Remote reads (range-read M2/M3: open_storage, HTTP/S3) and near-complete in-place editing #18

Merged
osobh merged 52 commits from feat/p3-remote-editor into main 2026-09-27 11:15:23 +00:00
Showing only changes of commit c2ae7846c9 - Show all commits
+6 -4
View File
@@ -98,10 +98,12 @@ Measured 2026-09-26 on tank with `cargo test -p clawhdf5-tools --test
edit_interop -- --ignored --nocapture measure_append_waste` (one editor for edit_interop -- --ignored --nocapture measure_append_waste` (one editor for
the whole workload; file sizes are deterministic): 1000 appends of 100 `f8` the whole workload; file sizes are deterministic): 1000 appends of 100 `f8`
values to a 1-D dataset with 1024-element chunks give 810 504 bytes values to a 1-D dataset with 1024-element chunks give 810 504 bytes
unfiltered, as libhdf5's file, and 306 780 bytes with gzip (307 210 before unfiltered, as libhdf5's file (`h5repack` of either: 810 360), and 306 780
reuse; libhdf5: 306 058; `h5repack`: 306 104); 2000 appends of 10 values bytes with gzip (307 210 before reuse; libhdf5's file: 306 058; `h5repack`
with 4096-element gzip chunks give 79 829 bytes (119 684 before reuse) of the editor's file: 306 104, of libhdf5's: 305 954); 2000 appends of 10
against libhdf5's 50 292 (`h5repack`: 49 930): the chunk being appended to values with 4096-element gzip chunks give 79 829 bytes (119 684 before
reuse) against libhdf5's 50 292 (`h5repack` of the editor's file: 49 930,
of libhdf5's: 50 188): the chunk being appended to
is followed by new index blocks and moves each time it grows, and the is followed by new index blocks and moves each time it grows, and the
space it leaves is too small for its next, larger version. space it leaves is too small for its next, larger version.