whisper.cpp/ggml/src/ggml-vulkan/vulkan-shaders
Jeff Bolz a375e4c4d2
vulkan: Fix validation failure in quantized flash attention (llama/16292)
2025-09-29 15:18:12 +03:00
..
CMakeLists.txt vulkan: Fix GGML_VULKAN_SHADER_DEBUG_INFO (llama/14427) 2025-07-01 17:54:53 +03:00
acc.comp
add.comp vulkan: optimize rms_norm, and allow the work to spread across multiple SMs (llama/15281) 2025-09-20 13:42:40 +03:00
add_id.comp llama : add gpt-oss (llama/15091) 2025-08-18 20:30:45 +03:00
argmax.comp vulkan : fix out-of-bounds access in argmax kernel (llama/15342) 2025-08-18 20:30:45 +03:00
argsort.comp vulkan: Optimize argsort (llama/15354) 2025-08-18 20:30:45 +03:00
clamp.comp
concat.comp
contig_copy.comp vulkan: Add bfloat16 support (llama/12554) 2025-05-07 15:39:32 +03:00
conv2d_dw.comp vulkan : fix lint (llama/0) 2025-05-07 15:39:32 +03:00
conv2d_mm.comp Vulkan: add conv_transpose_2d operation (llama/16022) 2025-09-29 15:18:09 +03:00
conv_transpose_1d.comp ggml-vulkan: adds support for op CONV_TRANSPOSE_1D (llama/13813) 2025-06-10 12:40:33 +03:00
copy.comp vulkan: Add bfloat16 support (llama/12554) 2025-05-07 15:39:32 +03:00
copy_from_quant.comp llama : add gpt-oss (llama/15091) 2025-08-18 20:30:45 +03:00
copy_to_quant.comp ggml : implement set_rows with i32 index (llama/16159) 2025-09-29 15:18:09 +03:00
cos.comp
count_equal.comp
dequant_f32.comp
dequant_funcs.comp vulkan: support GET_ROWS for k-quants (llama/16235) 2025-09-29 15:18:11 +03:00
dequant_funcs_cm2.comp llama : add gpt-oss (llama/15091) 2025-08-18 20:30:45 +03:00
dequant_head.comp
dequant_iq1_m.comp vulkan: fix warnings (llama/13626) 2025-05-27 18:03:00 +03:00
dequant_iq1_s.comp
dequant_iq2_s.comp vulkan: fix failing dequant shaders (llama/15862) 2025-09-20 13:45:29 +03:00
dequant_iq2_xs.comp
dequant_iq2_xxs.comp vulkan: fix failing dequant shaders (llama/15862) 2025-09-20 13:45:29 +03:00
dequant_iq3_s.comp vulkan: fix failing dequant shaders (llama/15862) 2025-09-20 13:45:29 +03:00
dequant_iq3_xxs.comp vulkan: fix failing dequant shaders (llama/15862) 2025-09-20 13:45:29 +03:00
dequant_iq4_nl.comp
dequant_iq4_xs.comp
dequant_mxfp4.comp llama : add gpt-oss (llama/15091) 2025-08-18 20:30:45 +03:00
dequant_q2_k.comp vulkan: fix noncontig check for mat_mul_id splitting (llama/14683) 2025-07-20 00:23:50 +03:00
dequant_q3_k.comp vulkan: fix noncontig check for mat_mul_id splitting (llama/14683) 2025-07-20 00:23:50 +03:00
dequant_q4_0.comp
dequant_q4_1.comp
dequant_q4_k.comp vulkan: fix noncontig check for mat_mul_id splitting (llama/14683) 2025-07-20 00:23:50 +03:00
dequant_q5_0.comp
dequant_q5_1.comp
dequant_q5_k.comp vulkan: fix noncontig check for mat_mul_id splitting (llama/14683) 2025-07-20 00:23:50 +03:00
dequant_q6_k.comp vulkan: fix noncontig check for mat_mul_id splitting (llama/14683) 2025-07-20 00:23:50 +03:00
dequant_q8_0.comp
diag_mask_inf.comp
div.comp
exp.comp vulkan: add RTE variants of exp shader (llama/16165) 2025-09-29 15:18:08 +03:00
flash_attn.comp vulkan: support arbitrary KV dimension in flash attention (llama/16160) 2025-09-29 15:18:12 +03:00
flash_attn_base.comp vulkan: Fix validation failure in quantized flash attention (llama/16292) 2025-09-29 15:18:12 +03:00
