diff --git a/shared/source/direct_submission/direct_submission_hw.h b/shared/source/direct_submission/direct_submission_hw.h
index 44d31a8f07..ec337519bb 100644
--- a/shared/source/direct_submission/direct_submission_hw.h
+++ b/shared/source/direct_submission/direct_submission_hw.h
@@ -48,7 +48,7 @@ enum class DirectSubmissionSfenceMode : int32_t {
 
 namespace UllsDefaults {
 inline constexpr bool defaultDisableCacheFlush = true;
-inline constexpr bool defaultDisableMonitorFence = false;
+inline constexpr bool defaultDisableMonitorFence = true;
 } // namespace UllsDefaults
 
 struct BatchBuffer;
@@ -145,8 +145,8 @@ class DirectSubmissionHw {
 
     void setReturnAddress(void *returnCmd, uint64_t returnAddress);
 
-    void *dispatchWorkloadSection(BatchBuffer &batchBuffer);
-    size_t getSizeDispatch(bool relaxedOrderingSchedulerRequired, bool returnPtrsRequired);
+    void *dispatchWorkloadSection(BatchBuffer &batchBuffer, bool dispatchMonitorFence);
+    size_t getSizeDispatch(bool relaxedOrderingSchedulerRequired, bool returnPtrsRequired, bool dispatchMonitorFence);
 
     void dispatchPrefetchMitigation();
     size_t getSizePrefetchMitigation();
diff --git a/shared/source/direct_submission/direct_submission_hw.inl b/shared/source/direct_submission/direct_submission_hw.inl
index ce44202079..17624f63f1 100644
--- a/shared/source/direct_submission/direct_submission_hw.inl
+++ b/shared/source/direct_submission/direct_submission_hw.inl
@@ -51,6 +51,9 @@ DirectSubmissionHw<GfxFamily, Dispatcher>::DirectSubmissionHw(const DirectSubmis
 
     disableCacheFlush = UllsDefaults::defaultDisableCacheFlush;
     disableMonitorFence = UllsDefaults::defaultDisableMonitorFence;
+    if (DebugManager.flags.DirectSubmissionDisableMonitorFence.get() != -1) {
+        this->disableMonitorFence = DebugManager.flags.DirectSubmissionDisableMonitorFence.get();
+    }
 
     if (DebugManager.flags.DirectSubmissionMaxRingBuffers.get() != -1) {
         this->maxRingBufferCount = DebugManager.flags.DirectSubmissionMaxRingBuffers.get();
@@ -505,7 +508,7 @@ bool DirectSubmissionHw<GfxFamily, Dispatcher>::startRingBuffer() {
         startSize += RelaxedOrderingHelper::getSizeRegistersInit<GfxFamily>();
     }
 
-    size_t requiredSize = startSize + getSizeDispatch(false, false) + getSizeEnd(false);
+    size_t requiredSize = startSize + getSizeDispatch(false, false, dispatchMonitorFenceRequired(true)) + getSizeEnd(false);
     if (ringCommandStream.getAvailableSpace() < requiredSize) {
         switchRingBuffers();
     }
@@ -662,7 +665,7 @@ inline size_t DirectSubmissionHw<GfxFamily, Dispatcher>::getSizeEnd(bool relaxed
 }
 
 template <typename GfxFamily, typename Dispatcher>
-inline size_t DirectSubmissionHw<GfxFamily, Dispatcher>::getSizeDispatch(bool relaxedOrderingSchedulerRequired, bool returnPtrsRequired) {
+inline size_t DirectSubmissionHw<GfxFamily, Dispatcher>::getSizeDispatch(bool relaxedOrderingSchedulerRequired, bool returnPtrsRequired, bool dispatchMonitorFence) {
     size_t size = getSizeSemaphoreSection(relaxedOrderingSchedulerRequired);
     if (workloadMode == 0) {
         size += getSizeStartSection();
@@ -677,7 +680,7 @@ inline size_t DirectSubmissionHw<GfxFamily, Dispatcher>::getSizeDispatch(bool re
     if (!disableCacheFlush) {
         size += Dispatcher::getSizeCacheFlush(rootDeviceEnvironment);
     }
-    if (!disableMonitorFence) {
+    if (dispatchMonitorFence) {
         size += Dispatcher::getSizeMonitorFence(rootDeviceEnvironment);
     }
 
