diff options
author | David Oberhollenzer <david.oberhollenzer@sigma-star.at> | 2020-02-21 03:28:55 +0100 |
---|---|---|
committer | David Oberhollenzer <david.oberhollenzer@sigma-star.at> | 2020-02-21 03:30:07 +0100 |
commit | 0841eb83042b59dfabed03f664d0a7b3d5921f59 (patch) | |
tree | e662f5d5e9f89a8571849a4bcb44b24218da5808 /doc | |
parent | 4c10e8326447216c7d79aa87668a7b00efdb7364 (diff) |
Update documentation
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
Diffstat (limited to 'doc')
-rw-r--r-- | doc/parallelism.txt | 96 |
1 files changed, 31 insertions, 65 deletions
diff --git a/doc/parallelism.txt b/doc/parallelism.txt index e509d86..142444f 100644 --- a/doc/parallelism.txt +++ b/doc/parallelism.txt @@ -60,80 +60,46 @@ - Completed blocks are inserted into a "done" queue, sorted by their sequence number. - The main thread that submits blocks also dequeues the completed ones, - keeping track of the sequence numbers, and calls the respective common - functions for processing completed blocks and fragments. - - If a fragment block is created, it is submitted with *the same* sequence - number as the fragment that caused it to overflow and the next expected - sequence number is reset to that. - - To make sure the queue doesn't fill all RAM, submitted blocks are counted. - The counter is decremented when dequeueing completed blocks. If it reaches - a maximum, signal/await is used to wait for the worker threads to complete - some blocks to process. Similarly, the worker threads use signal/await to - wait on the queue if it is empty. - - - 1.1) Problems - - The outlined approach performs sub-optimal, with an efficiency somewhere - between 50% to 75% on the bench mark data used. - - Profiling using perf shows that almost a third of the time, only one - worker thread is actually active, while the others are waiting. - - The current hypothesis is that this is caused by many small input files - being processed, causing a work load consisting primarily of fragments. - - - Fragments are only hashed, not compressed, so the work is - primarily I/O bound. - - After a number of fragments are consumed, a fragment block is created. - - The fragment block is submitted to the almost empty queue and the - I/O thread waits for it to be completed before doing anything else. - - One thread gets to handle the fragment block, which involves a lot more - work. Meanwhile the other threads starve on the empty queue. - - After that has finally been handed of to the I/O thread, another burst of - fragments comes in. - - Rinse and repeat. - - - 1.2) Proposed Solution - - It makes no sense for the main thread to block until the fragment block is - done. It can process further fragments (just not write blocks), creating - more fragment blocks on the way. - - A possible implementation might be to maintain 3 queues instead of 2: - - - A queue with submitted blocks. - - A queue with completed blocks. - - A queue for blocks ready to be written to disk ("I/O queue"). - - A second sequence number is needed for keeping order on the I/O queue: - - - Submit blocks as before with incremental processing sequence number. - - Dequeue completed blocks in order by processing sequence number. - - For regular blocks, add them to the I/O queue with incremental I/O - sequence number. - - For fragments, consolidate them into fragment blocks. On overflow, - dispatch a fragment block with incremental processing sequence number, - BUT give it an I/O queue sequence number NOW. - - For fragment blocks, add them to the I/O queue without allocating an - I/O sequence number, it already has one. - - Dequeue ordered by I/O sequence number from the I/O queue and send the - completed blocks to the block writer. - + in order of their sequence number. + - Data blocks are given an "i/o sequence number" and inserted into + an I/O queue in order by this number. + - For fragments, it and calls the respective function from common.c + to deduplicate and append to the current fragment block. + - If the current fragment block overflows, it is given the next free + **I/O sequence number** and inserted into the **processing queue**. + - Completed fragment blocks are inserted into the I/O queue with + regard to the I/O sequence number they already have + - The main thread also dequeues I/O blocks ordered by their I/O sequence + number, and passes them to the respective common.c function that eventually + forwards it to the block writer. + + To make sure the queue doesn't fill all RAM, submitted blocks are counted. The + counter is also incremented when submitting to the I/O queue. The counter is + decremented when dequeueing completed blocks or blocks from the I/O queue. + + The block submission function tries to add blocks only if the counter is below + a pre-set threshold. If it is above, it first tries to dequeue I/O block, and + if there are no I/O blocks, dequeue and handle "done" blocks. If the "done" + queue is also empty, it uses signal/await to wait for the worker threads to + complete some blocks to process. The main thread in turn uses signal/await to + tell the worker threads it has submitted block (they wait if the work queue is + empty). + + Profiling on small filesystems using perf shows that the outlined approach + seems to perform quite well for CPU bound compressors like XZ, but doesn't + add a lot for I/O bound compressors like zstd. Actual measurements still + need to be done. If you have a more insights or a better idea, please let me know. - 2) Benchmarks ************* TODO: benchmarks with the following images: - Debian live iso (2G) - - Arch Linux live iso (550M) - - Raspberry Pi 3 QT demo image (~300M) + - Arch Linux live iso (~550M) + - Raspberry Pi 3 QT demo image (~390M) sqfs2tar $IMAGE | tar2sqfs -j $NUM_CPU -f out.sqfs |