diff --git a/projects/hipblaslt/tensilelite/include/Tensile/AMDGPU.hpp b/projects/hipblaslt/tensilelite/include/Tensile/AMDGPU.hpp
index b1f09712f98..8caec03f87c 100644
--- a/projects/hipblaslt/tensilelite/include/Tensile/AMDGPU.hpp
+++ b/projects/hipblaslt/tensilelite/include/Tensile/AMDGPU.hpp
@@ -228,6 +228,7 @@ namespace TensileLite
         int         computeUnitCount = 0;
         int         skDynamicGrid    = 6;
         int         skDynamicWGM     = 0;
+        int         fixedWGM         = std::numeric_limits<int>::max();
         int         skMaxCUs         = 0;
         int         skGridMultiplier = 1;
         int         skFixedGrid      = 0;
@@ -263,6 +264,13 @@ namespace TensileLite
             return value;
         }
 
+        const int getFixedWGM() const
+        {
+            static const char* envStr = std::getenv("TENSILE_FIXED_WGM");
+            static const int   value  = (envStr == NULL ? std::numeric_limits<int>::max() : std::atoi(envStr));
+            return value;
+        }
+
         const int getSKMaxCUs() const
         {
             static const char* envStr = std::getenv("TENSILE_STREAMK_MAX_CUS");
diff --git a/projects/hipblaslt/tensilelite/src/AMDGPU.cpp b/projects/hipblaslt/tensilelite/src/AMDGPU.cpp
index 799076bfb1a..e93e72c8b3c 100644
--- a/projects/hipblaslt/tensilelite/src/AMDGPU.cpp
+++ b/projects/hipblaslt/tensilelite/src/AMDGPU.cpp
@@ -53,6 +53,7 @@ namespace TensileLite
         , deviceName(name)
         , skDynamicGrid(getSKDynamicGrid())
         , skDynamicWGM(getSKDynamicWGM())
+        , fixedWGM(getFixedWGM())
         , skMaxCUs(getSKMaxCUs())
         , skGridMultiplier(getSKGridMultiplier())
         , skFixedGrid(getSKFixedGrid())
diff --git a/projects/hipblaslt/tensilelite/src/ContractionSolution.cpp b/projects/hipblaslt/tensilelite/src/ContractionSolution.cpp
index 3b8a678adf6..680559f03ad 100644
--- a/projects/hipblaslt/tensilelite/src/ContractionSolution.cpp
+++ b/projects/hipblaslt/tensilelite/src/ContractionSolution.cpp
@@ -1204,7 +1204,11 @@ namespace TensileLite
             if(sizeMapping.streamK != 0)
             {
                 AMDGPU const* pAMDGPU = dynamic_cast<AMDGPU const*>(&hardware);
-                if(pAMDGPU->skDynamicWGM == 1)
+                if(pAMDGPU->fixedWGM >= -1024 && pAMDGPU->fixedWGM <= 1024)
+                {
+                    defaultWGM = pAMDGPU->fixedWGM;
+                }
+                else if(pAMDGPU->skDynamicWGM == 1)
                 {
                     hip::HipAMDGPU const* hipAMDGPU
                         = dynamic_cast<hip::HipAMDGPU const*>(&hardware);