[Openmp-commits] [openmp] 6563780 - [Libomptarget] Fix block and thread limit environment variables not being respected

[Joseph Huber via Openmp-commits]

Author: Joseph Huber
Date: 2023-03-01T14:12:46-06:00
New Revision: 656378085e7f52d67581245617474c5ce530dac4

URL: https://github.com/llvm/llvm-project/commit/656378085e7f52d67581245617474c5ce530dac4
DIFF: https://github.com/llvm/llvm-project/commit/656378085e7f52d67581245617474c5ce530dac4.diff

LOG: [Libomptarget] Fix block and thread limit environment variables not being respected

The next-gen plugins did not properly set the values from
`OMP_NUM_TEAMS` and `OMP_TEAMS_THREAD_LIMIT`. This is because these
maximum values are set by each plugin to its hardware maximum. This
happens *after* the previous initialization. Move it to the correct
place and then add a test.

Fixes https://github.com/llvm/llvm-project/issues/61082

Reviewed By: tianshilei1992

Differential Revision: https://reviews.llvm.org/D145105

Added: 
    openmp/libomptarget/test/api/omp_env_vars.c

Modified: 
    openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp

Removed: 
    


################################################################################
diff  --git a/openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp b/openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp
index c8c4a60d0bdf0..1d4d906bec4e8 100644
--- a/openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp
+++ b/openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp
@@ -352,15 +352,7 @@ GenericDeviceTy::GenericDeviceTy(int32_t DeviceId, int32_t NumDevices,
       OMPX_InitialNumEvents("LIBOMPTARGET_NUM_INITIAL_EVENTS", 32),
       DeviceId(DeviceId), GridValues(OMPGridValues),
       PeerAccesses(NumDevices, PeerAccessState::PENDING), PeerAccessesLock(),
-      PinnedAllocs(*this) {
-  if (OMP_NumTeams > 0)
-    GridValues.GV_Max_Teams =
-        std::min(GridValues.GV_Max_Teams, uint32_t(OMP_NumTeams));
-
-  if (OMP_TeamsThreadLimit > 0)
-    GridValues.GV_Max_WG_Size =
-        std::min(GridValues.GV_Max_WG_Size, uint32_t(OMP_TeamsThreadLimit));
-}
+      PinnedAllocs(*this) {}
 
 Error GenericDeviceTy::init(GenericPluginTy &Plugin) {
   if (auto Err = initImpl(Plugin))
@@ -385,6 +377,16 @@ Error GenericDeviceTy::init(GenericPluginTy &Plugin) {
     return HeapSizeEnvarOrErr.takeError();
   OMPX_TargetHeapSize = std::move(*HeapSizeEnvarOrErr);
 
+  // Update the maximum number of teams and threads after the device
+  // initialization sets the corresponding hardware limit.
+  if (OMP_NumTeams > 0)
+    GridValues.GV_Max_Teams =
+        std::min(GridValues.GV_Max_Teams, uint32_t(OMP_NumTeams));
+
+  if (OMP_TeamsThreadLimit > 0)
+    GridValues.GV_Max_WG_Size =
+        std::min(GridValues.GV_Max_WG_Size, uint32_t(OMP_TeamsThreadLimit));
+
   // Enable the memory manager if required.
   auto [ThresholdMM, EnableMM] = MemoryManagerTy::getSizeThresholdFromEnv();
   if (EnableMM)
@@ -1191,7 +1193,6 @@ __tgt_target_table *__tgt_rtl_load_binary(int32_t DeviceId,
   GenericPluginTy &Plugin = Plugin::get();
   GenericDeviceTy &Device = Plugin.getDevice(DeviceId);
 
-
   auto TableOrErr = Device.loadBinary(Plugin, TgtImage);
   if (!TableOrErr) {
     auto Err = TableOrErr.takeError();

diff  --git a/openmp/libomptarget/test/api/omp_env_vars.c b/openmp/libomptarget/test/api/omp_env_vars.c
new file mode 100644
index 0000000000000..2e78bb115beed
--- /dev/null
+++ b/openmp/libomptarget/test/api/omp_env_vars.c
@@ -0,0 +1,12 @@
+// RUN: %libomptarget-compile-generic
+// RUN: env OMP_NUM_TEAMS=1 OMP_TEAMS_THREAD_LIMIT=1 LIBOMPTARGET_INFO=16 \
+// RUN:   %libomptarget-run-generic 2>&1 | %fcheck-generic
+
+#define N 256
+
+int main() {
+  // CHECK: Launching kernel [[KERNEL:.+_main_.+]] with 1 blocks and 1 threads
+#pragma omp target teams
+#pragma omp parallel
+  {}
+}