## Canonical Aliases
## Maps canonical terms to known aliases for search and normalization.
## Format: canonical_term: [alias1, alias2, ...]

# Hardware features
tcgen05:
  - UMMA
  - tensor_core_gen05
  - "tensor core generation 5"
  - tcgen05.mma

tmem:
  - tensor-memory
  - "tensor memory"
  - TMEM

clc:
  - "cluster launch control"
  - "Cluster Launch Control"
  - CLC

tma:
  - "tensor memory accelerator"
  - "Tensor Memory Accelerator"
  - TMA
  - cp.async.bulk

2sm-cooperative:
  - "2-SM cooperative"
  - "two-SM cooperative"
  - "dual CTA"
  - "2CTA"
  - "cta_group::2"

pdl:
  - "programmatic dependent launch"
  - "Programmatic Dependent Launch"
  - PDL

gdc:
  - "grid dependency control"
  - "Grid Dependency Control"
  - GDC

nvfp4:
  - NVFP4
  - "nv_float4"
  - E2M1
  - "FP4 E2M1"

wgmma:
  - "wgmma.mma_async"
  - WGMMA

block-scale:
  - "block scaling"
  - "block-scaled"
  - UE8M0
  - microscaling
  - MX

# Architecture names
sm100:
  - Blackwell
  - blackwell
  - B200
  - B100
  - GB200
  - GB300
  - "SM100"

sm90:
  - Hopper
  - hopper
  - H100
  - H200
  - H800
  - "SM90"

dcu:
  - Hygon
  - hygon
  - DCU
  - "Hygon DCU"

wave64:
  - wavefront64
  - "wavefront size 64"
  - wavefront

mmac:
  - MMOP
  - "matrix core"
  - "v_mmac"
  - HCU

ck-tile:
  - "CK Tile"
  - Composable Kernel

aiter:
  - AITER
  - "AI Tensor Engine"

aotriton:
  - AOTriton
  - "AOT Triton"
  - "ahead of time triton"

conch:
  - Conch
  - conch-triton-kernels

flaggems:
  - FlagGems
  - flag_gems

liger-kernel:
  - Liger
  - "Liger Kernel"

huggingface-kernels:
  - "Hugging Face kernels"
  - kernels-community
  - hf-kernels

triton-distributed:
  - Triton-distributed
  - triton_dist
  - "triton distributed"

# Kernel types
moe:
  - MoE
  - "mixture of experts"
  - "Mixture of Experts"
  - "expert parallelism"

mla:
  - MLA
  - "multi-head latent attention"
  - "Multi-head Latent Attention"

gated-delta-net:
  - GDN
  - GatedDeltaNet
  - "Gated Delta Net"
  - "gated delta rule"

sparse-attention:
  - NSA
  - "native sparse attention"
  - "Native Sparse Attention"

# Technique aliases
warp-specialization:
  - "warp specialization"
  - warp-spec
  - "warp specialized"

persistent-kernel:
  - "persistent kernels"
  - "persistent scheduling"

fine-grained-quantization:
  - "tile-wise scaling"
  - "block-wise scaling"
  - "per-tile quantization"

cuda-core-promotion:
  - "Nc=128 promotion"
  - "FP22 accumulator fix"
  - "accumulation promotion"