@@ -695,7 +698,7 @@ void DirectSubmissionHw<GfxFamily, Dispatcher>::updateRelaxedOrderingQueueSize(u
 }
 
 template <typename GfxFamily, typename Dispatcher>
-void *DirectSubmissionHw<GfxFamily, Dispatcher>::dispatchWorkloadSection(BatchBuffer &batchBuffer) {
+void *DirectSubmissionHw<GfxFamily, Dispatcher>::dispatchWorkloadSection(BatchBuffer &batchBuffer, bool dispatchMonitorFence) {
     void *currentPosition = ringCommandStream.getSpace(0);
 
     if (DebugManager.flags.DirectSubmissionPrintBuffers.get()) {
@@ -765,7 +768,7 @@ void *DirectSubmissionHw<GfxFamily, Dispatcher>::dispatchWorkloadSection(BatchBu
         Dispatcher::dispatchCacheFlush(ringCommandStream, this->rootDeviceEnvironment, gpuVaForMiFlush);
     }
 
-    if (!disableMonitorFence || this->dispatchMonitorFenceRequired(batchBuffer.hasStallingCmds)) {
+    if (dispatchMonitorFence) {
         TagData currentTagData = {};
         getTagAddressValue(currentTagData);
         Dispatcher::dispatchMonitorFence(ringCommandStream, currentTagData.tagAddress, currentTagData.tagValue, this->rootDeviceEnvironment,
@@ -935,8 +938,9 @@ bool DirectSubmissionHw<GfxFamily, Dispatcher>::dispatchCommandBuffer(BatchBuffe
     this->startRingBuffer();
 
     bool relaxedOrderingSchedulerWillBeNeeded = (this->relaxedOrderingSchedulerRequired || batchBuffer.hasRelaxedOrderingDependencies);
+    bool dispatchMonitorFence = this->dispatchMonitorFenceRequired(batchBuffer.hasStallingCmds);
 
-    size_t dispatchSize = getSizeDispatch(relaxedOrderingSchedulerWillBeNeeded, batchBuffer.hasRelaxedOrderingDependencies);
+    size_t dispatchSize = getSizeDispatch(relaxedOrderingSchedulerWillBeNeeded, batchBuffer.hasRelaxedOrderingDependencies, dispatchMonitorFence);
 
     if (this->copyCommandBufferIntoRing(batchBuffer)) {
         dispatchSize += (batchBuffer.stream->getUsed() - batchBuffer.startOffset) - 2 * getSizeStartSection();
@@ -965,7 +969,7 @@ bool DirectSubmissionHw<GfxFamily, Dispatcher>::dispatchCommandBuffer(BatchBuffe
 
     handleNewResourcesSubmission();
 
-    void *currentPosition = dispatchWorkloadSection(batchBuffer);
+    void *currentPosition = dispatchWorkloadSection(batchBuffer, dispatchMonitorFence);
 
     cpuCachelineFlush(currentPosition, dispatchSize);
     handleResidency();
@@ -1084,7 +1088,7 @@ inline GraphicsAllocation *DirectSubmissionHw<GfxFamily, Dispatcher>::switchRing
 
 template <typename GfxFamily, typename Dispatcher>
 bool DirectSubmissionHw<GfxFamily, Dispatcher>::dispatchMonitorFenceRequired(bool hasStallingCmds) {
-    return false;
+    return !this->disableMonitorFence;
 }
 
 template <typename GfxFamily, typename Dispatcher>
diff --git a/shared/source/direct_submission/linux/drm_direct_submission.inl b/shared/source/direct_submission/linux/drm_direct_submission.inl
index a73360e548..af7895136d 100644
--- a/shared/source/direct_submission/linux/drm_direct_submission.inl
+++ b/shared/source/direct_submission/linux/drm_direct_submission.inl
@@ -28,12 +28,6 @@ template <typename GfxFamily, typename Dispatcher>
 DrmDirectSubmission<GfxFamily, Dispatcher>::DrmDirectSubmission(const DirectSubmissionInputParams &inputParams)
     : DirectSubmissionHw<GfxFamily, Dispatcher>(inputParams) {
 
