Method element_size

triton_kernels/tensor.py:152–153 · view source on GitHub ↗

(self)

Source from the content-addressed store, hash-verified

150	return self.storage.data.numel()
151
152	def element_size(self):
153	return bitwidth(self.dtype) // 8
154
155	@property
156	def data(self):

matmul_ogsFunction · 0.80

is_tma_compliantMethod · 0.80

matmul_launch_metadataFunction · 0.80

swiglu_launch_metadataFunction · 0.80

run_single_rankFunction · 0.80

run_single_rank_ar_rms_normFunction · 0.80

run_single_rank_ar_rms_norm_fp8Function · 0.80

run_single_rank_ub_passFunction · 0.80

run_single_rank_ar_rms_norm_fp4Function · 0.80

run_single_rank_ub_mm_add_passFunction · 0.80

run_single_rank_ub_pass_fp4Function · 0.80

row_linear_residual_norm_fusion_forwardFunction · 0.80

bitwidthFunction · 0.85

run_single_rankFunction · 0.64

run_single_rank_ar_rms_normFunction · 0.64

run_single_rank_ar_rms_norm_fp8Function · 0.64

run_single_rank_ub_passFunction · 0.64

run_single_rank_ar_rms_norm_fp4Function · 0.64

run_single_rank_ub_mm_add_passFunction · 0.64

run_single_rank_ub_pass_fp4Function · 0.64

row_linear_residual_norm_fusion_forwardFunction · 0.64

nvfp4_gemm_perf_testFunction · 0.64

test_linear_kvcache_updateFunction · 0.64

_create_memory_descs_from_tensorFunction · 0.64

test_nixl_write_transfer_multiple_chunksMethod · 0.64