flash_attn_cm1.comp vulkan: support arbitrary KV dimension in flash attention (llama/16160) 2025-09-29 15:18:12 +03:00
flash_attn_cm2.comp vulkan: clamp matmul and FA results to the max finite value (llama/15652) 2025-09-20 13:42:45 +03:00
flash_attn_split_k_reduce.comp vulkan: clamp matmul and FA results to the max finite value (llama/15652) 2025-09-20 13:42:45 +03:00
geglu.comp ggml : implement REGLU/GEGLU/SWIGLU ops (llama/14158) 2025-07-01 17:54:53 +03:00
geglu_erf.comp ggml : implement GEGLU_ERF and GEGLU_QUICK ops (llama/14445) 2025-07-12 19:23:56 +03:00
geglu_quick.comp ggml : implement GEGLU_ERF and GEGLU_QUICK ops (llama/14445) 2025-07-12 19:23:56 +03:00
gelu.comp
gelu_erf.comp add GELU_ERF (llama/14455) 2025-07-12 19:23:56 +03:00
gelu_quick.comp
generic_binary_head.comp vulkan: fuse adds (llama/15252) 2025-08-18 20:30:45 +03:00
generic_head.comp
generic_unary_head.comp
get_rows.comp vulkan: handle large sizes for get_rows (llama/15686) 2025-09-20 13:42:45 +03:00
get_rows_quant.comp vulkan: handle large sizes for get_rows (llama/15686) 2025-09-20 13:42:45 +03:00
glu_head.comp llama : add gpt-oss (llama/15091) 2025-08-18 20:30:45 +03:00
glu_main.comp ggml : implement REGLU/GEGLU/SWIGLU ops (llama/14158) 2025-07-01 17:54:53 +03:00
group_norm.comp
hardsigmoid.comp ggml vulkan: add hardsigmoid and hardswish operations (llama/15762) 2025-09-20 13:42:48 +03:00
hardswish.comp ggml vulkan: add hardsigmoid and hardswish operations (llama/15762) 2025-09-20 13:42:48 +03:00
im2col.comp vulkan: 64-bit im2col (llama/16135) 2025-09-29 15:18:12 +03:00
im2col_3d.comp vulkan: 64-bit im2col (llama/16135) 2025-09-29 15:18:12 +03:00
l2_norm.comp
leaky_relu.comp
mul.comp
mul_mat_split_k_reduce.comp
mul_mat_vec.comp vulkan: Add bfloat16 support (llama/12554) 2025-05-07 15:39:32 +03:00
mul_mat_vec_base.comp Vulkan: Add Integer Dot Product mul_mat_vec shader for legacy quants (llama/14903) 2025-09-20 13:42:46 +03:00
mul_mat_vec_iq1_m.comp
mul_mat_vec_iq1_s.comp
mul_mat_vec_iq2_s.comp
mul_mat_vec_iq2_xs.comp
mul_mat_vec_iq2_xxs.comp
mul_mat_vec_iq3_s.comp
mul_mat_vec_iq3_xxs.comp
mul_mat_vec_nc.comp vulkan: Support ne[3]>1 in noncontig matrix-vector multiply (llama/15015) 2025-08-18 20:30:45 +03:00
mul_mat_vec_p021.comp
mul_mat_vec_q2_k.comp
mul_mat_vec_q3_k.comp
mul_mat_vec_q4_k.comp
mul_mat_vec_q5_k.comp
mul_mat_vec_q6_k.comp
mul_mat_vecq.comp Vulkan: Add Integer Dot Product mul_mat_vec shader for legacy quants (llama/14903) 2025-09-20 13:42:46 +03:00
mul_mm.comp vulkan: vec dot matrix multiplication fix (llama/16151) 2025-09-29 15:18:08 +03:00
mul_mm_cm2.comp vulkan: handle mat_mul with A matrix > 4GB (llama/16176) 2025-09-29 15:18:12 +03:00
mul_mm_funcs.comp vulkan: vec dot matrix multiplication fix (llama/16151) 2025-09-29 15:18:08 +03:00
mul_mmq.comp Vulkan: Add Integer Dot Product mul_mat_vec shader for legacy quants (llama/14903) 2025-09-20 13:42:46 +03:00
mul_mmq_funcs.comp Vulkan: Add Integer Dot Product mul_mat_vec shader for legacy quants (llama/14903) 2025-09-20 13:42:46 +03:00
