Skip to content

Add Apple GPU backend for count_overlaps #152

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.

Sign up for GitHub

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

Open
mwiewior wants to merge 6 commits into
base: master
Choose a base branch
from
Open

Add Apple GPU backend for count_overlaps #152

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
a68addc
Add Apple GPU count_overlaps backend
mwiewior Apr 27, 2026
937d89a
Support count_overlaps projection pushdown
mwiewior Apr 27, 2026
509b51e
Add count_overlaps benchmark timings
mwiewior Apr 27, 2026
25b28aa
Prune count_overlaps right-side columns
mwiewior Apr 27, 2026
239d976
Optimize Apple GPU count_overlaps batching
mwiewior Apr 27, 2026
00f5795
Optimize count_overlaps rank coordinate access
mwiewior Apr 27, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
72 changes: 72 additions & 0 deletions Cargo.lock
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

25 changes: 25 additions & 0 deletions datafusion/bio-function-ranges/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,14 @@ repository.workspace = true
homepage.workspace = true
edition.workspace = true

[features]
default = []
apple-gpu = [
"dep:objc2",
"dep:objc2-foundation",
"dep:objc2-metal",
]

[dependencies]
datafusion = { workspace = true }
tokio = { workspace = true }
Expand All @@ -21,6 +29,23 @@ rust-lapper = "1.1.0"
superintervals = { path = "superintervals" }
parking_lot = "0.12.3"

[target.'cfg(target_os = "macos")'.dependencies]
objc2 = { version = "0.6.4", optional = true }
objc2-foundation = { version = "0.3.2", optional = true, features = ["NSError", "NSString"] }
objc2-metal = { version = "0.3.2", optional = true, default-features = false, features = [
"MTLAllocation",
"MTLBuffer",
"MTLCommandBuffer",
"MTLCommandEncoder",
"MTLCommandQueue",
"MTLComputeCommandEncoder",
"MTLComputePipeline",
"MTLDevice",
"MTLLibrary",
"MTLResource",
"MTLTypes",
] }

[dev-dependencies]
tokio = { workspace = true, features = ["rt-multi-thread", "macros"] }
rstest = "0.26.1"
10 changes: 10 additions & 0 deletions datafusion/bio-function-ranges/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -89,6 +89,16 @@ Counts overlapping intervals. Same interface as `coverage`, but returns the coun
SELECT * FROM count_overlaps('reads', 'targets')
```

When built with the `apple-gpu` Cargo feature on macOS, `count_overlaps` can use an Apple Metal endpoint-rank backend:

```sql
SET bio.count_overlaps_backend = auto; -- default
SET bio.count_overlaps_backend = cpu;
SET bio.count_overlaps_backend = apple_gpu;
```

`auto` keeps the CPU COITree backend for small inputs and currently tries the Apple GPU backend for large non-coverage `count_overlaps` workloads when the collected left side has at least 5M intervals. If Metal initialization fails before output is emitted, execution falls back to CPU.

### `nearest(left_table, right_table [, k] [, overlap] [, columns...] [, filter_op])`

Returns up to `k` nearest left intervals for each right interval.
Expand Down
Loading
Loading