Code
Hub
Workspaces
Connect
Indexed graphs
Engine
MCP
copy
hub
/
github.com/NVIDIA/TensorRT-LLM
/ stride
Method
stride
triton_kernels/tensor.py:143–144 ·
view source on GitHub ↗
(self, i=None)
Source
from the content-addressed store, hash-verified
141
return
self.storage.device
142
143
def
stride(self, i=None):
144
return
self.storage.data.stride()
if
i is None
else
self.storage.data.stride(i)
145
146
def
data_ptr(self):
147
return
self.storage.data.data_ptr()
Callers
15
matmul_ogs
Function · 0.95
topk_forward
Function · 0.95
can_overflow_int32
Function · 0.80
_canonicalize_storage
Function · 0.80
reduce_grouped
Function · 0.80
forward
Method · 0.80
topk_backward
Function · 0.80
forward
Method · 0.80
forward
Method · 0.80
compute_expt_data
Function · 0.80
compaction
Function · 0.80
is_tma_compliant
Method · 0.80
Calls
no outgoing calls
Tested by
7
test_quantization_dequantization_activation
Method · 0.64
test_quantization_dequantization_weight
Method · 0.64
test_indexer_k_cache_scatter_custom_op
Function · 0.64
test_update_paged_kv_cache
Function · 0.64
test_attention_kv_paged_flash_decoding
Function · 0.64
run
Function · 0.64
test_context_attention_kv_paged
Function · 0.64