoUrl":1058,"updatedAt":1082},"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},[1077,1078,1079],{"name":1067,"slug":1068,"type":15},{"name":946,"slug":947,"type":15},{"name":1080,"slug":1081,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1084,"name":1084,"fn":1085,"description":1086,"org":1087,"tags":1088,"stars":1057,"repoUrl":1058,"updatedAt":1096},"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},[1089,1092,1093],{"name":1090,"slug":1091,"type":15},"Debugging","debugging",{"name":1080,"slug":1081,"type":15},{"name":1094,"slug":1095,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1098,"name":1098,"fn":1099,"description":1100,"org":1101,"tags":1102,"stars":1057,"repoUrl":1058,"updatedAt":1109},"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},[1103,1106],{"name":1104,"slug":1105,"type":15},"Best Practices","best-practices",{"name":1107,"slug":1108,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1111,"name":1111,"fn":1112,"description":1113,"org":1114,"tags":1115,"stars":1057,"repoUrl":1058,"updatedAt":1123},"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},[1116,1119,1122],{"name":1117,"slug":1118,"type":15},"Machine Learning","machine-learning",{"name":1120,"slug":1121,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1125,"name":1125,"fn":1126,"description":1127,"org":1128,"tags":1129,"stars":1057,"repoUrl":1058,"updatedAt":1134},"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},[1130,1133],{"name":1131,"slug":1132,"type":15},"QA","qa",{"name":976,"slug":977,"type":15},"2026-07-14T05:25:53.673039",{"slug":1136,"name":1136,"fn":1137,"description":1138,"org":1139,"tags":1140,"stars":1057,"repoUrl":1058,"updatedAt":1143},"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},[1141,1142],{"name":946,"slug":947,"type":15},{"name":949,"slug":950,"type":15},"2026-07-14T05:25:49.362534",{"slug":1145,"name":1145,"fn":1146,"description":1147,"org":1148,"tags":1149,"stars":1057,"repoUrl":1058,"updatedAt":1157},"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},[1150,1153,1154],{"name":1151,"slug":1152,"type":15},"Code Review","code-review",{"name":1080,"slug":1081,"type":15},{"name":1155,"slug":1156,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1159,"name":1159,"fn":1160,"description":1161,"org":1162,"tags":1163,"stars":1057,"repoUrl":1058,"updatedAt":1166},"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},[1164,1165],{"name":1131,"slug":1132,"type":15},{"name":976,"slug":977,"type":15},"2026-07-14T05:25:54.928983",{"slug":1168,"name":1168,"fn":1169,"description":1170,"org":1171,"tags":1172,"stars":1057,"repoUrl":1058,"updatedAt":1175},"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},[1173,1174],{"name":986,"slug":987,"type":15},{"name":1067,"slug":1068,"type":15},"2026-07-30T05:29:03.275638",496]