Vukasin Milovanovic

Results 33 issues of Vukasin Milovanovic

## Description ## Checklist - [x] I am familiar with the [Contributing Guidelines](https://github.com/rapidsai/cudf/blob/HEAD/CONTRIBUTING.md). - [x] New or existing tests cover these changes. - [ ] The documentation is up to...

libcudf
improvement
non-breaking

When reading binary files that have independently encoded segments (e.g. pages, row groups) we often end up reading each unit of encoding separately. While readers coalesce adjacent read calls, this...

feature request
cuIO

## Description ## Checklist - [ ] I am familiar with the [Contributing Guidelines](https://github.com/rapidsai/cudf/blob/HEAD/CONTRIBUTING.md). - [ ] New or existing tests cover these changes. - [ ] The documentation is...

feature request
libcudf
Python
CMake
Java
non-breaking

## Description Fixed the logic in the CSV reader that led to empty output instead of producing a table with a single column and one row. Also did some small...

bug
libcudf
Python
cuIO
non-breaking

## Description Upgrade to nvcomp 4.0. As part of that change we need to refactor the `nvcomp.cmake` to be aware of the new CMake files and targets that nvcomp generates....

feature request
non-breaking
5 - DO NOT MERGE

## Description TBD ## Checklist - [x] I am familiar with the [Contributing Guidelines](https://github.com/rapidsai/cudf/blob/HEAD/CONTRIBUTING.md). - [x] New or existing tests cover these changes. - [ ] The documentation is up...

libcudf
Performance
improvement
non-breaking

## Description Some random data generators in cuDF default to `seed=None`, which means that an OS or time dependent seed is used, leading to different test data between systems/runs. This...

0 - Waiting on Author
tests
Python
improvement
non-breaking

Benchmarks should cover different workloads (size/number of reads/writes). Multi-threaded benchmarks can also be considered. Is possible, the throughput can be compared to the results from utilities like fio and gdsio.

feature request
non-breaking
c++

Memory mapping could be faster in some use cases and/or some systems. This is a proposal to add an option to replace file reads with copying from the mapped-to buffer....

The host resource would be used for the bounce buffer in `value()`. The potential optimization would be to pass a pinned memory resource to avoid a pageable copy in `value()`....

feature request