All pull requests

Pull requests

• [transform.air] multi_reduction_tree: reduce a vector's leading dim in halves

  #2324·erwei-xilinx opened Nov 12, 2026Collaborator·

  ·
  5

• [air] Keep a channel put's broadcast_set when a canonicalization rebuilds it

  #2323·erwei-xilinx opened Nov 12, 2026Collaborator·

  ·
  1

• [transform.air] widen_elementwise_vectors: compute narrow 1-D elementwise vector ops at a wider lane count

  #2322·erwei-xilinx opened Nov 12, 2026Collaborator·

  ·
  5

• [transform.air] unroll_small_loops: fully unroll short loops without contractions

  #2321·erwei-xilinx opened Nov 12, 2026Collaborator·

  ·
  3

• [transform.air] accumulate_into_loop_init: start a contraction sum at what it is added to

  #2320·erwei-xilinx opened Nov 12, 2026Collaborator·

  ·

• [transform.air] hoist_loop_invariant_reads: move unchanged vector reads out of loops

  #2319·erwei-xilinx opened Nov 12, 2026Collaborator·

  ·
  5

• [transform.air.fuse_elementwise_linalg] no_recompute: don't fuse into a loop the producer is outside of

  #2318·erwei-xilinx opened Nov 12, 2026Collaborator·

  ·
  1

• [gpu] Megakernel decode on AMD GPUs: whole-model persistent launch, chiplet-aware scheduling

  #2011·erwei-xilinx opened Sep 21, 2026Collaborator·

  ·

• [fused_decode] Spell data movement as air.dma_memcpy_nd and derive the far half

  #1977·erwei-xilinx opened Sep 3, 2026Collaborator·

  ·

• [multi-gpu] Phase 7: aircc integration (--multi-gpu flag)

  #1582·erwei-xilinx opened May 3, 2026Collaborator·

  ·

• Fix air-to-std crash for DMA ops with >4 offset/size dimensions

  #1517·erwei-xilinx opened Apr 8, 2026Collaborator·3 tasks done·

  ·
  2

• Reduce Python Code Duplication

  #1476·hunhoffe opened Mar 28, 2026Contributor·

  ·

• Enable control packet reconfiguration tests on ci

  #1275·erwei-xilinx opened Feb 3, 2026Collaborator·

  ·
  1