Tim Armstrong has posted comments on this change. Change subject: IMPALA-3203: Part 2: per-core free lists in buffer pool ......................................................................
Patch Set 10: (7 comments) Addressed Taras's comments and rebased onto the IMPALA-5113 fix http://gerrit.cloudera.org:8080/#/c/6414/10/be/src/common/init.cc File be/src/common/init.cc: PS10, Line 96: Maintenance thread > Log Maintenance thread Done PS10, Line 97: 1) > It's kind of weird to have "1)" here. Maybe condense everything into a sing Done PS10, Line 224: from > for Done http://gerrit.cloudera.org:8080/#/c/6414/10/be/src/runtime/bufferpool/system-allocator.cc File be/src/runtime/bufferpool/system-allocator.cc: Line 28: DCHECK_GE(len, min_buffer_len_); > Should we also check that len is smaller than some maximum? Done Line 29: DCHECK(BitUtil::IsPowerOf2(len)) << len; > In the .h file comment you mentioned that len must be a power of two multip It was sort-of implied that min_buffer_len is a power-of-two. added a DCHECK in the constructor for that. Give that assumption the two DCHECKs imply that len is a multiple of min_buffer_len, because min_buffer_len = 2^n and len = 2^(n+m) implies that len is divisible by min_buffer_len. http://gerrit.cloudera.org:8080/#/c/6414/10/be/src/runtime/mem-tracker.cc File be/src/runtime/mem-tracker.cc: PS10, Line 312: curr_consumption - max_consumption + EXTRA_BYTES_TO_FREE > I think it would be clearer if the result of this calculation was put into Done Line 315: if (curr_consumption <= max_consumption) break; > Should we take into account EXTRA_BYTES_TO_FREE here? Good point -done (although reordered the calculation slightly so that it matches the order of varaibles in the bytes_to_free calculation). -- To view, visit http://gerrit.cloudera.org:8080/6414 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I612bd1cd0f0e87f7d8186e5bedd53a22f2d80832 Gerrit-PatchSet: 10 Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-Owner: Tim Armstrong <[email protected]> Gerrit-Reviewer: Taras Bobrovytsky <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: Yes
