Skip to main content

mlx_native/ops/
mod.rs

1//! GPU kernel host-side dispatch functions.
2//!
3//! Each submodule implements dispatch for a specific kernel family.
4
5pub mod argmax;
6pub mod argsort;
7pub mod top_k;
8pub mod copy;
9pub mod cumsum;
10pub mod dense_gemm;
11pub mod dense_gemv_bf16;
12pub mod dense_mm_bf16;
13pub mod elementwise;
14pub mod embedding;
15pub mod gather;
16pub mod gather_bench;
17pub mod hadamard;
18pub mod hadamard_quantize_kv;
19pub mod encode_helpers;
20pub mod fused_head_norm_rope;
21pub mod fused_norm_add;
22pub mod fused_residual_norm;
23pub mod gelu;
24pub mod kv_cache_copy;
25pub mod l2_norm;
26pub mod moe_dispatch;
27pub mod moe_gate;
28pub mod moe_softmax_topk;
29pub mod moe_weighted_reduce;
30pub mod quantized_matmul;
31pub mod quantized_matmul_ggml;
32pub mod quantized_matmul_id;
33pub mod quantized_matmul_id_ggml;
34pub mod rms_norm;
35pub mod rope;
36pub mod rope_multi;
37pub mod scale_mask_softmax;
38pub mod sigmoid_mul;
39pub mod silu_mul;
40pub mod compute_g_beta;
41pub mod ssm_norm_gate;
42pub mod flash_attn_prefill;
43pub mod flash_attn_prefill_blk;
44pub mod flash_attn_prefill_d512;
45pub mod flash_attn_prefill_mask;
46pub mod flash_attn_vec;
47pub mod flash_attn_vec_tq;
48pub mod flash_attn_vec_tq_hb;
49pub mod fwht_standalone;
50pub mod gated_delta_net;
51pub mod tq_dequantize_kv;
52pub mod sdpa;
53pub mod sdpa_decode;
54pub mod sdpa_sliding;
55pub mod softcap;
56pub mod softmax;
57pub mod softmax_sample;
58pub mod ssm_conv;
59pub mod transpose;
60pub mod tri_solve;