-    this->disableMonitorFence = true;
-
-    if (DebugManager.flags.DirectSubmissionDisableMonitorFence.get() != -1) {
-        this->disableMonitorFence = DebugManager.flags.DirectSubmissionDisableMonitorFence.get();
-    }
-
     if (DebugManager.flags.OverrideUserFenceStartValue.get() != -1) {
         this->completionFenceValue = static_cast<decltype(completionFenceValue)>(DebugManager.flags.OverrideUserFenceStartValue.get());
     }
diff --git a/shared/source/direct_submission/windows/wddm_direct_submission.inl b/shared/source/direct_submission/windows/wddm_direct_submission.inl
index e7ec61ca64..3c0a336ef2 100644
--- a/shared/source/direct_submission/windows/wddm_direct_submission.inl
+++ b/shared/source/direct_submission/windows/wddm_direct_submission.inl
@@ -41,12 +41,6 @@ WddmDirectSubmission<GfxFamily, Dispatcher>::WddmDirectSubmission(const DirectSu
     if (this->miMemFenceRequired) {
         this->gpuVaForAdditionalSynchronizationWA = this->completionFenceAllocation->getGpuAddress() + 8u;
     }
-
-    this->disableMonitorFence = true;
-
-    if (DebugManager.flags.DirectSubmissionDisableMonitorFence.get() != -1) {
-        this->disableMonitorFence = DebugManager.flags.DirectSubmissionDisableMonitorFence.get();
-    }
 }
 
 template <typename GfxFamily, typename Dispatcher>
@@ -157,7 +151,7 @@ uint64_t WddmDirectSubmission<GfxFamily, Dispatcher>::updateTagValue(bool hasSta
 
 template <typename GfxFamily, typename Dispatcher>
 bool WddmDirectSubmission<GfxFamily, Dispatcher>::dispatchMonitorFenceRequired(bool hasStallingCmds) {
-    return hasStallingCmds;
+    return !this->disableMonitorFence || hasStallingCmds;
 }
 
 template <typename GfxFamily, typename Dispatcher>
diff --git a/shared/test/common/mocks/mock_direct_submission_hw.h b/shared/test/common/mocks/mock_direct_submission_hw.h
index 7b4e2b4be7..d225268e41 100644
--- a/shared/test/common/mocks/mock_direct_submission_hw.h
+++ b/shared/test/common/mocks/mock_direct_submission_hw.h
@@ -32,6 +32,7 @@ struct MockDirectSubmissionHw : public DirectSubmissionHw<GfxFamily, Dispatcher>
     using BaseClass::disableCpuCacheFlush;
     using BaseClass::disableMonitorFence;
     using BaseClass::dispatchDisablePrefetcher;
+    using BaseClass::dispatchMonitorFenceRequired;
     using BaseClass::dispatchPartitionRegisterConfiguration;
     using BaseClass::dispatchPrefetchMitigation;
     using BaseClass::dispatchRelaxedOrderingReturnPtrRegs;
diff --git a/shared/test/unit_test/direct_submission/direct_submission_tests_1.cpp b/shared/test/unit_test/direct_submission/direct_submission_tests_1.cpp
index 46cdf0abbf..ffb8d3b1a9 100644
--- a/shared/test/unit_test/direct_submission/direct_submission_tests_1.cpp
+++ b/shared/test/unit_test/direct_submission/direct_submission_tests_1.cpp
@@ -452,12 +452,12 @@ HWTEST_F(DirectSubmissionTest,
     using Dispatcher = RenderDispatcher<FamilyType>;
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> regularDirectSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
+    regularDirectSubmission.disableMonitorFence = false;
     size_t regularSizeEnd = regularDirectSubmission.getSizeEnd(false);
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
 
     bool ret = directSubmission.allocateResources();
-    directSubmission.disableMonitorFence = true;
     directSubmission.ringStart = true;
 
