Split layouts into a commons crate - #9401
Conversation
Signed-off-by: Adam Gutglick <adam@spiraldb.com>
Merging this PR will regress 3 benchmarks
|
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ❌ | Simulation | decompress[u64, (10000, 4)] |
310.2 µs | 401.6 µs | -22.76% |
| ❌ | Simulation | take[small_m/shuffled/primitive/nonnull/chunks=16384/indices=16] |
1 ms | 1.2 ms | -11.4% |
| ❌ | Simulation | decompress[u64, (1000, 16)] |
64.4 µs | 72.6 µs | -11.3% |
| ⚡ | Simulation | cold_misaligned[(64, 256)] |
5.1 ms | 4.4 ms | +16.53% |
| ⚡ | WallTime | words_gather_scalar[65536] |
9.4 µs | 8.3 µs | +13.67% |
| ⚡ | Simulation | compress_fsst[(1000, 64, 8)] |
1.2 ms | 1 ms | +12.98% |
| ⚡ | Simulation | compress_fsst[(10000, 64, 4)] |
9.7 ms | 8.6 ms | +11.83% |
Tip
Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.
Comparing adamg/layout-split (996bb2c) with develop (93b0535)
Footnotes
-
46 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩
Signed-off-by: Adam Gutglick <adam@spiraldb.com>
|
On naming I like vortex-layout and I don't like core but it matches the editions that these features should be added too. |
|
I also think core is more appropriate here |
Rationale for this change
Step 1 trying to break some of the core crates out into smaller core crates that pull less stuff.
This will both improve compile times and cache usage as the core crates will change less often.
I've tried to make sure the public API of vortex-layout doesn't change, and I'll verify it again before this PR is ready for review.