Drop cuda 11 usages

build.sh

@@ -114,7 +114,7 @@ function buildAll {
 }
 
 function buildLibCudfJniInDocker {
-    local cudaVersion="11.8.0"
+    local cudaVersion="12.9.1"
     local imageName="cudf-build:${cudaVersion}-devel-rocky8"
     local CMAKE_GENERATOR="${CMAKE_GENERATOR:-Ninja}"
     local workspaceDir="/rapids"

docs/cudf/source/cudf_pandas/benchmarks.md

@@ -43,7 +43,7 @@ source pandas/py-pandas/bin/activate
 4. Install cudf:
 
 ```bash
-pip install --extra-index-url=https://pypi.nvidia.com cudf-cu12  # or cudf-cu11
+pip install --extra-index-url=https://pypi.nvidia.com cudf-cu12
 ```
 
 5. Modify pandas join/group code to use `cudf.pandas` and remove the `dtype_backend` keyword argument (not supported):

docs/cudf/source/user_guide/io/io.md

@@ -90,7 +90,7 @@ IO operations.  GDS enables a direct data path for direct memory access
 (DMA) transfers between GPU memory and storage, which avoids a bounce
 buffer through the CPU. The SDK is available for download
 [here](https://developer.nvidia.com/gpudirect-storage). GDS is also
-included in CUDA Toolkit 11.4 and higher.
+included in CUDA Toolkit.
 
 Use of GDS in cuDF is controlled by KvikIO's environment variable `KVIKIO_COMPAT_MODE`. It has
 3 options (case-insensitive):

java/src/test/java/ai/rapids/cudf/RmmTest.java

@@ -547,13 +547,7 @@ public void testPoolSize() {
   @Tag("noSanitizer")
   @Test
   public void testCudaAsyncMemoryResourceSize() {
-    try {
-      Rmm.initialize(RmmAllocationMode.CUDA_ASYNC, Rmm.logToStderr(), 1024);
-    } catch (CudfException e) {
-      // CUDA 11.2 introduced cudaMallocAsync, older CUDA Toolkit will skip this test.
-      assumeFalse(e.getMessage().contains("cudaMallocAsync not supported"));
-      throw e;
-    }
+    Rmm.initialize(RmmAllocationMode.CUDA_ASYNC, Rmm.logToStderr(), 1024);
     try (DeviceMemoryBuffer ignored1 = Rmm.alloc(1024)) {
       assertThrows(OutOfMemoryError.class,
           () -> {

python/cudf/cudf/tests/test_groupby.py

@@ -15,8 +15,6 @@
 from numba import cuda
 from numpy.testing import assert_array_equal
 
-import rmm
-
 import cudf
 from cudf import DataFrame, Series
 from cudf.api.extensions import no_default
@@ -1062,13 +1060,9 @@ def test_groupby_agg_decimal(num_groups, nelem_per_group, func):
     )
 
     expect_df = pdf.groupby("idx", sort=True).agg(func)
-    if rmm._cuda.gpu.runtimeGetVersion() < 11000:
-        with pytest.raises(RuntimeError):
-            got_df = gdf.groupby("idx", sort=True).agg(func)
-    else:
-        got_df = gdf.groupby("idx", sort=True).agg(func)
-        assert_eq(expect_df["x"], got_df["x"], check_dtype=False)
-        assert_eq(expect_df["y"], got_df["y"], check_dtype=False)
+    got_df = gdf.groupby("idx", sort=True).agg(func)
+    assert_eq(expect_df["x"], got_df["x"], check_dtype=False)
+    assert_eq(expect_df["y"], got_df["y"], check_dtype=False)
 
 
 @pytest.mark.parametrize(

python/cudf/cudf/tests/test_parquet.py

@@ -4547,9 +4547,6 @@ def my_pdf(request):
 
 @pytest.mark.parametrize("compression", ["brotli", "gzip", "snappy", "zstd"])
 def test_parquet_decompression(set_decomp_env_vars, my_pdf, compression):
-    if compression == "snappy":
-        pytest.skip("Skipping because of a known issue on CUDA 11.8")
-
     # PANDAS returns category objects whereas cuDF returns hashes
     expect = my_pdf.drop(columns=["col_category"])
 

python/cudf/cudf/utils/gpu_utils.py

@@ -108,14 +108,14 @@ def validate_setup():
 
         cuda_runtime_version = runtimeGetVersion()
 
-        if cuda_runtime_version < 11000:
-            # Require CUDA Runtime version 11.0 or greater.
+        if cuda_runtime_version < 12000:
+            # Require CUDA Runtime version 12.0 or greater.
             major_version = cuda_runtime_version // 1000
             minor_version = (cuda_runtime_version % 1000) // 10
             raise UnsupportedCUDAError(
                 "Detected CUDA Runtime version is "
                 f"{major_version}.{minor_version}. "
-                "Please update your CUDA Runtime to 11.0 or above."
+                "Please update your CUDA Runtime to 12.0 or above."
             )
 
         cuda_driver_supported_rt_version = driverGetVersion()
@@ -142,13 +142,13 @@ def validate_setup():
             # Driver Runtime version is >= Runtime version
             pass
         elif (
-            cuda_driver_supported_rt_version >= 11000
-            and cuda_runtime_version >= 11000
+            cuda_driver_supported_rt_version >= 12000
+            and cuda_runtime_version >= 12000
         ):
             # With cuda enhanced compatibility any code compiled
-            # with 11.x version of cuda can now run on any
-            # driver >= 450.80.02. 11000 is the minimum cuda
-            # version 450.80.02 supports.
+            # with 12.x version of cuda can now run on any
+            # driver >= 525.60.13. 12000 is the minimum cuda
+            # version 525.60.13 supports.
             pass
         else:
             raise UnsupportedCUDAError(

python/cudf_polars/cudf_polars/callback.py

@@ -104,8 +104,7 @@ def default_memory_resource(
         ):
             raise ComputeError(
                 "GPU engine requested, but incorrect cudf-polars package installed. "
-                "If your system has a CUDA 11 driver, please uninstall `cudf-polars-cu12` "
-                "and install `cudf-polars-cu11`"
+                "cudf-polars requires CUDA 12.0+ to installed."
             ) from None
         else:
             raise

python/custreamz/README.md

@@ -48,8 +48,8 @@ Please see the [Demo Docker Repository](https://hub.docker.com/r/rapidsai/rapids
 
 ### CUDA/GPU requirements
 
-* CUDA 11.0+
-* NVIDIA driver 450.80.02+
+* CUDA 12.0+
+* NVIDIA driver 525.60.13+
 * Volta architecture or better (Compute Capability >=7.0)
 
 ### Conda