     EXPECT_TRUE(ret);
@@ -567,10 +567,9 @@ HWTEST_F(DirectSubmissionTest, givenDirectSubmissionWhenGetDispatchSizeThenExpec
 
     size_t expectedSize = directSubmission.getSizeStartSection() +
                           Dispatcher::getSizeCacheFlush(directSubmission.rootDeviceEnvironment) +
-                          Dispatcher::getSizeMonitorFence(directSubmission.rootDeviceEnvironment) +
                           directSubmission.getSizeSemaphoreSection(false) + directSubmission.getSizeNewResourceHandler();
 
-    size_t actualSize = directSubmission.getSizeDispatch(false, false);
+    size_t actualSize = directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(true));
     EXPECT_EQ(expectedSize, actualSize);
 }
 
@@ -584,9 +583,8 @@ HWTEST_F(DirectSubmissionTest,
 
     size_t expectedSize = Dispatcher::getSizeStoreDwordCommand() +
                           Dispatcher::getSizeCacheFlush(directSubmission.rootDeviceEnvironment) +
-                          Dispatcher::getSizeMonitorFence(directSubmission.rootDeviceEnvironment) +
                           directSubmission.getSizeSemaphoreSection(false) + directSubmission.getSizeNewResourceHandler();
-    size_t actualSize = directSubmission.getSizeDispatch(false, false);
+    size_t actualSize = directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false));
     EXPECT_EQ(expectedSize, actualSize);
 }
 
@@ -599,9 +597,8 @@ HWTEST_F(DirectSubmissionTest,
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
     directSubmission.workloadMode = 2;
     size_t expectedSize = Dispatcher::getSizeCacheFlush(directSubmission.rootDeviceEnvironment) +
-                          Dispatcher::getSizeMonitorFence(directSubmission.rootDeviceEnvironment) +
                           directSubmission.getSizeSemaphoreSection(false) + directSubmission.getSizeNewResourceHandler();
-    size_t actualSize = directSubmission.getSizeDispatch(false, false);
+    size_t actualSize = directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false));
     EXPECT_EQ(expectedSize, actualSize);
 }
 
@@ -612,10 +609,9 @@ HWTEST_F(DirectSubmissionTest,
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
     directSubmission.disableCacheFlush = true;
     size_t expectedSize = directSubmission.getSizeStartSection() +
-                          Dispatcher::getSizeMonitorFence(directSubmission.rootDeviceEnvironment) +
                           directSubmission.getSizeSemaphoreSection(false) + directSubmission.getSizeNewResourceHandler();
 
-    size_t actualSize = directSubmission.getSizeDispatch(false, false);
+    size_t actualSize = directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false));
     EXPECT_EQ(expectedSize, actualSize);
 }
 
@@ -630,7 +626,7 @@ HWTEST_F(DirectSubmissionTest,
                           Dispatcher::getSizeCacheFlush(directSubmission.rootDeviceEnvironment) +
                           directSubmission.getSizeSemaphoreSection(false) + directSubmission.getSizeNewResourceHandler();
 
-    size_t actualSize = directSubmission.getSizeDispatch(false, false);
+    size_t actualSize = directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false));
     EXPECT_EQ(expectedSize, actualSize);
 }
 
