P4 to Git Change 1228628 by wchau@wchau_WIN_OCL_HSA on 2016/01/18 17:02:40

SWDEV-79308 -  Resubmit of CL1228064 with restriction of mininum scratch buffer size of 64K if a scratch buffer is needed.
	  Reduce the total scratch buffer size by a factor of 4, which in effect reducing the max. scratch waves from 32 to 8, to avoid the required total scratch buffer size exceeds the available local memory.
	  Made sure the scratch buffer size is aligned with 64K boundary

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuresource.cpp#235 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp#156 edit

rocclr/runtime/device/gpu/gpuresource.cpp

@@ -1937,8 +1937,8 @@ ResourceCache::findCalResource(Resource::CalResourceDesc* desc)
     GslResourceReference* ref = NULL;
     size_t  size = getResourceSize(desc);
 
-    // Early exit if resource is too big
-    if (size >= cacheSizeLimit_ || desc->skipRsrcCache_) {
+    // Early exit if resource is too big or it is for scratch buffer
+    if (size >= cacheSizeLimit_ || desc->skipRsrcCache_ || desc->scratch_) {
         //! \note we may need to free the cache here to reduce memory pressure
         return ref;
     }

rocclr/runtime/device/gpu/gslbe/src/rt/GSLDevice.cpp

@@ -1217,7 +1217,7 @@ CALGSLDevice::calcScratchBufferSize(uint32 regNum) const
 {
     gslProgramTargetEnum target = GSL_COMPUTE_PROGRAM;
 
-    // Determine the scratch size we need to allocate.
+    // Determine the scratch size we need to allocate
     cmScratchSpaceNeededPerShaderStage scratchSpacePerShaderStage;
     memset(&scratchSpacePerShaderStage, 0, sizeof(scratchSpacePerShaderStage));
     uint32 scratchBufferSizes[gslProgramTarget_COUNT];
@@ -1231,6 +1231,21 @@ CALGSLDevice::calcScratchBufferSize(uint32 regNum) const
 
     m_cs->CalcAllScratchBufferSizes(enabledShadersFlag, scratchSpacePerShaderStage,
                                 scratchBufferSizes);
+
+    // SWDEV-79308:
+    //   Reduce the total scratch buffer size by a factor of 4, which in effect reducing the
+    //   max. scratch waves from 32 to 8. This will avoid the required total scratch buffer
+    //   size exceeds the available local memory. (Note: the scratch buffer size needs to
+    //   be 64K alignment)
+    if (scratchBufferSizes[target] > 0)
+    {
+        scratchBufferSizes[target] = (scratchBufferSizes[target] >> 2) & 0xFFFF0000;
+
+        if (scratchBufferSizes[target] == 0) {  // assign minimum scratch buffer size of 64K
+            scratchBufferSizes[target] = 0x10000;
+        }
+    }
+
     return scratchBufferSizes[target];
 }