For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Mojo module

iglp

IGroupLP sched_group_barrier aggregate-pair helpers for AMD MHA.

Comptime-recursive expansions of the reference's sched_barrier_pairs<Pairs, VALU_CNT, Group> and sched_barrier_exp_pairs<...> C++ templates (see the reference attn/gqa_causal/kernel.cpp:44-56).

These helpers prescribe IGroupLP groupings to LLVM's AMDGPU instruction scheduler via the llvm.amdgcn.sched.group.barrier intrinsic. They were defined identically across several attention kernels and pulled here to a shared module: one place to fix as the language evolves, no duplication. Current consumers: mha_prefill_v2, mla_prefill, mla_prefill_v2, mla_components.

Per-kernel hint-pair parameters (which N, M for QK / PV / EXP cluster types) are tuned via parameter sweep at the kernel. Only the helper expansion logic is shared; the per-cluster (N, M) defaults belong with each kernel (they're shape-dependent and kernel-specific).

Structs

• ​AMDIGLPStrategy: Preset strategy values for the llvm.amdgcn.iglp.opt intrinsic.

Functions

• ​sched_barrier_exp_pairs: Emits pairs schedule groups of shape [1 MFMA, exp_cnt TRANS].
• ​sched_barrier_pairs: Emits pairs schedule groups of shape [1 MFMA, valu_cnt VALU].
• ​sched_dsread_valu_pairs: Emits pairs schedule groups of shape [1 DS_READ, valu_cnt VALU].

Was this page helpful?

Thank you! We'll create more content like this.

Thank you for helping us improve!

😔 What went wrong?

Some code doesn’t work

It includes inaccurate information

It's missing information I need

It was difficult to understand

Other

Submit