@@ -638,6 +634,7 @@ HWTEST_F(DirectSubmissionTest, givenDirectSubmissionWhenGetEndSizeThenExpectCorr
     using Dispatcher = RenderDispatcher<FamilyType>;
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
+    directSubmission.disableMonitorFence = false;
 
     size_t expectedSize = Dispatcher::getSizeStopCommandBuffer() +
                           Dispatcher::getSizeCacheFlush(directSubmission.rootDeviceEnvironment) +
@@ -821,9 +818,9 @@ HWTEST_F(DirectSubmissionTest,
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
     EXPECT_TRUE(UllsDefaults::defaultDisableCacheFlush);
-    EXPECT_FALSE(UllsDefaults::defaultDisableMonitorFence);
+    EXPECT_TRUE(UllsDefaults::defaultDisableMonitorFence);
     EXPECT_TRUE(directSubmission.disableCacheFlush);
-    EXPECT_FALSE(directSubmission.disableMonitorFence);
+    EXPECT_TRUE(directSubmission.disableMonitorFence);
     EXPECT_EQ(0u, directSubmission.workloadMode);
     EXPECT_EQ(nullptr, directSubmission.diagnostic.get());
     bool ret = directSubmission.initialize(false, false);
@@ -855,9 +852,9 @@ HWTEST_F(DirectSubmissionTest,
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
     EXPECT_TRUE(UllsDefaults::defaultDisableCacheFlush);
-    EXPECT_FALSE(UllsDefaults::defaultDisableMonitorFence);
+    EXPECT_TRUE(UllsDefaults::defaultDisableMonitorFence);
     EXPECT_TRUE(directSubmission.disableCacheFlush);
-    EXPECT_FALSE(directSubmission.disableMonitorFence);
+    EXPECT_TRUE(directSubmission.disableMonitorFence);
     EXPECT_EQ(0u, directSubmission.workloadMode);
     EXPECT_EQ(nullptr, directSubmission.diagnostic.get());
     bool ret = directSubmission.initialize(true, false);
@@ -895,9 +892,9 @@ HWTEST_F(DirectSubmissionTest,
     NEO::IoFunctions::mockFcloseCalled = 0u;
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
     EXPECT_TRUE(UllsDefaults::defaultDisableCacheFlush);
-    EXPECT_FALSE(UllsDefaults::defaultDisableMonitorFence);
+    EXPECT_TRUE(UllsDefaults::defaultDisableMonitorFence);
     EXPECT_TRUE(directSubmission.disableCacheFlush);
-    EXPECT_FALSE(directSubmission.disableMonitorFence);
+    EXPECT_TRUE(directSubmission.disableMonitorFence);
     EXPECT_EQ(0u, directSubmission.workloadMode);
     EXPECT_EQ(nullptr, directSubmission.diagnostic.get());
     bool ret = directSubmission.initialize(false, false);
@@ -935,7 +932,7 @@ HWTEST_F(DirectSubmissionTest,
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
     uint32_t expectedSemaphoreValue = directSubmission.currentQueueWorkCount;
     EXPECT_TRUE(UllsDefaults::defaultDisableCacheFlush);
-    EXPECT_FALSE(UllsDefaults::defaultDisableMonitorFence);
+    EXPECT_TRUE(UllsDefaults::defaultDisableMonitorFence);
     EXPECT_TRUE(directSubmission.disableCacheFlush);
     EXPECT_TRUE(directSubmission.disableMonitorFence);
     EXPECT_EQ(1u, directSubmission.workloadMode);
@@ -946,7 +943,7 @@ HWTEST_F(DirectSubmissionTest,
     size_t expectedSize = Dispatcher::getSizePreemption() +
                           directSubmission.getSizeSemaphoreSection(false) +
                           directSubmission.getDiagnosticModeSection();
-    expectedSize += expectedExecCount * (directSubmission.getSizeDispatch(false, false) - directSubmission.getSizeNewResourceHandler());
+    expectedSize += expectedExecCount * (directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false)) - directSubmission.getSizeNewResourceHandler());
 
     if (directSubmission.miMemFenceRequired) {
         expectedSize += directSubmission.getSizeSystemMemoryFenceAddress();
@@ -967,7 +964,7 @@ HWTEST_F(DirectSubmissionTest,
     EXPECT_EQ(expectedSemaphoreValue, directSubmission.currentQueueWorkCount);
 
