[WIP] cuDF integration into XGBoost

CMakeLists.txt

@@ -9,6 +9,8 @@ msvc_use_static_runtime()
 
 # Options
 option(USE_CUDA  "Build with GPU acceleration")
+option(USE_CUDF "Build with cuDF support. Requires USE_CUDA." OFF)
+option(USE_NCCL "Build using NCCL for multi-GPU. Also requires USE_CUDA")
 option(JVM_BINDINGS "Build JVM bindings" OFF)
 option(GOOGLE_TEST "Build google tests" OFF)
 option(R_LIB "Build shared library for R package" OFF)
@@ -140,6 +142,12 @@ if(USE_CUDA)
     add_definitions(-DXGBOOST_USE_NCCL)
   endif()
 
+  if(USE_CUDF)
+    find_package(CUDF REQUIRED)
+    include_directories(${CUDF_INCLUDE_DIR})
+    add_definitions(-DXGBOOST_USE_CUDF)
+  endif()
+
   set(GENCODE_FLAGS "")
   format_gencode_flags("${GPU_COMPUTE_VER}" GENCODE_FLAGS)
   message("cuda architecture flags: ${GENCODE_FLAGS}")

cmake/modules/FindCUDF.cmake

@@ -0,0 +1,31 @@
+# Tries to find cuDF headers and libraries.
+#
+# Usage of this module as follows:
+#
+#  find_package(CUDF)
+#
+# Variables used by this module, they can change the default behaviour and need
+# to be set before calling find_package:
+#
+#  CUDF_ROOT - When set, this path is inspected instead of standard library
+#              locations as the root of the CUDF installation.
+#              The environment variable CUDF_ROOT overrides this variable.
+#
+# This module defines
+#  CUDF_FOUND, whether cuDF has been found
+#  CUDF_INCLUDE_DIR, directory containing header
+#
+# This module assumes that the user has already called find_package(CUDA)
+
+
+find_path(CUDF_INCLUDE_DIR
+  NAMES cudf.h cudf/types.h
+  PATHS $ENV{CUDF_ROOT}/include ${CUDF_ROOT}/include ${CUDA_INCLUDE_DIRS} /usr/include)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(CUDF DEFAULT_MSG
+                                  CUDF_INCLUDE_DIR)
+
+mark_as_advanced(
+  CUDF_INCLUDE_DIR
+)

cmake/modules/FindGDF.cmake

@@ -0,0 +1,45 @@
+# Tries to find GDF headers and libraries.
+#
+# Usage of this module as follows:
+#
+#  find_package(GDF)
+#
+# Variables used by this module, they can change the default behaviour and need
+# to be set before calling find_package:
+#
+#  GDF_ROOT - When set, this path is inspected instead of standard library
+#              locations as the root of the GDF installation.
+#              The environment variable GDF_ROOT overrides this variable.
+#
+# This module defines
+#  GDF_FOUND, whether nccl has been found
+#  GDF_INCLUDE_DIR, directory containing header
+#  GDF_LIBRARY, directory containing nccl library
+#  GDF_LIB_NAME, nccl library name
+#
+# This module assumes that the user has already called find_package(CUDA)
+
+
+set(GDF_LIB_NAME cudf)
+
+find_path(GDF_INCLUDE_DIR
+  NAMES cudf.h
+  PATHS $ENV{GDF_ROOT}/include ${GDF_ROOT}/include ${CUDA_INCLUDE_DIRS} /usr/include)
+
+find_library(GDF_LIBRARY
+  NAMES ${GDF_LIB_NAME}
+  PATHS $ENV{GDF_ROOT}/lib ${GDF_ROOT}/lib ${CUDA_INCLUDE_DIRS}/../lib /usr/lib)
+
+if (GDF_INCLUDE_DIR AND GDF_LIBRARY)
+  get_filename_component(GDF_LIBRARY ${GDF_LIBRARY} PATH)
+endif ()
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(GDF DEFAULT_MSG
+                                  GDF_INCLUDE_DIR GDF_LIBRARY)
+
+mark_as_advanced(
+  GDF_INCLUDE_DIR
+  GDF_LIBRARY
+  GDF_LIB_NAME
+)

include/xgboost/c_api.h

@@ -16,6 +16,10 @@
 #include <stdint.h>
 #endif
 
+#ifdef XGBOOST_USE_CUDF
+#include <cudf/types.h>
+#endif
+
 // XGBoost C API will include APIs in Rabit C API
 #include <rabit/c_api.h>
 
@@ -28,7 +32,6 @@
 // manually define unsigned long
 typedef uint64_t bst_ulong;  // NOLINT(*)
 
