Search
By facebookexperimental
Skills
Sort:BestMost starsTrending todayTrending this weekTrending this monthNewestRecently updatedName
| # | Skill | Repository | Stars | Used in | Tokens | Auto-check | Licence | Updated |
|---|---|---|---|---|---|---|---|---|
| 1 | Collect, validate, package, and inspect rocprofv3 Advanced Thread Trace bundles for AMD GPU kernels. | facebookexperimental/ | 201 | — | ~733 | Automated safety check: Pass | MIT | today |
| 2 | Design and run Triton TTGIR debugging ablations using iroverride. | facebookexperimental/ | 201 | — | ~978 | Automated safety check: Pass | MIT | today |
| 3 | Execute the TLX Kernel Optimization Agent CLI on a Triton or TLX kernel. | facebookexperimental/ | 201 | — | ~3.5k | Automated safety check: Pass | MIT | today |
| 4 | Run NVIDIA compute-sanitizer (memcheck, racecheck, initcheck, synccheck) against a Triton/TLX kernel to find runtime memory and synchronization bugs. | facebookexperimental/ | 201 | — | ~1.6k | Automated safety check: Pass | MIT | today |
| 5 | Recover from GPU-busy / GPU-unavailable failures. An agent skill from facebookexperimental/triton. | facebookexperimental/ | 201 | — | ~709 | Automated safety check: Pass | MIT | today |
| 6 | Debug Triton compilation by dumping IR at each stage (TTIR, TTGIR, LLVM, PTX). | facebookexperimental/ | 201 | — | ~644 | Automated safety check: Pass | MIT | today |
| 7 | Run TLX kernel performance benchmarks on Hopper, Blackwell, and AMD (gfx950/CDNA4, gfx1250) GPUs. | facebookexperimental/ | 201 | — | ~1.1k | Automated safety check: Pass | MIT | today |
| 8 | A skill your agent uses when working on fence-related compiler passes, TMA store lowering, proxy fence insertion, investigating missing or spurious fences, or debugging correctness issues in TLX… | facebookexperimental/ | 201 | — | ~1.2k | Automated safety check: Pass | MIT | today |
| 9 | How to build and run GPU targets under Buck in fbcode. An agent skill from facebookexperimental/triton. | facebookexperimental/ | 201 | — | ~998 | Automated safety check: Pass | MIT | today |
| 10 | Run the sched2tlx perf/correctness harness over the modulo-scheduling example corpus (case1-9: GEMM, persistent GEMM, FA fwd/bwd, addmm+bias, LayerNorm, wgrad+bias, multiphase GEMM, scaledmm). | facebookexperimental/ | 201 | — | ~1.9k | Automated safety check: Pass | MIT | yesterday |
| 11 | Test and run TLX-AMD tutorial kernels (gfx950/CDNA4 and gfx1250) and understand their CI. | facebookexperimental/ | 201 | — | ~1.3k | Automated safety check: Notes | MIT | yesterday |
| 12 | TLX DSL API reference for low-level GPU primitives. An agent skill from facebookexperimental/triton. | facebookexperimental/ | 201 | — | ~3.9k | Automated safety check: Pass | MIT | yesterday |
| 13 | Author Triton kernels with automatic warp specialization (AutoWS). | facebookexperimental/ | 201 | — | ~7.1k | Automated safety check: Pass | MIT | yesterday |
| 14 | Run autoWS (automatic warp specialization) correctness tests. | facebookexperimental/ | 201 | — | ~467 | Automated safety check: Pass | MIT | yesterday |
| 15 | Produce a structured barrier report for AutoWS (automatic warp specialization) IR. | facebookexperimental/ | 201 | — | ~6.8k | Automated safety check: Pass | MIT | yesterday |
| 16 | Consult and maintain AutoWS documentation. An agent skill from facebookexperimental/triton. | facebookexperimental/ | 201 | — | ~968 | Automated safety check: Pass | MIT | yesterday |
| 17 | Diagnose CUDA "illegal instruction" / kernel crashes on Triton kernels that reference to TMA loads or stores (maketensordescriptor, TensorDescriptor, descriptor.load, descriptor.store… | facebookexperimental/ | 201 | — | ~1.1k | Automated safety check: Pass | MIT | yesterday |
| 18 | Compare Triton, TLX, and inductor kernel numerics across A/B compiler builds with per-config isolation tests and grouped impact reporting. | facebookexperimental/ | 201 | — | ~792 | Automated safety check: Pass | MIT | yesterday |