     EXPECT_TRUE(directSubmission.disableCacheFlush);
-    EXPECT_FALSE(directSubmission.disableMonitorFence);
+    EXPECT_TRUE(directSubmission.disableMonitorFence);
     EXPECT_EQ(0u, directSubmission.workloadMode);
     EXPECT_EQ(nullptr, directSubmission.diagnostic.get());
 
@@ -1040,7 +1037,7 @@ HWTEST_F(DirectSubmissionTest,
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
     uint32_t expectedSemaphoreValue = directSubmission.currentQueueWorkCount;
     EXPECT_TRUE(UllsDefaults::defaultDisableCacheFlush);
-    EXPECT_FALSE(UllsDefaults::defaultDisableMonitorFence);
+    EXPECT_TRUE(UllsDefaults::defaultDisableMonitorFence);
     EXPECT_TRUE(directSubmission.disableCacheFlush);
     EXPECT_TRUE(directSubmission.disableMonitorFence);
     EXPECT_EQ(2u, directSubmission.workloadMode);
@@ -1051,7 +1048,7 @@ HWTEST_F(DirectSubmissionTest,
     size_t expectedSize = Dispatcher::getSizePreemption() +
                           directSubmission.getSizeSemaphoreSection(false);
     size_t expectedDispatch = directSubmission.getSizeSemaphoreSection(false);
-    EXPECT_EQ(expectedDispatch, directSubmission.getSizeDispatch(false, false) - directSubmission.getSizeNewResourceHandler());
+    EXPECT_EQ(expectedDispatch, directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false)) - directSubmission.getSizeNewResourceHandler());
     expectedSize += expectedExecCount * expectedDispatch;
 
     if (directSubmission.miMemFenceRequired) {
@@ -1073,7 +1070,7 @@ HWTEST_F(DirectSubmissionTest,
     EXPECT_EQ(expectedSemaphoreValue, directSubmission.currentQueueWorkCount);
 
     EXPECT_TRUE(directSubmission.disableCacheFlush);
-    EXPECT_FALSE(directSubmission.disableMonitorFence);
+    EXPECT_TRUE(directSubmission.disableMonitorFence);
     EXPECT_EQ(0u, directSubmission.workloadMode);
     EXPECT_EQ(nullptr, directSubmission.diagnostic.get());
 
diff --git a/shared/test/unit_test/direct_submission/direct_submission_tests_2.cpp b/shared/test/unit_test/direct_submission/direct_submission_tests_2.cpp
index c3c62711db..268300991a 100644
--- a/shared/test/unit_test/direct_submission/direct_submission_tests_2.cpp
+++ b/shared/test/unit_test/direct_submission/direct_submission_tests_2.cpp
@@ -244,7 +244,7 @@ HWCMDTEST_F(IGFX_XE_HP_CORE, DirectSubmissionDispatchBufferTest,
     EXPECT_EQ(1u, directSubmission.submitCount);
     EXPECT_EQ(2u, directSubmission.handleResidencyCount);
 
-    EXPECT_EQ(directSubmission.getSizeDispatch(false, false) - directSubmission.getSizeNewResourceHandler(), directSubmission.ringCommandStream.getUsed());
+    EXPECT_EQ(directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false)) - directSubmission.getSizeNewResourceHandler(), directSubmission.ringCommandStream.getUsed());
     EXPECT_TRUE(directSubmission.ringStart);
 
