Hi Doug, I have seen this. This may improve the alignment performance than the earlier code. Actually, I think alignment for CellMap is useless as I mentioned before. We alloc the CellMap's memory from the end of the buffer(This address is aligned natrually). As long as the CellMap's element's size is aligned, the memory is aligned.
2008/12/19 Doug Judd <[email protected]> > Thanks, Phoenix. I'll add this fix into my next commit. BTW, not sure if > you saw this, but I fixed the CellCachePool alignment logic with the > following code: > > if (is_map) { > unsigned long alignment_offset = (unsigned long)(m_tail_ptr-size) > & (CCP_WORD_SIZE-1); > if (alignment_offset) > size += alignment_offset; > } > > This did seem to give the system a minor performance improvement. > > - Doug > > > On Thu, Dec 18, 2008 at 9:02 AM, Phoenix <[email protected]>wrote: > >> >> Hi Doug, >> >> In src/cc/Hypertable/RangeServer/MetadataRoot.cc:78 >> MetadataRoot::get_next_files >> >> The parameter ag_name is not assigned a valid value which would cause >> function Range::load_cell_stores (src/cc/Hypertable/RangeServer/ >> Range.c:137) to fail when the root range is recovering from its >> CellStores. <Though not used at present> >> >> We should add a line in the function. >> ... >> String attrname = (String)"files." + m_agnames[m_next]; >> ag_name = m_agnames[m_next]; // add this line >> m_next++; >> ... >> >> >> >> > > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Hypertable Development" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/hypertable-dev?hl=en -~----------~----~----~----~------~----~------~--~---
