Update HiFi1 to use nnlib (#17676)

[mcremon-meta]

Summary:

As titled. Was falling back on xtensa.
Adds a few fixes to some operators to enable the build for all cases.

Differential Revision: D94179206

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/17676

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (6 Unrelated Failures)

As of commit 034a40e with merge base 2548ee1 ():

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

• pull / test-models-linux-basic (mv3, portable, buck2, linux.2xlarge, executorch-ubuntu-22.04-clang12) / linux-job (gh) (detected as infra flaky with no log or failing log classifier)
• pull / test-models-linux-basic (mv3, xnnpack-quantization-delegation, buck2, linux.2xlarge, executorch-u... / linux-job (gh) (detected as infra flaky with no log or failing log classifier)
• pull / test-static-llama-qnn-linux (stories_110m) / linux-job (gh) (detected as infra flaky with no log or failing log classifier)
• pull / test-vulkan-models-linux / linux-job (gh) (detected as infra flaky with no log or failing log classifier)
• pull / unittest / windows / windows-job (gh) (matched win rule in flaky-rules.json)
  ##[error]The operation was canceled.

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• pull / unittest-editable / windows / windows-job (gh) (trunk failure)
  ##[error]The operation was canceled.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[meta-codesync]

@mcremon-meta has exported this pull request. If you are a Meta employee, you can view the originating Diff in D94179206.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[mcremon-meta]

@dijopaul @cad-audio can you take a look please? :)

[dijopaul]

I think this is not needed. Can remove this line
xtbool2 con = int32_rtor_xtbool2(0x00000003);
the vaiable not used

[dijopaul]

Can use the below code

#ifndef AE_MOVAB
#ifdef AE_MOVAB1
#define AE_MOVAB AE_MOVAB1
#else
#define AE_MOVAB(x) (x)
#endif
#endif

#ifndef AE_MOVAB2
#define AE_MOVAB2(x) (x)
#endif

[dijopaul]

@dijopaul @cad-audio can you take a look please? :)

The root cause here is that we are using local implementations outside nnlib (which is hifi4 specific). The latest nnlib has all these implementation inside including hifi1 support. So better solution is to upgrade to latest nnlib and avoid these local files. We will make a PR for the same. This may involve some changes in API calls as well due to minor changes in API names.

[mcremon-meta]

@dijopaul @cad-audio can you take a look please? :)

The root cause here is that we are using local implementations outside nnlib (which is hifi4 specific). The latest nnlib has all these implementation inside including hifi1 support. So better solution is to upgrade to latest nnlib and avoid these local files. We will make a PR for the same. This may involve some changes in API calls as well due to minor changes in API names.

I see, makes sense. Are some of those changes on our side, or if you open a PR to update the lib, we get things for free on our end? If that's the case, we can wait a little bit

[cad-audio]

@dijopaul @cad-audio can you take a look please? :)

The root cause here is that we are using local implementations outside nnlib (which is hifi4 specific). The latest nnlib has all these implementation inside including hifi1 support. So better solution is to upgrade to latest nnlib and avoid these local files. We will make a PR for the same. This may involve some changes in API calls as well due to minor changes in API names.

I see, makes sense. Are some of those changes on our side, or if you open a PR to update the lib, we get things for free on our end? If that's the case, we can wait a little bit

We will create a PR for nnlib update after verifying the build on HiFi1. We will target this on this week

[mcremon-meta]

@dijopaul @cad-audio can you take a look please? :)

The root cause here is that we are using local implementations outside nnlib (which is hifi4 specific). The latest nnlib has all these implementation inside including hifi1 support. So better solution is to upgrade to latest nnlib and avoid these local files. We will make a PR for the same. This may involve some changes in API calls as well due to minor changes in API names.

I see, makes sense. Are some of those changes on our side, or if you open a PR to update the lib, we get things for free on our end? If that's the case, we can wait a little bit

We will create a PR for nnlib update after verifying the build on HiFi1. We will target this on this week

Sounds good! Thanks :)