Michael Ho has posted comments on this change.

Change subject: IMPALA-4432: Handle internal codegen disabling properly
......................................................................


Patch Set 5:

(3 comments)

http://gerrit.cloudera.org:8080/#/c/5105/5/be/src/exec/exec-node.cc
File be/src/exec/exec-node.cc:

PS5, Line 463: CodegenDisabledByHint(
> is this correct? we might still codegen if there's an IR udf, right? See al
Good point. Fixed in the new patch.


http://gerrit.cloudera.org:8080/#/c/5105/5/be/src/runtime/runtime-state.h
File be/src/runtime/runtime-state.h:

PS5, Line 171: codegen is disabled due to hints
> is this true?  doesn't this just indicate whether the hint would like to di
Rephrased to "if there is hint to disable codegen."


PS5, Line 175: CodegenDisabledByHint
> if my above question is right, then this should be called CodegenDisableHin
The new patch introduces both CodegenDisableHint() and CodegenDisabledByHint().

The former is needed in ScalarFnCall::Prepare(). Ideally, if there is hint to 
disable codegen, we should only codegen expressions which cannot be 
interpreted.  If we use CodegenDisabledByHint() (with the suggested change) in 
ScalarFnCall::Prepare() and once an expression is added to 
scalar_fn_calls_to_codegen_, all subsequent ScalarFnCall expressions will be 
codegen'd. This makes it a bit non-deterministic which expressions will be 
codegen'd in the presence of the disabling hints as it depends on the order of 
when Prepare() is called for the non-interpretable expression.

That said, the expressions may still be codegen'd when the operator is 
codegen'd but this logic may affect fe-support.cc and the above may matter in 
that case.


-- 
To view, visit http://gerrit.cloudera.org:8080/5105
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I0b6a9ed723c64ba21b861608583cc9b6607d3397
Gerrit-PatchSet: 5
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Michael Ho <[email protected]>
Gerrit-Reviewer: Dan Hecht <[email protected]>
Gerrit-Reviewer: Michael Ho <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-HasComments: Yes

Reply via email to