Re: [apache/tvm] [VOTE] Transition Main to Unity (Issue #16368)

2024-01-08 Thread Christoph Gerum
+1 -- Reply to this email directly or view it on GitHub: https://github.com/apache/tvm/issues/16368#issuecomment-1882491242 You are receiving this because you are subscribed to this thread. Message ID:

Re: [apache/tvm-rfcs] [RFC] DietCode: An Auto-Scheduler for Dynamic Tensor Programs (PR #72)

2022-05-31 Thread Christoph Gerum
@areusch It's good to merge from my side -- Reply to this email directly or view it on GitHub: https://github.com/apache/tvm-rfcs/pull/72#issuecomment-1142585916 You are receiving this because you are subscribed to this thread. Message ID:

Re: [apache/tvm-rfcs] [RFC] UMA Universal Modular Accelerator Interface (PR #60)

2022-05-15 Thread Christoph Gerum
> The main downside I see with this, is that all attributes are treated as > strings since the type is hardcoded. However, I'm not sure if we can avoid > this at all. This could probably be solved by adding `type` and/or `default` arguments to the argument parser, e.g.: ``` self._registe

Re: [apache/tvm-rfcs] [RFC] UMA Universal Modular Accelerator Interface (PR #60)

2022-03-16 Thread Christoph Gerum
> Im wondering whether should we provide an optional partitioning hook as well > -- so then it can be anything (i.e. any Sequential) and let the default be a > Sequential of MergeComposite, AnnotateTarget, MergeCompilerRegions, > ParititionGraph. WDYT ? Considering how partitioning is handled i

Re: [apache/tvm-rfcs] [RFC] UMA Universal Modular Accelerator Interface (PR #60)

2022-03-14 Thread Christoph Gerum
> One thing I'd like anwsered here what sort of control it will allow on the > passes run there : MergeComposite, AnnotateTarget, MergeCompilerRegions and > ParititionGraph. So far we had planned to standardize on MergeComposite, AnnotateTarget, MergeCompilerRegions and ParititionGraph. To get

[Apache TVM Discuss] [Development/pre-RFC] [RFC] UMA: Universal Modular Accelerator Interface

2022-03-02 Thread Christoph Gerum via Apache TVM Discuss
[quote="manupa-arm, post:11, topic:12039"] [quote="MJKlaiber, post:1, topic:12039"] ``` > class MyCustomAcceleratorPartitioner(UMAPartitioner): @property def target_name(self): return "my_custom_accelerator" def _register_patterns(self): self._register_pattern("conv1d

[Apache TVM Discuss] [Development/RFC] [Guideline] Relay AOT

2020-11-18 Thread Christoph Gerum via Apache TVM Discuss
@areusch Thanks for sharing the progress. I absolutely agree that P1 would be the better solution. Is there a migration Path from your work to P1 or would one need to implement it from scratch. --- [Visit Topic](https://discuss.tvm.apache.org/t/guideline-relay-aot/5977/6) to respond. You

[Apache TVM Discuss] [Development/RFC] [Guideline] Relay AOT

2020-11-11 Thread Christoph Gerum via Apache TVM Discuss
Not sure whether it makes sense to revive such an old RFC, but has there been any progress on a relay AOT? IMHO this would be very much needed, especially for microcontrollers. --- [Visit Topic](https://discuss.tvm.apache.org/t/guideline-relay-aot/5977/2) to respond. You are receiving th