-
 /*! \brief handle to DMatrix */
 typedef void *DMatrixHandle;  // NOLINT(*)
 /*! \brief handle to Booster */
@@ -105,6 +108,20 @@ XGB_DLL const char *XGBGetLastError(void);
  */
 XGB_DLL int XGBRegisterLogCallback(void (*callback)(const char*));
 
+#ifdef XGBOOST_USE_CUDF
+
+/*!
+ * \bried create a data matrix from a CUDA data frame (CUDF)
+ * \param cols array of CUDF columns
+ * \param n_cols number of CUDF columns
+ * \param[out] out handle for the DMatrix built
+ * \return 0 when success, -1 when failure happens
+ */
+XGB_DLL int XGDMatrixCreateFromCUDF
+(gdf_column **cols, size_t n_cols, DMatrixHandle *out);
+
+#endif
+
 /*!
  * \brief load a data matrix
  * \param fname the name of the file
@@ -282,6 +299,25 @@ XGB_DLL int XGDMatrixSetFloatInfo(DMatrixHandle handle,
                                   const char *field,
                                   const float *array,
                                   bst_ulong len);
+
+#ifdef XGBOOST_USE_CUDF
+
+/*!
+ * \brief set a vector to 
+ * \param handle a instance of data matrix
+ * \param field field name, can be label, weight
+ * \param array pointer to float vector
+ * \param len length of array
+ * \return 0 when success, -1 when failure happens
+ */
+
+XGB_DLL int XGDMatrixSetInfoCUDF(DMatrixHandle handle,
+                                const char *field,
+                                gdf_column** gdf,
+                                size_t n_cols);
+
+#endif
+
 /*!
  * \brief set uint32 vector to a content in info
  * \param handle a instance of data matrix

include/xgboost/data.h

@@ -9,6 +9,9 @@
 
 #include <dmlc/base.h>
 #include <dmlc/data.h>
+#ifdef XGBOOST_USE_CUDF
+#include <cudf/types.h>
+#endif
 #include <cstring>
 #include <memory>
 #include <numeric>
@@ -21,6 +24,10 @@
 
 #include "../../src/common/host_device_vector.h"
 
+#include "../../src/common/host_device_vector.h"
+
+#include "../../src/common/host_device_vector.h"
+
 namespace xgboost {
 // forward declare learner.
 class LearnerImpl;
@@ -121,6 +128,16 @@ class MetaInfo {
    * \param num Number of elements in the source array.
    */
   void SetInfo(const char* key, const void* dptr, DataType dtype, size_t num);
+
+#ifdef XGBOOST_USE_CUDF
+  /*!
+   * \brief Set information in the meta info from CUDF columns.
+   * \param key The key of the information.
+   * \param cols The CUDF columns used to set the info.
+   * \param n_cols The number of CUDF columns.
+   */
+  void SetInfoCUDF(const char* key, gdf_column** cols, size_t n_cols);
+#endif
 
  private:
   /*! \brief argsort of labels */

python-package/xgboost/core.py

@@ -15,7 +15,10 @@
 import re
 import sys
 
+import cudf
+from libgdf_cffi import ffi
 import numpy as np
+import cudf.dataframe as gdf
 import scipy.sparse
 
 from .compat import (STRING_TYPES, PY3, DataFrame, MultiIndex, py_str,
@@ -227,7 +230,7 @@ def c_array(ctype, values):
 def _maybe_pandas_data(data, feature_names, feature_types):
     """ Extract internal data from pd.DataFrame for DMatrix data """
 
-    if not isinstance(data, DataFrame):
+    if not isinstance(data, (DataFrame, cudf.DataFrame)):
         return data, feature_names, feature_types
 
     data_dtypes = data.dtypes
@@ -240,18 +243,24 @@ def _maybe_pandas_data(data, feature_names, feature_types):
         raise ValueError(msg + ', '.join(bad_fields))
 
     if feature_names is None:
-        if isinstance(data.columns, MultiIndex):
+        data_columns = data.columns
+        if isinstance(data_columns, MultiIndex):
             feature_names = [
                 ' '.join(map(str, i))
-                for i in data.columns
+                for i in data_columns
             ]
+        elif (isinstance(data_columns, (tuple, list)) and len(data_columns) > 0 and
+              isinstance(data_columns[0], str)):
+            feature_names = list(data_columns)
         else:
-            feature_names = data.columns.format()
+            feature_names = data_columns.format()
 
     if feature_types is None:
         feature_types = [PANDAS_DTYPE_MAPPER[dtype.name] for dtype in data_dtypes]
 