     HardwareParse hwParse;
@@ -415,22 +415,22 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
     DebugManager.flags.DirectSubmissionDisableCacheFlush.set(0);
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> regularDirectSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
-    size_t regularSizeDispatch = regularDirectSubmission.getSizeDispatch(false, false);
+    regularDirectSubmission.disableMonitorFence = false;
+    size_t regularSizeDispatch = regularDirectSubmission.getSizeDispatch(false, false, regularDirectSubmission.dispatchMonitorFenceRequired(false));
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
-    directSubmission.disableMonitorFence = true;
 
     bool ret = directSubmission.allocateResources();
     EXPECT_TRUE(ret);
 
     size_t tagUpdateSize = Dispatcher::getSizeMonitorFence(directSubmission.rootDeviceEnvironment);
 
-    size_t disabledSizeDispatch = directSubmission.getSizeDispatch(false, false);
+    size_t disabledSizeDispatch = directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false));
     EXPECT_EQ(disabledSizeDispatch, (regularSizeDispatch - tagUpdateSize));
 
     directSubmission.tagValueSetValue = 0x4343123ull;
     directSubmission.tagAddressSetValue = 0xBEEF00000ull;
-    directSubmission.dispatchWorkloadSection(batchBuffer);
+    directSubmission.dispatchWorkloadSection(batchBuffer, directSubmission.dispatchMonitorFenceRequired(batchBuffer.hasStallingCmds));
     size_t expectedDispatchSize = disabledSizeDispatch - directSubmission.getSizeNewResourceHandler();
     EXPECT_EQ(expectedDispatchSize, directSubmission.ringCommandStream.getUsed());
 
@@ -464,7 +464,7 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
     DebugManager.flags.DirectSubmissionDisableCacheFlush.set(0);
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> regularDirectSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
-    size_t regularSizeDispatch = regularDirectSubmission.getSizeDispatch(false, false);
+    size_t regularSizeDispatch = regularDirectSubmission.getSizeDispatch(false, false, regularDirectSubmission.dispatchMonitorFenceRequired(false));
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
 
@@ -474,10 +474,10 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
 
     size_t flushSize = Dispatcher::getSizeCacheFlush(directSubmission.rootDeviceEnvironment);
 
-    size_t disabledSizeDispatch = directSubmission.getSizeDispatch(false, false);
+    size_t disabledSizeDispatch = directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false));
     EXPECT_EQ(disabledSizeDispatch, (regularSizeDispatch - flushSize));
 
-    directSubmission.dispatchWorkloadSection(batchBuffer);
+    directSubmission.dispatchWorkloadSection(batchBuffer, directSubmission.dispatchMonitorFenceRequired(batchBuffer.hasStallingCmds));
     size_t expectedDispatchSize = disabledSizeDispatch - directSubmission.getSizeNewResourceHandler();
     EXPECT_EQ(expectedDispatchSize, directSubmission.ringCommandStream.getUsed());
 
@@ -511,8 +511,7 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
     using Dispatcher = RenderDispatcher<FamilyType>;
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> regularDirectSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
-
-    size_t regularSizeDispatch = regularDirectSubmission.getSizeDispatch(false, false);
+    size_t regularSizeDispatch = regularDirectSubmission.getSizeDispatch(false, false, regularDirectSubmission.dispatchMonitorFenceRequired(false));
 
     MockDirectSubmissionHw<FamilyType, Dispatcher> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
 
@@ -523,12 +522,12 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
     size_t startSize = directSubmission.getSizeStartSection();
     size_t storeDataSize = Dispatcher::getSizeStoreDwordCommand();
 
-    size_t debugSizeDispatch = directSubmission.getSizeDispatch(false, false);
+    size_t debugSizeDispatch = directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false));
     EXPECT_EQ(debugSizeDispatch, (regularSizeDispatch - startSize + storeDataSize));
 
     directSubmission.workloadModeOneExpectedValue = 0x40u;
     directSubmission.semaphoreGpuVa = 0xAFF0000;
-    directSubmission.dispatchWorkloadSection(batchBuffer);
+    directSubmission.dispatchWorkloadSection(batchBuffer, directSubmission.dispatchMonitorFenceRequired(batchBuffer.hasStallingCmds));
     size_t expectedDispatchSize = debugSizeDispatch - directSubmission.getSizeNewResourceHandler();
     EXPECT_EQ(expectedDispatchSize, directSubmission.ringCommandStream.getUsed());
 
