-
Notifications
You must be signed in to change notification settings - Fork 127
CLI for measuring execute_cuda encoding perf #6381
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
base: develop
Are you sure you want to change the base?
Conversation
0249a54 to
a4c923c
Compare
| let compressor = BtrBlocksCompressorBuilder::empty() | ||
| .include_int([ | ||
| IntCode::Uncompressed, | ||
| IntCode::Constant, | ||
| IntCode::BitPacking, | ||
| IntCode::For, | ||
| IntCode::Sequence, | ||
| IntCode::ZigZag, | ||
| IntCode::Dict, | ||
| ]) | ||
| .include_float([ | ||
| FloatCode::Uncompressed, | ||
| FloatCode::Constant, | ||
| FloatCode::Alp, | ||
| FloatCode::AlpRd, | ||
| FloatCode::RunEnd, | ||
| ]) | ||
| // Don't compress strings, this is b/c we don't have any BtrBlocks encodings that support | ||
| // strings. | ||
| .include_string([StringCode::Uncompressed]) | ||
| .build(); |
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.
I @onursatici added something like this?
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.
yea @a10y you should be able to just use this for the strategy: https://github.com/vortex-data/vortex/blob/develop/vortex-file/src/strategy.rs#L184
Signed-off-by: Andrew Duffy <andrew@a10y.dev>
a4c923c to
74926d3
Compare
Signed-off-by: Andrew Duffy <andrew@a10y.dev>
fe10865 to
e52bb67
Compare
Signed-off-by: Andrew Duffy <andrew@a10y.dev>
e52bb67 to
21537a6
Compare
Signed-off-by: Andrew Duffy <andrew@a10y.dev>
Signed-off-by: Andrew Duffy <andrew@a10y.dev>
This adds a new binary which takes as input a Vortex file with values.
It will recompress the file using only encodings that we implemented on the GPU, scan it back, and collect timings for how long each column scan took. The results are printed as a CSV to stdout, which can be piped into duckdb or similar for analysis
Example usage
Taking the CSV file and loading it into DuckDB: