Files
pytorch/torch
Chirag Pandya fd90991790 [rfc] opentelemetry in pytorch (#122999)
1. Add current latest version (opentelemetry-cpp version v1.14.2) to PyTorch library.
Steps:
```
$cd pytorch
$git submodule add https://github.com/open-telemetry/opentelemetry-cpp.git third_party/opentelemetry-cpp
$cd third_party/opentelemetry-cpp
$git checkout v1.14.2
$git add third_party/opentelemetry-cpp .gitmodules
$git commit
```
Expected change in checkout size:
```
(/home/cpio/local/a/pytorch-env) [cpio@devvm17556.vll0 ~/local/pytorch (gh/c-p-i-o/otel)]$ git count-objects -vH
count: 654
size: 3.59 MiB
in-pack: 1229701
packs: 17
size-pack: 1.17 GiB
prune-packable: 76
garbage: 0
size-garbage: 0 bytes
```

2.

TODO
- [x] Figure out how dynamic linking works. App builders will somehow need to `target_include` opentelemetry-cpp at runtime.
- [ ] Examples on how to use opentelemetry + pytorch
- [ ] Tests + documentation (e.g. using null opentelemetry implementation).

Pull Request resolved: https://github.com/pytorch/pytorch/pull/122999
Approved by: https://github.com/ezyang
2024-04-21 15:20:21 +00:00
..
2024-01-29 23:42:35 +00:00
2024-04-21 14:06:23 +00:00
2024-04-21 14:06:23 +00:00
2024-04-21 14:06:23 +00:00
2024-04-21 14:06:23 +00:00
2024-03-30 13:07:53 +00:00
2024-02-29 08:29:14 +00:00

Note [TH abstraction violation]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

TH/THC provide some hpp headers, which are proper C++ headers rather than
C headers.  These headers serve double duty as *internal implementation
detail* headers, whose contents should largely not be used by external
clients.

Ideally, we would not install these headers at all; instead, you should
use public functions (in headers like `THTensor.h`, NOT `THTensor.hpp`)
to manipulate these structs.  However, there are a few places
in torch/csrc where we violate this abstraction.  They are marked with
a pointer to this note.  Each of those sites will have to be refactored
when we refactor the guts of THTensor and related structures.