MCPcopy Create free account

hub / github.com/HazyResearch/HipKittens / functions

Functions1,340 in github.com/HazyResearch/HipKittens

↓ 1 callersMethodblock
kernels/attn/gqa_backwards/attn_fwd_non_causal.cpp:76
↓ 1 callersMethodblock
kernels/attn/gqa_backwards/archive/GQA_bkwd_asm.cpp:29
↓ 1 callersMethodblock
kernels/attn/gqa_backwards/archive/GQA_bkwd_4warps.cpp:39
↓ 1 callersMethodblock
kernels/attn/gqa_backwards/archive/GQA_bkwd_8warps.cpp:40
↓ 1 callersMethodblock
kernels/attn/gqa_backwards/archive/GQA_fwd.cpp:50
↓ 1 callersMethodblock
kernels/attn/gqa_causal_backwards/attn_bkwd_causal.cpp:44
↓ 1 callersMethodblock
kernels/attn/gqa_causal_backwards/attn_fwd_causal.cpp:156
↓ 1 callersMethodblock
kernels/attn/gqa/kernel.cpp:98
↓ 1 callersMethodblock
kernels/attn/gqa/kernel_d64.cpp:98
↓ 1 callersMethodblock
kernels/attn/gqa_causal/kernel.cpp:177
↓ 1 callersMethodblock
kernels/attn/gqa_causal/kernel_d64.cpp:177
↓ 1 callersMethodblock
kernels/rotary/kernel.cpp:40
↓ 1 callersMethodblock
training/llama/csrc/attn_bkwd_causal_HBN.cpp:44
↓ 1 callersMethodblock
training/llama/csrc/attn_fwd_causal.cpp:156
↓ 1 callersMethodblock
training/llama/csrc/attn_bkwd_causal_HNB.cpp:44
↓ 1 callersFunctionbuffer_load_vec4
analysis/bf16_gemm/mi325x/utils.cpp:32
↓ 1 callersFunctionbuffer_load_vec4
kernels/gemm/bf16fp32/archive/utils.cpp:36
↓ 1 callersFunctionceil_div
analysis/paper_experiments/grid_micro/kernel_8192_w32.cpp:24
↓ 1 callersFunctionceil_div
analysis/paper_experiments/grid_micro/kernel_8192_w2.cpp:24
↓ 1 callersFunctionceil_div
analysis/paper_experiments/grid_micro/kernel_8192_w8.cpp:24
↓ 1 callersFunctionceil_div
analysis/paper_experiments/grid_micro/kernel_8192_w4.cpp:24
↓ 1 callersMethodcheck_args
(self, q, k, v, o)
analysis/baselines/attn/triton_baseline_v02.py:99
↓ 1 callersFunctioncoalesce_columns
(df)
analysis/bf16_gemm/mi350x/analyze_prof.py:60
↓ 1 callersFunctioncoalesce_columns
(df)
docs/profiling/analyze_pmc_counter_output.py:60
↓ 1 callersMethodcompute
(self)
training/llama/train/metrics/num_tokens.py:30
↓ 1 callersFunctioncpu_gemm_abt_ref
kernels/gemm/bf16fp32/gfx1250/harness.h:40
↓ 1 callersFunctioncreate_block
(config, layer_idx=None, device=None, dtype=None, **kwargs)
training/llama/llama/models/gpt.py:85
↓ 1 callersFunctioncreate_mha_cls
(config, layer_idx=None, device=None, dtype=None)
training/llama/llama/models/gpt.py:25
↓ 1 callersFunctioncreate_mlp_cls
(config, layer_idx=None, device=None, dtype=None)
training/llama/llama/models/gpt.py:70
↓ 1 callersFunctioncross_entropy_loss
Arguments: logits: (batch, vocab_size) labels: (batch,) label_smoothing: float logit_scale: float. Multiply logit
training/llama/llama/ops/triton/cross_entropy.py:284
↓ 1 callersMethoddata_ptr
distributed-kernels/iris_py.cpp:28
↓ 1 callersFunctiondictconfig_filter_key
Only keep keys where fn(key) is True. Support nested DictConfig.
training/llama/train/run.py:23
↓ 1 callersFunctiondropout_mask
(philox_seed, philox_offset, dropout_p, m, n, stride)
analysis/baselines/attn/triton_baseline_v02.py:164
↓ 1 callersFunctiondropout_mask
analysis/layernorm/mi350x/kernel_4096.cpp:21
↓ 1 callersFunctiondropout_mask
analysis/layernorm/mi350x/kernel_1024.cpp:21
↓ 1 callersFunctiondropout_mask
analysis/layernorm/mi350x/kernel_2048.cpp:21
↓ 1 callersFunctiondropout_mask
analysis/layernorm/mi350x/kernel_16384.cpp:21
↓ 1 callersFunctiondropout_mask
analysis/layernorm/mi350x/kernel_8192.cpp:21
↓ 1 callersFunctiondropout_mask
kernels/layernorm/kernel.cpp:21
↓ 1 callersFunctiondropout_offsets
(philox_seed, philox_offset, dropout_p, m, n, stride)
analysis/baselines/attn/triton_baseline_v02.py:150
↓ 1 callersFunctiondropout_rng
(philox_seed, philox_offset, dropout_p, m, n, stride)
analysis/baselines/attn/triton_baseline_v02.py:157
↓ 1 callersMethoddynamic_shared_memory
distributed-kernels/bf16_gemm/kernel.cpp:86
↓ 1 callersMethoddynamic_shared_memory
analysis/paper_experiments/phases/ds_write_b64/kernel.cpp:15
↓ 1 callersMethoddynamic_shared_memory
analysis/paper_experiments/phases/ds_read_b128/kernel.cpp:15
↓ 1 callersMethoddynamic_shared_memory
analysis/paper_experiments/phases/ds_read_b64/kernel.cpp:15
↓ 1 callersMethoddynamic_shared_memory
analysis/paper_experiments/phases/ds_read_b96/kernel.cpp:15
↓ 1 callersMethoddynamic_shared_memory
analysis/paper_experiments/grid_micro/kernel_8192_w0.cpp:34
↓ 1 callersMethoddynamic_shared_memory
analysis/paper_experiments/grid_micro/kernel_8192_w32.cpp:34
↓ 1 callersMethoddynamic_shared_memory
analysis/paper_experiments/grid_micro/kernel_8192_w2.cpp:34
↓ 1 callersMethoddynamic_shared_memory
