Merge pull request #167 from sterrettm2/taskwait

Fixes bug with nested OpenMP, fixed task threshold, extended tests range

Author: Raghuveer Devulapalli

.github/workflows/build-numpy.yml

@@ -14,7 +14,7 @@ jobs:
   np-multiarray-tgl:
 
     if: github.repository == 'intel/x86-simd-sort'
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - name: Checkout x86-simd-sort
@@ -80,7 +80,7 @@ jobs:
   np-multiarray-spr:
 
     if: github.repository == 'intel/x86-simd-sort'
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - name: Checkout x86-simd-sort

.github/workflows/c-cpp.yml

@@ -11,7 +11,7 @@ permissions: read-all
 jobs:
   SKL-gcc9:
 
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
@@ -41,7 +41,7 @@ jobs:
 
   SKX-gcc10:
 
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
@@ -71,7 +71,7 @@ jobs:
 
   TGL-gcc11:
 
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
@@ -100,7 +100,7 @@ jobs:
 
   SPR-gcc13:
 
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
@@ -137,7 +137,7 @@ jobs:
 
   SKX-SKL-openmp:
 
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
@@ -169,7 +169,7 @@ jobs:
 
   SPR-gcc13-special-cases:
 
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
@@ -204,7 +204,7 @@ jobs:
 
   manylinux-32bit:
 
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
@@ -216,7 +216,7 @@ jobs:
 
   SPR-icpx:
 
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
@@ -226,8 +226,7 @@ jobs:
         echo "deb https://apt.repos.intel.com/oneapi all main" | sudo tee /etc/apt/sources.list.d/oneAPI.list
         sudo add-apt-repository -y "deb https://apt.repos.intel.com/oneapi all main"
         sudo apt update --allow-insecure-repositories
-        sudo apt --allow-unauthenticated -y install intel-oneapi-compiler-dpcpp-cpp libgtest-dev curl git python3-pip
-        sudo pip3 install meson ninja
+        sudo apt --allow-unauthenticated -y install intel-oneapi-compiler-dpcpp-cpp libgtest-dev curl git python3-pip meson
 
     - name: Install Intel SDE
       run: |

.github/workflows/linting.yml

@@ -11,7 +11,7 @@ permissions: read-all
 jobs:
   clang-format:
 
-    runs-on: intel-ubuntu-latest
+    runs-on: intel-ubuntu-24.04
 
     steps:
     - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1

.github/workflows/scorecard.yml

@@ -22,7 +22,7 @@ jobs:
 
     name: Scorecard analysis
     if: github.repository == 'intel/x86-simd-sort'
-    runs-on: ubuntu-latest
+    runs-on: intel-ubuntu-24.04
     permissions:
       # Needed to upload the results to code-scanning dashboard.
       security-events: write

src/xss-common-keyvaluesort.hpp

@@ -559,9 +559,14 @@ X86_SIMD_SORT_INLINE void xss_qsort_kv(
                     keys, indexes, 0, arrsize - 1, maxiters, task_threshold);
         }
         else {
-            kvsort_<keytype, valtype>(
-                    keys, indexes, 0, arrsize - 1, maxiters, 0);
+            kvsort_<keytype, valtype>(keys,
+                                      indexes,
+                                      0,
+                                      arrsize - 1,
+                                      maxiters,
+                                      std::numeric_limits<arrsize_t>::max());
         }
+#pragma omp taskwait
 #else
         kvsort_<keytype, valtype>(keys, indexes, 0, arrsize - 1, maxiters, 0);
 #endif

tests/test-keyvalue.cpp

@@ -15,6 +15,10 @@ class simdkvsort : public ::testing::Test {
     simdkvsort()
     {
         std::iota(arrsize.begin(), arrsize.end(), 1);
+        arrsize.push_back(10'000);
+        arrsize.push_back(100'000);
+        arrsize.push_back(1'000'000);
+
         arrtype = {"random",
                    "constant",
                    "sorted",