-    data = data.values.astype('float')
+    # only convert pandas.DataFrame, CUDF conversion happens elsewhere
+    if isinstance(data, DataFrame):
+        data = data.values.astype('float')
 
     return data, feature_names, feature_types
 
@@ -386,6 +395,8 @@ def __init__(self, data, label=None, missing=None,
             _check_call(_LIB.XGDMatrixCreateFromFile(c_str(data),
                                                      ctypes.c_int(silent),
                                                      ctypes.byref(self.handle)))
+        elif isinstance(data, cudf.DataFrame):
+            self._init_from_cudf(data)
         elif isinstance(data, scipy.sparse.csr_matrix):
             self._init_from_csr(data)
         elif isinstance(data, scipy.sparse.csc_matrix):
@@ -405,17 +416,37 @@ def __init__(self, data, label=None, missing=None,
         if label is not None:
             if isinstance(label, np.ndarray):
                 self.set_label_npy2d(label)
+            elif isinstance(label, cudf.dataframe.column.Column):
+                self.set_info_cudf('label', label)
+            elif isinstance(label, cudf.DataFrame):
+                self.set_info_cudf('label', label)
             else:
                 self.set_label(label)
         if weight is not None:
             if isinstance(weight, np.ndarray):
                 self.set_weight_npy2d(weight)
+            elif isinstance(weight, cudf.dataframe.column.Column):
+                self.set_info_cudf('weight', weight)
+            elif isinstance(weight, cudf.DataFrame):
+                self.set_info_cudf('weight', weight)
             else:
                 self.set_weight(weight)
 
         self.feature_names = feature_names
         self.feature_types = feature_types
 
+    def _init_from_cudf(self, df):
+        """
+        Initialize data from a GPU data frame.
+        """
+        self.handle = ctypes.c_void_p()
+        col_ptrs = [df[col]._column.cffi_view for col in df.columns]
+        col_ptr_arr = ffi.new('gdf_column*[]', col_ptrs)
+        _check_call(_LIB.XGDMatrixCreateFromCUDF
+                    (ctypes.c_void_p(int(ffi.cast('uintptr_t', col_ptr_arr))),
+                     ctypes.c_size_t(len(df.columns)),
+                     ctypes.byref(self.handle)))
+
     def _init_from_csr(self, csr):
         """
         Initialize data from a CSR matrix.
@@ -592,6 +623,19 @@ def set_float_info_npy2d(self, field, data):
                                                c_data,
                                                c_bst_ulong(len(data))))
 
+    def set_info_cudf(self, field, data):
+        col_ptrs = []
+        if isinstance(data, cudf.DataFrame):
+            col_ptrs = [data[col]._column.cffi_view for col in data.columns]
+        else:
+            # data is a single CUDF column
+            col_ptrs = [data.cffi_view]
+        col_ptr_arr = ffi.new('gdf_column*[]', col_ptrs)
+        _check_call(_LIB.XGDMatrixSetInfoCUDF
+                    (self.handle, c_str(field),
+                     ctypes.c_void_p(int(ffi.cast('uintptr_t', col_ptr_arr))),
+                     ctypes.c_size_t(len(col_ptrs))))
+
     def set_uint_info(self, field, data):
         """Set uint type property into the DMatrix.
 

src/c_api/c_api.cc

@@ -260,6 +260,19 @@ int XGDMatrixCreateFromDataIter(
   API_END();
 }
 
+#ifdef XGBOOST_USE_CUDF
+
+int XGDMatrixCreateFromCUDF
+(gdf_column **cols, size_t n_cols, DMatrixHandle *out) {
+  API_BEGIN();
+  std::unique_ptr<data::SimpleCSRSource> source(new data::SimpleCSRSource());
+  source->InitFromCUDF(cols, n_cols);
+  *out = new std::shared_ptr<DMatrix>(DMatrix::Create(std::move(source)));
+  API_END();
+}
+
+#endif
+
 XGB_DLL int XGDMatrixCreateFromCSREx(const size_t* indptr,
                                      const unsigned* indices,
                                      const bst_float* data,
@@ -778,6 +791,21 @@ XGB_DLL int XGDMatrixSetFloatInfo(DMatrixHandle handle,
   API_END();
 }
 
+#ifdef XGBOOST_USE_CUDF
+
+XGB_DLL int XGDMatrixSetInfoCUDF(DMatrixHandle handle,
+                                const char *field,
+                                gdf_column **cols,
+                                size_t n_cols) {
+  API_BEGIN();
+  CHECK_HANDLE();
+  static_cast<std::shared_ptr<DMatrix>*>(handle)
+    ->get()->Info().SetInfoCUDF(field, cols, n_cols);
+  API_END();
+}
+
+#endif
+
 XGB_DLL int XGDMatrixSetUIntInfo(DMatrixHandle handle,
                          const char* field,
                          const unsigned* info,