[EPLB] Nixl communicator optimization. Zero-copy transfers (#41633)

Signed-off-by: ilmarkov <markovilya197@gmail.com>
Signed-off-by: Markov Ilya <markovilya19@gmail.com>
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Markov Ilya <markovilya19@gmail.com>
Co-authored-by: Nicolò Lucchesi <nlucches@redhat.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
This commit is contained in:
Ilya Markov
2026-06-04 03:40:34 +00:00
committed by GitHub
co-authored by Markov Ilya Nicolò Lucchesi Tyler Michael Smith Tyler Michael Smith
parent f0cd590d62
commit 4f423bd5bc
12 changed files with 287 additions and 219 deletions
+14
View File
@@ -64,6 +64,20 @@ def divide(numerator, denominator):
return numerator // denominator
def is_weak_contiguous(inp: torch.Tensor) -> bool:
"""Check that *inp* occupies a single contiguous block of memory.
Unlike ``torch.Tensor.is_contiguous()``, this also accepts tensors
whose strides are not strictly C-contiguous (e.g. column-major) as
long as the underlying storage from the tensor's offset onward is
exactly ``numel * element_size`` bytes.
"""
return inp.is_contiguous() or (
inp.storage().nbytes() - inp.storage_offset() * inp.element_size()
== inp.numel() * inp.element_size()
)
def split_tensor_along_last_dim(
tensor: torch.Tensor,
num_partitions: int,