Skip to content

hrx: keep the Q6 row-range hint local to tiled prefill #191

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
AaronStGeorge wants to merge 1 commit into from
Closed

hrx: keep the Q6 row-range hint local to tiled prefill #191

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 1 addition & 2 deletions ggml/src/ggml-hrx/kernel-corpus/kernels/loom-libs/motifs/dequant.loom
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2719,7 +2719,7 @@ func.def inline @ggml_q6k_f32_vector4(%weight: buffer, %weight_row_byte_base: of
func.return %values : vector<4xf32>
}

func.def inline @ggml_q6k_row64_tile_lane_f32_vector4(%weight: buffer, %tile: index, %row_lane0: index, %q6_group: index, %packet: index) -> (vector<4xf32>) {
func.def inline @ggml_q6k_row64_tile_lane_f32_vector4(%weight: buffer, %tile: index, %row_lane: index, %q6_group: index, %packet: index) -> (vector<4xf32>) {
%c2 = index.constant 2 : index
%c4 = index.constant 4 : index
%tile_bytes = index.constant 13440 : offset
Expand All @@ -2731,7 +2731,6 @@ func.def inline @ggml_q6k_row64_tile_lane_f32_vector4(%weight: buffer, %tile: in
%c32_f32v = vector.constant 32.0 : vector<4xf32>
%bounded_group = index.assume %q6_group [range(%q6_group, 0, 7)] : index
%bounded_packet = index.assume %packet [range(%packet, 0, 7)] : index
%row_lane = index.assume %row_lane0 [range(%row_lane0, 0, 63)] : index
%tile_base = index.scale %tile, %tile_bytes : index, offset -> offset
%scale_byte_base = index.add %tile_base, %scale_byte_add : offset
%raw_byte_base = index.add %tile_base, %raw_byte_add : offset
Expand Down
4 changes: 3 additions & 1 deletion ...kernel-corpus/kernels/loom-libs/motifs/matmul/tiled_input_f32_dequant_accumulate_f32.loom
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -232,7 +232,9 @@ func.def inline @ggml_matmul_tiled_input_f32_dequant_accumulate_f32(%full_token_
scf.yield %half : vector<4xf16>
} else {
%q6_or_generic = scf.if %weight_is_q6_row64 -> (vector<4xf16>) {
%wide = func.call @ggml_q6k_row64_tile_lane_f32_vector4(%weight_noalias, %row64_tile, %local_row, %quant_group, %load_packet) : (buffer, index, index, index, index) -> (vector<4xf32>)
// Keep this hint local to prefill to preserve vector decode code generation.
%row_lane = index.assume %local_row [range(%local_row, 0, 63)] : index
%wide = func.call @ggml_q6k_row64_tile_lane_f32_vector4(%weight_noalias, %row64_tile, %row_lane, %quant_group, %load_packet) : (buffer, index, index, index, index) -> (vector<4xf32>)
%half = vector.fptrunc %wide : vector<4xf32> to vector<4xf16>
scf.yield %half : vector<4xf16>
} else {
Expand Down
Loading