vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

little stuff

Browse source
This commit is contained in:
Eve 2025-01-09 21:58:26 -05:00
parent 6145fc79e5
commit 845d572b87
2 changed files with 2 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
View file

@ -19,7 +19,7 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint itid,
if (i < num_blocks_per_row) { if (i < num_blocks_per_row) {
sccache[ix][0][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8] >> v_im*4) & 0xF); // lower 8 bytes sccache[ix][0][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8] >> v_im*4) & 0xF); // lower 8 bytes
sccache[ix][1][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8+8] >> v_im*4) & 0xF); // upper 8 bytes sccache[ix][1][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8+8] >> v_im*4) & 0xF); // upper 8 bytes
} }
barrier(); barrier();
if (i >= num_blocks_per_row) if (i >= num_blocks_per_row)

1
ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
View file

@ -6,6 +6,7 @@
layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][8]; shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][8];
FLOAT_TYPE temp[NUM_COLS][NUM_ROWS]; FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
void calc_superblock(const uint a_offset, const uint b_offset, const uint ix, const uint itid8, const uint v_im, const uint v_im4, const uint v_in, const uint32_t hm_m[4], const uint q_offset, const uint y_offset, const uint s_shift, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) { void calc_superblock(const uint a_offset, const uint b_offset, const uint ix, const uint itid8, const uint v_im, const uint v_im4, const uint v_in, const uint32_t hm_m[4], const uint q_offset, const uint y_offset, const uint s_shift, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) {