Fix MXNet R package build (#13952)

* fix mxnet r package build

* add ci

* remove mkldnn-gpu test for R

* add minimal test for MKLDNN-R

* pick mlp as minimal R test

Makefile

@@ -599,6 +599,13 @@ rpkg:
 	mkdir -p R-package/inst/libs
 	cp src/io/image_recordio.h R-package/src
 	cp -rf lib/libmxnet.so R-package/inst/libs
+
+	if [ -e "lib/libmkldnn.so.0" ]; then \
+		cp -rf lib/libmkldnn.so.0 R-package/inst/libs; \
+		cp -rf lib/libiomp5.so R-package/inst/libs; \
+		cp -rf lib/libmklml_intel.so R-package/inst/libs; \
+	fi
+
 	mkdir -p R-package/inst/include
 	cp -rf include/* R-package/inst/include
 	rm R-package/inst/include/dmlc

ci/docker/runtime_functions.sh

@@ -902,6 +902,36 @@ unittest_ubuntu_cpu_R() {
     make rpkgtest R_LIBS=/tmp/r-site-library
 }
 
+unittest_ubuntu_minimal_R() {
+    set -ex
+    mkdir -p /tmp/r-site-library
+    # build R packages in parallel
+    mkdir -p ~/.R/
+    build_ccache_wrappers
+    echo  "MAKEFLAGS = -j"$(nproc) > ~/.R/Makevars
+    # make -j not supported
+    make rpkg                           \
+        USE_BLAS=openblas               \
+        R_LIBS=/tmp/r-site-library
+
+    R CMD INSTALL --library=/tmp/r-site-library R-package
+    # pick mlp as minimal R test
+    R_LIBS=/tmp/r-site-library \
+        Rscript -e "library(mxnet); require(mlbench); \
+                    data(Sonar, package=\"mlbench\"); \
+                    Sonar[,61] = as.numeric(Sonar[,61])-1; \
+                    train.ind = c(1:50, 100:150); \
+                    train.x = data.matrix(Sonar[train.ind, 1:60]); \
+                    train.y = Sonar[train.ind, 61]; \
+                    test.x = data.matrix(Sonar[-train.ind, 1:60]); \
+                    test.y = Sonar[-train.ind, 61]; \
+                    model = mx.mlp(train.x, train.y, hidden_node = 10, \
+                                   out_node = 2, out_activation = \"softmax\", \
+                                   learning.rate = 0.1, \
+                                   array.layout = \"rowmajor\"); \
+                    preds = predict(model, test.x, array.layout = \"rowmajor\")"
+}
+
 unittest_ubuntu_gpu_R() {
     set -ex
     mkdir -p /tmp/r-site-library

ci/jenkins/Jenkins_steps.groovy

@@ -927,6 +927,20 @@ def test_unix_r_cpu() {
     }]
 }
 
+def test_unix_r_mkldnn_cpu() {
+    return ['R: MKLDNN-CPU': {
+      node(NODE_LINUX_CPU) {
+        ws('workspace/ut-r-mkldnn-cpu') {
+          timeout(time: max_time, unit: 'MINUTES') {
+            utils.unpack_and_init('mkldnn_cpu', mx_mkldnn_lib, true)
+            utils.docker_run('ubuntu_cpu', 'unittest_ubuntu_minimal_R', false)
+            utils.publish_test_coverage()
+          }
+        }
+      }
+    }]
+}
+
 def test_unix_perl_cpu() {
     return ['Perl: CPU': {
       node(NODE_LINUX_CPU) {

ci/jenkins/Jenkinsfile_unix_cpu

@@ -55,6 +55,7 @@ core_logic: {
     custom_steps.test_unix_clojure_integration_cpu(),
     custom_steps.test_unix_perl_cpu(),
     custom_steps.test_unix_r_cpu(),
+    custom_steps.test_unix_r_mkldnn_cpu(),
     custom_steps.test_unix_julia07_cpu(),
     custom_steps.test_unix_julia10_cpu(),
     custom_steps.test_unix_onnx_cpu(),