You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
To run Sirius CLI, simply start the shell with `./build/release/duckdb {DATABASE_NAME}.duckdb`.
171
190
From the duckdb shell, initialize the Sirius buffer manager with `call gpu_buffer_init`. This API accepts 2 parameters, the GPU caching region size and the GPU processing region size. The GPU caching region is a memory region where the raw data is stored in GPUs, whereas the GPU processing region is where intermediate results are stored in GPUs (hash tables, join results .etc).
@@ -248,11 +267,17 @@ con.execute('''
248
267
```
249
268
250
269
## Correctness Testing
251
-
Sirius provides a unit test that compares Sirius against DuckDB for correctness across all 22 TPC-H queries. To run the unittest, generate SF=1 TPC-H dataset using the method described [here](https://github.com/sirius-db/sirius?tab=readme-ov-file#generating-tpc-h-dataset) and run the unittest using the following command:
270
+
Sirius provides a unit test that compares Sirius against DuckDB for correctness across many test queries. To run the unittest, generate the datasets using the method described [here](https://github.com/sirius-db/sirius?tab=readme-ov-file#generating-tpc-h-dataset) and run the unittest using the following command:
252
271
```
253
272
make test
254
273
```
255
274
275
+
To run a specific test run the command from the root directory:
Make sure to build the duckdb-python package before running this test using the method described [here](https://github.com/sirius-db/sirius?tab=readme-ov-file#building-sirius). To test Sirius performance against DuckDB across all 22 TPC-H queries, run the following command (replace {SF} with the desired scale factor):
0 commit comments