@tqchen As you have proposed to eliminate "bfloat16" dtype after the legalization pass, I have two concerns.
* The legalization pass will be more compilicated. It has to check every TIR node to replace the bf16 dtype * As the casting between bf16 and fp32 will be lowered as function calls as is in your proposal, I am not sure if TVM can correctly vectorize the casting function. I think we'd better to treat bf16 as a "first class" type as "int16/fp32", not in the way of treating custom data types. --- [Visit Topic](https://discuss.tvm.ai/t/rfc-add-bfloat16-data-type/6778/2) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe from these emails, [click here](https://discuss.tvm.ai/email/unsubscribe/00f452d1f900eba1f6899174a2de771c1340392242fa811b52f89b0260474a0c).
