-
Notifications
You must be signed in to change notification settings - Fork 26
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
feat: Support unaligned block memory accesses #718
Conversation
INT-2494 Implement new memory data structure
Needs to support indexing at arbitrary locations, not just block-aligned |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
4bc96b5
to
7ce8c8d
Compare
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
7ce8c8d
to
df5659c
Compare
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
a781f09
to
e4095e1
Compare
Replaces previous Memory data structure with a new one. The new data structure represents the partition via a BST (using BTreeMap) rather than a segment tree. The new implementation is signficantly simpler than the previous one. It also separates the partition representation from the underlying data, reducing the amount of data copying.
e4095e1
to
51ea5f0
Compare
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
Detailed Metrics
Commit: a10d82e |
Detailed Metrics
Commit: a10d82e |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The BTreeMap usage could probably be optimized with more iterator usage, however I added a criterion benchmark with flamegraph of fibonacci runtime, and the flamegraph now indicates memory access does not dominate the execution, so there is no need to overoptimize this.
Detailed Metrics
Commit: a10d82e |
* feat: Support unaligned block memory accesses Replaces previous Memory data structure with a new one. The new data structure represents the partition via a BST (using BTreeMap) rather than a segment tree. The new implementation is signficantly simpler than the previous one. It also separates the partition representation from the underlying data, reducing the amount of data copying. * Use hashmap and other small changes * Reduce # of insertions and deletions * Use usize for address space * chore: switch to FxHashMap * feat: add fibonacci execute criterion benchmark and more FxHashMap * chore: use FxHashSet * chore: update readme * chore: update comment --------- Co-authored-by: Jonathan Wang <[email protected]>
* feat: Support unaligned block memory accesses Replaces previous Memory data structure with a new one. The new data structure represents the partition via a BST (using BTreeMap) rather than a segment tree. The new implementation is signficantly simpler than the previous one. It also separates the partition representation from the underlying data, reducing the amount of data copying. * Use hashmap and other small changes * Reduce # of insertions and deletions * Use usize for address space * chore: switch to FxHashMap * feat: add fibonacci execute criterion benchmark and more FxHashMap * chore: use FxHashSet * chore: update readme * chore: update comment --------- Co-authored-by: Jonathan Wang <[email protected]>
Replaces previous Memory data structure with a new one. The new data structure represents the partition via a BST (using BTreeMap) rather than a segment tree. The new implementation is signficantly simpler than the previous one. It also separates the partition representation from the underlying data, reducing the amount of data copying.
Closes INT-2494