MasterJH5574 opened a new pull request, #14038: URL: https://github.com/apache/tvm/pull/14038
This PR introduces the static memory planning pass on binding block level, as well as an analysis function that estimate the memory usage after the memory planning pass. * The static memory planning supports nested tuples. * The static memory planning supports reuse memory of the input of reshape ops. * The estimation will return the total memory size needed to be allocated before and after memory planning, as well as the number of tensors / memory blocks to be allocated before and after memory planning. Notes for the estimate function: * The estimation is static -- it does not consider control flows (such as “if” and cross-function calls). It simply accumulates the size of every alloc_tensor and alloc_storage. * We regards “`relax.memory.alloc_tensor/storage`” as the results produced by memory planning. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