analysis/paper_experiments/grid_micro/kernel_8192_w8.cpp:34
↓ 1 callersMethoddynamic_shared_memory
analysis/paper_experiments/grid_micro/kernel_8192_w4.cpp:34
↓ 1 callersMethoddynamic_shared_memory
analysis/layernorm/mi350x/kernel_4096.cpp:81
↓ 1 callersMethoddynamic_shared_memory
analysis/layernorm/mi350x/kernel_1024.cpp:82
↓ 1 callersMethoddynamic_shared_memory
analysis/layernorm/mi350x/kernel_2048.cpp:81
↓ 1 callersMethoddynamic_shared_memory
analysis/layernorm/mi350x/kernel_16384.cpp:81
↓ 1 callersMethoddynamic_shared_memory
analysis/layernorm/mi350x/kernel_8192.cpp:81
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/bkwd/benchmark/attn_bkwd_causal_HBN.cpp:45
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/bkwd/benchmark/attn_bkwd_non_causal_8_warps.cpp:54
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/bkwd/benchmark/attn_fwd_causal.cpp:157
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/bkwd/benchmark/attn_bkwd_causal_HNB.cpp:45
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/bkwd/benchmark/attn_fwd_non_causal.cpp:77
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/bkwd/benchmark/attn_bkwd_non_causal_4_warps.cpp:53
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/bkwd/benchmark/attn_bkwd_non_causal_HBN.cpp:44
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/bkwd/benchmark/attn_bkwd_non_causal_HNB.cpp:44
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/fwd/benchmark/attn_fwd_causal.cpp:178
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/fwd/benchmark/attn_fwd_causal_d64.cpp:178
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/fwd/benchmark/attn_fwd_non_causal.cpp:98
↓ 1 callersMethoddynamic_shared_memory
analysis/attn/fwd/benchmark/attn_fwd_non_causal_d64.cpp:99
↓ 1 callersMethoddynamic_shared_memory
analysis/rotary/mi350x/kernel_4096.cpp:41
↓ 1 callersMethoddynamic_shared_memory
analysis/rotary/mi350x/kernel_1024.cpp:41
↓ 1 callersMethoddynamic_shared_memory
analysis/rotary/mi350x/kernel_2048.cpp:41
↓ 1 callersMethoddynamic_shared_memory
analysis/rotary/mi350x/kernel_16384.cpp:41
↓ 1 callersMethoddynamic_shared_memory
analysis/rotary/mi350x/kernel_8192.cpp:41
↓ 1 callersMethoddynamic_shared_memory
analysis/bf16_gemm/mi325x/kernel_4096.cpp:31
↓ 1 callersMethoddynamic_shared_memory
analysis/bf16_gemm/mi325x/kernel_1024.cpp:31
↓ 1 callersMethoddynamic_shared_memory
analysis/bf16_gemm/mi325x/kernel_2048.cpp:31
↓ 1 callersMethoddynamic_shared_memory
analysis/bf16_gemm/mi325x/kernel_16384.cpp:31
↓ 1 callersMethoddynamic_shared_memory
analysis/bf16_gemm/mi325x/kernel_8192.cpp:31
↓ 1 callersMethoddynamic_shared_memory
analysis/bf16_gemm/mi350x/kernel.cpp:35
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/256_256_64_32_with32x16.cpp:31
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/256_256_64_32_with16x32.cpp:35
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/micro_04_2stage_12c4p.cpp:32
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/micro_09_async.cpp:33
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/micro_03_3stage_8c4p.cpp:33
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/micro_02_2stage_8c4p.cpp:33
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/timing/micro_01_syncthreads.cpp:25
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/archive/micro_06_2stage_8c4p_64x96.cpp:34
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/archive/micro_06_2stage_8c4p_96x64.cpp:33
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/archive/micro_07_2stage_8c4p_nblock8.cpp:31
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/archive/micro_08_2stage_4c4p.cpp:35
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/archive/micro_08_async.cpp:33
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/archive/micro_08_4stage_4c4p.cpp:34
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/32x16/archive/micro_05_2stage_16c2p.cpp:32
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/16x32/micro_04_2stage_12c4p.cpp:34
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/16x32/micro_03_3stage_8c4p.cpp:34
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/16x32/micro_02_2stage_8c4p.cpp:34
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/16x32/micro_05_async.cpp:33
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/16x32/archive/micro_06_async.cpp:33
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/16x32/archive/micro_03_3stage_8c4p_async.cpp:34
↓ 1 callersMethoddynamic_shared_memory
kernels/gemm/bf16fp32/micros/producer_consumer/16x32/archive/micro_07_async.cpp:33
← previousnext →301–400 of 1,340, ranked by callers