Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Do not merge 2: Everything Else that is needed for the evaluation paper. #419

Closed
wants to merge 37 commits into from

Conversation

joka921
Copy link
Member

@joka921 joka921 commented Jun 24, 2021

Currently includes the parallel sort via STL #418 , the parallel Join #417 and the parallel pattern Trick #342

-The degree of parallelism is configurable at compiletime
Since it is task-based it also performs well under heavy load.

-Updated Dockerfile to use Ubuntu 20.04. This is required for sufficient OpenMP support.
* Since C++17, the STL supports parallel Algorithms like std::sort.

* This commit uses this sort as the new default.

* For "dynamic" IdTables (IdTable<0>) we still use the GNU parallel extensions, because the IDTables do not support the STL sorting currently (TODO: figure out why, and remove the reason).
This allows more efficient sorting also for intermediate results with a larger number of columns.
… the input into chunks.

TODO: test, and tweak the performance.
…a std::vector (or similar data structure)

which can cheaply create new Elements (like ints) without initializing (zeroing) them.

This makes e.g. resizing the Vector to a bigger size much cheaper.
This commit allows the export of query results directly in the ID space in a dense binary format.
This is very useful for debugging.
(This is actually a separate PR which should be merged before this).
* Integrated the Zstandard compression library.

* Compress the PSO and POS relations in Blocks of 8 MBs.

* When performing an Index Scan, there is much less access to disk and the decompression of the Blocks is performed in parallel.
This allows using/building a docker container without recompilation in case of a test failure.
Also comment out the SortPerformanceEstimatorTest because it has spurious failures which are nasty.
@joka921 joka921 closed this Feb 28, 2024
@joka921 joka921 deleted the f.evaluationPaper2 branch February 28, 2024 13:06
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants