[BUG] 25.02 submodule sync failed apply patch disable_shared_agg.patch

[pxLi]

Describe the bug
first seen at, spark-rapids-jni_submodule-sync-pre_release run:1013

with cudf commit 4b2ce98745187f1367cc8c461c6a6a18f42e2d1b

[2025-02-11T22:59:16.409Z] [INFO] --- maven-antrun-plugin:3.0.0:run (cudf patch) @ spark-rapids-jni ---

[2025-02-11T22:59:16.409Z] [INFO] Executing tasks

[2025-02-11T22:59:16.409Z] [INFO]      [exec] /home/jenkins/agent/workspace/jenkins-spark-rapids-jni_submodule-sync-pre_release-1013/thirdparty/cudf /home/jenkins/agent/workspace/jenkins-spark-rapids-jni_submodule-sync-pre_release-1013

[2025-02-11T22:59:16.409Z] [INFO]      [exec] patching with: /home/jenkins/agent/workspace/jenkins-spark-rapids-jni_submodule-sync-pre_release-1013/patches/disable_shared_agg.patch

[2025-02-11T22:59:16.409Z] [INFO]      [exec] Checking patch cpp/src/groupby/hash/compute_aggregations.cuh...

[2025-02-11T22:59:16.409Z] [INFO]      [exec] error: while searching for:

[2025-02-11T22:59:16.409Z] [INFO]      [exec]   auto const available_shmem_size = get_available_shared_memory_size(grid_size);

[2025-02-11T22:59:16.409Z] [INFO]      [exec]   auto const offsets_buffer_size  = compute_shmem_offsets_size(flattened_values.num_columns()) * 2;

[2025-02-11T22:59:16.409Z] [INFO]      [exec]   auto const data_buffer_size     = available_shmem_size - offsets_buffer_size;

[2025-02-11T22:59:16.409Z] [INFO]      [exec]   auto const is_shared_memory_compatible = std::all_of(

[2025-02-11T22:59:16.409Z] [INFO]      [exec]     requests.begin(), requests.end(), [&](cudf::groupby::aggregation_request const& request) {

[2025-02-11T22:59:16.409Z] [INFO]      [exec]       if (cudf::is_dictionary(request.values.type())) { return false; }

[2025-02-11T22:59:16.409Z] [INFO]      [exec]       // Ensure there is enough buffer space to store local aggregations up to the max cardinality

[2025-02-11T22:59:16.409Z] [INFO]      [exec]       // for shared memory aggregations

[2025-02-11T22:59:16.409Z] [INFO]      [exec]       auto const size = cudf::type_dispatcher<cudf::dispatch_storage_type>(request.values.type(),

[2025-02-11T22:59:16.409Z] [INFO]      [exec]                                                                            size_of_functor{});

[2025-02-11T22:59:16.409Z] [INFO]      [exec]       return static_cast<size_type>(data_buffer_size) >= (size * GROUPBY_CARDINALITY_THRESHOLD);

[2025-02-11T22:59:16.409Z] [INFO]      [exec]     });

[2025-02-11T22:59:16.409Z] [INFO]      [exec] 

[2025-02-11T22:59:16.409Z] [INFO]      [exec]   // Performs naive global memory aggregations when the workload is not compatible with shared

[2025-02-11T22:59:16.409Z] [INFO]      [exec]   // memory, such as when aggregating dictionary columns or when there is insufficient dynamic

[2025-02-11T22:59:16.409Z] [INFO]      [exec] 

[2025-02-11T22:59:16.409Z] [INFO]      [exec] error: patch failed: cpp/src/groupby/hash/compute_aggregations.cuh:69

[2025-02-11T22:59:16.409Z] [INFO]      [exec] error: cpp/src/groupby/hash/compute_aggregations.cuh: patch does not apply

Steps/Code to reproduce bug
Please provide a list of steps or a code sample to reproduce the issue.
Avoid posting private or sensitive data.

Expected behavior
A clear and concise description of what you expected to happen.

Environment details (please complete the following information)

• Environment location: [Standalone, YARN, Kubernetes, Cloud(specify cloud provider)]
• Spark configuration settings related to the issue

Additional context
Add any other context about the problem here.

[pxLi]

context
#2909
#2910

[pxLi]

closed with #2912