[BACKEND][TLE] add sunrise backend with tle

[xdali-adora]

No description provided.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

lixingda seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[zhzhcookie]

Necessary change?

[xdali-adora]

structured bindings is not available in c++17 (until c++20)，but the CMAKE_CXX_STANDARD set to 17

[sunnycase]

Thanks for pointing this out. I may be missing something, but I do not see any structured bindings in this snippet. Also, structured bindings are a C++17 feature, so they should be allowed with CMAKE_CXX_STANDARD set to 17.

Could you point me to the exact expression you are concerned about? If there is any C++20-only construct here, I am happy to adjust it.

[xdali-adora]

for (auto &[dst, store] : storeByDst) { if (!store || duplicateStores.contains(dst)) continue; if (!isSafeToPromote(store, forOp, domInfo)) continue; OpBuilder builder(store); auto stage = builder.create<ttg::LocalAllocOp>( store.getLoc(), store.getDst().getType(), store.getSrc()); Operation *storeOp = store.getOperation(); dst.replaceUsesWithIf(stage.getResult(), [&](OpOperand &use) { Operation *owner = use.getOwner(); if (owner == storeOp) return false; if (!forOp->isAncestor(owner)) return false; return domInfo.properlyDominates(storeOp, owner); });

with store in the lambda function，there reports an error as follows
TlePromoteLocalStoreStaging.cpp:171:24: error: captured structured bindings are a C++20 extension [-Werror,-Wc++20-extensions]
171 | if (owner == store.getOperation())
| ^
TlePromoteLocalStoreStaging.cpp:159:24: note: 'store' declared here
159 | for (auto &[dst, store] : storeByDst) {
| ^

[sunnycase]

Thanks for the contribution and for adding TLE support for the sunrise backend.

Could you please update the PR description or add supporting documentation to explain which TLE primitives are implemented/supported by this work, and include performance benefit data so reviewers can evaluate whether the implementation scope matches the expected value?

It would be helpful to include:

• The list of implemented TLE primitives, their semantic coverage, and any partial support or known limitations.
• The lowering/runtime path for each key primitive, especially where it differs from the native Triton path.
• Performance data: benchmark cases, input sizes, hardware/driver environment, baseline, before/after results, improvement ratio, and any regression cases.
• If this PR is currently only structural enablement and has no measurable performance gain yet, please state that explicitly and describe the follow-up validation plan.

[zhzhcookie]

Which dir without __init__.py does sunrise backend need?

[xdali-adora]

done