multi_add.comp vulkan: workaround MoltenVK compile failure in multi_add (llama/15506) 2025-09-20 13:42:40 +03:00
norm.comp
opt_step_adamw.comp
opt_step_sgd.comp finetune: SGD optimizer, more CLI args (llama/13873) 2025-08-18 20:30:45 +03:00
pad.comp vulkan: Support pad_ext (llama/15794) 2025-09-20 13:42:51 +03:00
pool2d.comp
quantize_q8_1.comp Vulkan: Add Integer Dot Product mul_mat_vec shader for legacy quants (llama/14903) 2025-09-20 13:42:46 +03:00
reglu.comp ggml : implement REGLU/GEGLU/SWIGLU ops (llama/14158) 2025-07-01 17:54:53 +03:00
relu.comp vulkan: Additional type support for unary, binary, and copy (llama/13266) 2025-05-07 21:00:32 +03:00
repeat.comp
repeat_back.comp
rms_norm.comp vulkan: optimize rms_norm, and allow the work to spread across multiple SMs (llama/15281) 2025-09-20 13:42:40 +03:00
rms_norm_back.comp
rms_norm_partials.comp vulkan: optimize rms_norm, and allow the work to spread across multiple SMs (llama/15281) 2025-09-20 13:42:40 +03:00
roll.comp vulkan : implement ggml_roll (ggml/1290) 2025-07-12 19:23:56 +03:00
rope_head.comp vulkan: add RTE variants for glu/add/sub/mul/div (llama/14653) 2025-07-20 00:23:50 +03:00
rope_multi.comp vulkan : fix rope with partial rotation and non-cont src (llama/14582) 2025-07-12 19:23:56 +03:00
rope_neox.comp vulkan : fix rope with partial rotation and non-cont src (llama/14582) 2025-07-12 19:23:56 +03:00
rope_norm.comp vulkan : fix rope with partial rotation and non-cont src (llama/14582) 2025-07-12 19:23:56 +03:00
rope_vision.comp
rte.comp vulkan: add RTE variants for glu/add/sub/mul/div (llama/14653) 2025-07-20 00:23:50 +03:00
scale.comp ggml : add ggml_scale_bias (llama/14417) 2025-07-12 19:23:56 +03:00
sigmoid.comp vulkan: Additional type support for unary, binary, and copy (llama/13266) 2025-05-07 21:00:32 +03:00
silu.comp
silu_back.comp
sin.comp
soft_max.comp llama : add gpt-oss (llama/15091) 2025-08-18 20:30:45 +03:00
soft_max_back.comp vulkan: Fix OOB accesses in soft_max_back (llama/15861) 2025-09-20 13:42:52 +03:00
sqrt.comp vulkan: support sqrt (llama/15370) 2025-08-18 20:30:45 +03:00
square.comp
sub.comp
sum_rows.comp vulkan : support ggml_mean (llama/15393) 2025-09-20 13:42:40 +03:00
swiglu.comp ggml : implement REGLU/GEGLU/SWIGLU ops (llama/14158) 2025-07-01 17:54:53 +03:00
swiglu_oai.comp llama : add gpt-oss (llama/15091) 2025-08-18 20:30:45 +03:00
tanh.comp vulkan: Additional type support for unary, binary, and copy (llama/13266) 2025-05-07 21:00:32 +03:00
test_bfloat16_support.comp vulkan: Add bfloat16 support (llama/12554) 2025-05-07 15:39:32 +03:00
test_coopmat2_support.comp
test_coopmat_support.comp
test_integer_dot_support.comp Vulkan: Add DP4A MMQ and Q8_1 quantization shader (llama/12135) 2025-04-02 15:51:57 +03:00
timestep_embedding.comp ggml : fix padding in timestep embedding kernels (llama/15932) 2025-09-20 13:45:30 +03:00
types.comp vulkan: 64-bit im2col (llama/16135) 2025-09-29 15:18:12 +03:00
upscale.comp sync : resolve conflicts (ggml/0) 2025-07-12 19:23:56 +03:00
utils.comp vulkan: fuse adds (llama/15252) 2025-08-18 20:30:45 +03:00
vulkan-shaders-gen.cpp vulkan: 64-bit im2col (llama/16135) 2025-09-29 15:18:12 +03:00
wkv6.comp
wkv7.comp