Use mixin for binops #4
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…pidsai#10317) In file `jit/cache.cpp`, a program cache always internally reserves a `std::unordered_map` using a size set by an environment variable `LIBCUDF_KERNEL_CACHE_LIMIT_PER_PROCESS`. If that environment variable does not exist, a default value (`std::numeric_limit<size_t>::max`) is used. Such default value is huge, leading to allocating a huge (impossible) size of memory chunk that crashes the system. This PR changes that default value from `std::numeric_limit<size_t>::max` to `1024^2`. This is essentially a reverse of the PR rapidsai#10312 but set the default value to `1024` instead of `100`. Note that `1024^2` is just some random number, not based on any specific calculation. Closes rapidsai#10312 and closes rapidsai#9362. Authors: - Nghia Truong (https://github.com/ttnghia) Approvers: - Mark Harris (https://github.com/harrism) - Karthikeyan (https://github.com/karthikeyann) URL: rapidsai#10317
addresses parts of rapidsai#5773 - Add `create_sequence_table` which creates sequences in device (only numeric types supported) with/without nulls. - Add `create_random_null_mask` to create random null mask with given probability. (0.0-1.0 null probability) ~- add gnu++17 to generate_input.cu (temporarily for int128 STL support).~ - renamed `repeat_dtypes` to `cycle_dtypes` and moved out of create_* methods - updated ast bench, search, scatter , binary ops bench Splitting PR rapidsai#10109 for review Authors: - Karthikeyan (https://github.com/karthikeyann) Approvers: - Conor Hoekstra (https://github.com/codereport) - Nghia Truong (https://github.com/ttnghia) - Robert Maynard (https://github.com/robertmaynard) - MithunR (https://github.com/mythrocks) - Bradley Dice (https://github.com/bdice) URL: rapidsai#10300
This PR removes `__pycache__`, `.so` & `.pyc` files. Authors: - GALI PREM SAGAR (https://github.com/galipremsagar) Approvers: - Vyas Ramasubramani (https://github.com/vyasr) URL: rapidsai#10355
This PR catches or silences warnings in `test_categorical.py`. (I am working through one test file at a time so we can enable `-Werr` in the future.) Most of the warnings come from deprecated `inplace` arguments to pandas' categorical functions. The `inplace` argument will be removed in pandas 2.0. Until then, we should just hide the warning. Additionally, I refactored some `inplace` behavior to make the expected behavior of the test clearer. Authors: - Bradley Dice (https://github.com/bdice) Approvers: - Ashwin Srinath (https://github.com/shwina) URL: rapidsai#10354
…#10346) This PR builds on rapidsai#10217 and rapidsai#10287 to bring full ufunc support for Index types, expanding well beyond the small set previously supported in the `cudf.core.ops` namespace. By using most of the machinery introduced for IndexedFrame in the prior two PRs we avoid duplicating much logic so that all ufunc dispatches flow through a relatively standard path of known methods prior to a common cupy dispatch. With this change we are also able to deprecate the various ufunc operations defined in cudf/core/ops.py that exist only for this purpose as well as a number of Frame methods that are not defined for the corresponding pandas types. Users of those APIs are recommended to calling the corresponding numpy/cupy ufuncs instead to leverage the new dispatch. This PR also fixes a bug where index binary operations that output booleans would previously return instances of GenericIndex, whereas those pandas operations would return numpy arrays. cudf now returns cupy arrays in those cases. Resolves rapidsai#9083. Contributes to rapidsai#9038. Authors: - Vyas Ramasubramani (https://github.com/vyasr) Approvers: - GALI PREM SAGAR (https://github.com/galipremsagar) URL: rapidsai#10346
This PR implements a factory for mixins classes based on the common pattern in cuDF of categories of similar functions all calling a common method implementing some standard pre/post-processing before calling a lower-level API of either one of its members (e.g. `Frames` calling `Column` methods) or the C++ libcudf library. When added to another class, these mixins support customization of which methods are exposed via a class member set of method names. Documentation for these methods is generated by formatting the docstring for the common internal method, e.g. `_reduce` for reductions. As a first pass, this PR generates a single mixin for reductions and applies it to all the relevant classes. Future PRs will use this to generate classes for scans, binary operations, and unary operations, and perhaps other similar categories as they are uncovered. This approach assumes a great deal of API homogeneity between the different methods in a category. `Frame` violates this assumption because similar operations often support slightly different parameters (for instance, some reductions support a `min_count` parameter), so for now `Frame` was not made `Reducible`. That decision could be revisited if 1) the degree of homogeneity of these function signatures increases over time, or 2) we can introduce greater customization into these mixins without adding too much complexity. A first attempt of (2) can be seen in [this branch](https://github.com/vyasr/cudf/tree/refactor/reductions_extended), but the degree of additional complexity just to support `Frame` isn't really justifiable at this stage, so unless we can come up with a simpler solution I recommend leaving `Frame` as is for now. Authors: - Vyas Ramasubramani (https://github.com/vyasr) - Ashwin Srinath (https://github.com/shwina) Approvers: - Michael Wang (https://github.com/isVoid) - Ashwin Srinath (https://github.com/shwina) URL: rapidsai#9925
… process op names.
vyasr
force-pushed
the
refactor/binops_mixin
branch
from
1025fea
to
70a13a0
Compare
vyasr
pushed a commit
that referenced
this pull request
Jun 13, 2023
This implements stacktrace and adds a stacktrace string into any exception thrown by cudf. By doing so, the exception carries information about where it originated, allowing the downstream application to trace back with much less effort. Closes rapidsai#12422. ### Example: ``` #0: cudf/cpp/build/libcudf.so : std::unique_ptr<cudf::column, std::default_delete<cudf::column> > cudf::detail::sorted_order<false>(cudf::table_view, std::vector<cudf::order, std::allocator<cudf::order> > const&, std::vector<cudf::null_order, std::allocator<cudf::null_order> > const&, rmm::cuda_stream_view, rmm::mr::device_memory_resource*)+0x446 #1: cudf/cpp/build/libcudf.so : cudf::detail::sorted_order(cudf::table_view const&, std::vector<cudf::order, std::allocator<cudf::order> > const&, std::vector<cudf::null_order, std::allocator<cudf::null_order> > const&, rmm::cuda_stream_view, rmm::mr::device_memory_resource*)+0x113 #2: cudf/cpp/build/libcudf.so : std::unique_ptr<cudf::column, std::default_delete<cudf::column> > cudf::detail::segmented_sorted_order_common<(cudf::detail::sort_method)1>(cudf::table_view const&, cudf::column_view const&, std::vector<cudf::order, std::allocator<cudf::order> > const&, std::vector<cudf::null_order, std::allocator<cudf::null_order> > const&, rmm::cuda_stream_view, rmm::mr::device_memory_resource*)+0x66e #3: cudf/cpp/build/libcudf.so : cudf::detail::segmented_sort_by_key(cudf::table_view const&, cudf::table_view const&, cudf::column_view const&, std::vector<cudf::order, std::allocator<cudf::order> > const&, std::vector<cudf::null_order, std::allocator<cudf::null_order> > const&, rmm::cuda_stream_view, rmm::mr::device_memory_resource*)+0x88 #4: cudf/cpp/build/libcudf.so : cudf::segmented_sort_by_key(cudf::table_view const&, cudf::table_view const&, cudf::column_view const&, std::vector<cudf::order, std::allocator<cudf::order> > const&, std::vector<cudf::null_order, std::allocator<cudf::null_order> > const&, rmm::mr::device_memory_resource*)+0xb9 #5: cudf/cpp/build/gtests/SORT_TEST : ()+0xe3027 rapidsai#6: cudf/cpp/build/lib/libgtest.so.1.13.0 : void testing::internal::HandleExceptionsInMethodIfSupported<testing::Test, void>(testing::Test*, void (testing::Test::*)(), char const*)+0x8f rapidsai#7: cudf/cpp/build/lib/libgtest.so.1.13.0 : testing::Test::Run()+0xd6 rapidsai#8: cudf/cpp/build/lib/libgtest.so.1.13.0 : testing::TestInfo::Run()+0x195 rapidsai#9: cudf/cpp/build/lib/libgtest.so.1.13.0 : testing::TestSuite::Run()+0x109 rapidsai#10: cudf/cpp/build/lib/libgtest.so.1.13.0 : testing::internal::UnitTestImpl::RunAllTests()+0x44f rapidsai#11: cudf/cpp/build/lib/libgtest.so.1.13.0 : bool testing::internal::HandleExceptionsInMethodIfSupported<testing::internal::UnitTestImpl, bool>(testing::internal::UnitTestImpl*, bool (testing::internal::UnitTestImpl::*)(), char const*)+0x87 rapidsai#12: cudf/cpp/build/lib/libgtest.so.1.13.0 : testing::UnitTest::Run()+0x95 rapidsai#13: cudf/cpp/build/gtests/SORT_TEST : ()+0xdb08c rapidsai#14: /lib/x86_64-linux-gnu/libc.so.6 : ()+0x29d90 rapidsai#15: /lib/x86_64-linux-gnu/libc.so.6 : __libc_start_main()+0x80 rapidsai#16: cudf/cpp/build/gtests/SORT_TEST : ()+0xdf3d5 ``` ### Usage In order to retrieve a stacktrace with fully human-readable symbols, some compiling options must be adjusted. To make such adjustment convenient and effortless, a new cmake option (`CUDF_BUILD_STACKTRACE_DEBUG`) has been added. Just set this option to `ON` before building cudf and it will be ready to use. For downstream applications, whenever a cudf-type exception is thrown, it can retrieve the stored stacktrace and do whatever it wants with it. For example: ``` try { // cudf API calls } catch (cudf::logic_error const& e) { std::cout << e.what() << std::endl; std::cout << e.stacktrace() << std::endl; throw e; } // similar with catching other exception types ``` ### Follow-up work The next step would be patching `rmm` to attach stacktrace into `rmm::` exceptions. Doing so will allow debugging various memory exceptions thrown from libcudf using their stacktrace. ### Note: * This feature doesn't require libcudf to be built in Debug mode. * The flag `CUDF_BUILD_STACKTRACE_DEBUG` should not be turned on in production as it may affect code optimization. Instead, libcudf compiled with that flag turned on should be used only when needed, when debugging cudf throwing exceptions. * This flag removes the current optimization flag from compiling (such as `-O2` or `-O3`, if in Release mode) and replaces by `-Og` (optimize for debugging). * If this option is not set to `ON`, the stacktrace will not be available. This is to avoid expensive stracktrace retrieval if the throwing exception is expected. Authors: - Nghia Truong (https://github.com/ttnghia) Approvers: - AJ Schmidt (https://github.com/ajschmidt8) - Robert Maynard (https://github.com/robertmaynard) - Vyas Ramasubramani (https://github.com/vyasr) - Jason Lowe (https://github.com/jlowe) URL: rapidsai#13298
vyasr
pushed a commit
that referenced
this pull request
Sep 22, 2023
Pin conda packages to `aws-sdk-cpp<1.11`. The recent upgrade in version `1.11.*` has caused several issues with cleaning up (more details on changes can be read in [this link](https://github.com/aws/aws-sdk-cpp#version-111-is-now-available)), leading to Distributed and Dask-CUDA processes to segfault. The stack for one of those crashes looks like the following: ``` (gdb) bt #0 0x00007f5125359a0c in Aws::Utils::Logging::s_aws_logger_redirect_get_log_level(aws_logger*, unsigned int) () from /opt/conda/envs/dask/lib/python3.9/site-packages/pyarrow/../../.././libaws-cpp-sdk-core.so #1 0x00007f5124968f83 in aws_event_loop_thread () from /opt/conda/envs/dask/lib/python3.9/site-packages/pyarrow/../../../././libaws-c-io.so.1.0.0 #2 0x00007f5124ad9359 in thread_fn () from /opt/conda/envs/dask/lib/python3.9/site-packages/pyarrow/../../../././libaws-c-common.so.1 #3 0x00007f519958f6db in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0 #4 0x00007f5198b1361f in clone () from /lib/x86_64-linux-gnu/libc.so.6 ``` Such segfaults now manifest frequently in CI, and in some cases are reproducible with a hit rate of ~30%. Given the approaching release time, it's probably the safest option to just pin to an older version of the package while we don't pinpoint the exact cause for the issue and a patched build is released upstream. The `aws-sdk-cpp` is statically-linked in the `pyarrow` pip package, which prevents us from using the same pinning technique. cuDF is currently pinned to `pyarrow=12.0.1` which seems to be built against `aws-sdk-cpp=1.10.*`, as per [recent build logs](https://github.com/apache/arrow/actions/runs/6276453828/job/17046177335?pr=37792#step:6:1372). Authors: - Peter Andreas Entschev (https://github.com/pentschev) Approvers: - GALI PREM SAGAR (https://github.com/galipremsagar) - Ray Douglass (https://github.com/raydouglass) URL: rapidsai#14173
galipremsagar
added a commit
that referenced
this pull request
Nov 8, 2023
…#4) Fixes: rapidsai#14148 This PR resolves a RunTimeError by raising when a pd.PeriodIndex is passed to the column constructor, the reason being there is no cudf.PeriodIndex implemented yet.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This builds on rapidsai#9925, using that approach to define binary operations. Note that I plan to apply this mixin to ColumnBase as well, but that will require more work to clean up binary operations for column types, which are currently quite a mess. This PR will also be affected by some of the changes in rapidsai#10346, which rewrites some binops internals.