Is Databend's execution model better than the volcano model used by Datafusion? #7954
-
|
I understand that the volcano model has some obvious drawbacks, such as large number of virtual function calls, conditional branch prediction failures, and so on, so Databend doesn't use it, but I'm not sure what the advantages of the Databend execution model are right now, so please help me introduce it. I have read this article, but I still don't quite understand the advantages of DAG Scheduler. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
Datafusion is iterating on a new execution model, which you can learn more about in this document: https://docs.google.com/document/d/1txX60thXn1tQO1ENNT8rwfU3cXLofa7ZccnvP4jD6AA/edit#heading=h.3iwlbn2gzs29 The new execution model has many similarities to the current execution model of databend (both largely avoiding the drawbacks of the valcano model) If you want to know more about the advantages of DAG Scheduler in databend, I'll ping @zhang2014 for help :) |
Beta Was this translation helpful? Give feedback.
-
|
Thanks @xudong963. |
Beta Was this translation helpful? Give feedback.
Datafusion is iterating on a new execution model, which you can learn more about in this document: https://docs.google.com/document/d/1txX60thXn1tQO1ENNT8rwfU3cXLofa7ZccnvP4jD6AA/edit#heading=h.3iwlbn2gzs29
The new execution model has many similarities to the current execution model of databend (both largely avoiding the drawbacks of the valcano model)
If you want to know more about the advantages of DAG Scheduler in databend, I'll ping @zhang2014 for help :)