@@ -556,7 +555,7 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
     using MI_STORE_DATA_IMM = typename FamilyType::MI_STORE_DATA_IMM;
 
     MockDirectSubmissionHw<FamilyType, RenderDispatcher<FamilyType>> regularDirectSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
-    size_t regularSizeDispatch = regularDirectSubmission.getSizeDispatch(false, false);
+    size_t regularSizeDispatch = regularDirectSubmission.getSizeDispatch(false, false, regularDirectSubmission.dispatchMonitorFenceRequired(false));
 
     MockDirectSubmissionHw<FamilyType, RenderDispatcher<FamilyType>> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
 
@@ -566,11 +565,11 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
 
     size_t startSize = directSubmission.getSizeStartSection();
 
-    size_t debugSizeDispatch = directSubmission.getSizeDispatch(false, false);
+    size_t debugSizeDispatch = directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false));
     EXPECT_EQ(debugSizeDispatch, (regularSizeDispatch - startSize));
 
     directSubmission.currentQueueWorkCount = 0x40u;
-    directSubmission.dispatchWorkloadSection(batchBuffer);
+    directSubmission.dispatchWorkloadSection(batchBuffer, directSubmission.dispatchMonitorFenceRequired(batchBuffer.hasStallingCmds));
     size_t expectedDispatchSize = debugSizeDispatch - directSubmission.getSizeNewResourceHandler();
     EXPECT_EQ(expectedDispatchSize, directSubmission.ringCommandStream.getUsed());
 
@@ -624,7 +623,7 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
     EXPECT_EQ(1u, directSubmission.submitCount);
     EXPECT_EQ(2u, directSubmission.handleResidencyCount);
 
-    EXPECT_EQ(sizeUsed + directSubmission.getSizeDispatch(false, false) - directSubmission.getSizeNewResourceHandler(), directSubmission.ringCommandStream.getUsed());
+    EXPECT_EQ(sizeUsed + directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false)) - directSubmission.getSizeNewResourceHandler(), directSubmission.ringCommandStream.getUsed());
     EXPECT_TRUE(directSubmission.ringStart);
 
     HardwareParse hwParse;
@@ -666,7 +665,7 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
     EXPECT_EQ(oldRingAllocation->getGpuAddress(), directSubmission.submitGpuAddress);
     EXPECT_EQ(2u, directSubmission.handleResidencyCount);
 
-    size_t dispatchSize = submitSize + directSubmission.getSizeDispatch(false, false) - directSubmission.getSizeNewResourceHandler();
+    size_t dispatchSize = submitSize + directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false)) - directSubmission.getSizeNewResourceHandler();
 
     EXPECT_EQ(dispatchSize, directSubmission.ringCommandStream.getUsed());
     EXPECT_TRUE(directSubmission.ringStart);
@@ -709,7 +708,7 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
     EXPECT_EQ(1u, directSubmission.submitCount);
     EXPECT_EQ(2u, directSubmission.handleResidencyCount);
 
-    EXPECT_EQ(directSubmission.getSizeDispatch(false, false) - directSubmission.getSizeNewResourceHandler(), directSubmission.ringCommandStream.getUsed());
+    EXPECT_EQ(directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false)) - directSubmission.getSizeNewResourceHandler(), directSubmission.ringCommandStream.getUsed());
     EXPECT_TRUE(directSubmission.ringStart);
 }
 
@@ -746,7 +745,7 @@ HWTEST_F(DirectSubmissionDispatchBufferTest,
     EXPECT_EQ(submitSize, directSubmission.submitSize);
     EXPECT_EQ(2u, directSubmission.handleResidencyCount);
 
-    size_t dispatchSize = submitSize + directSubmission.getSizeDispatch(false, false) - directSubmission.getSizeNewResourceHandler();
+    size_t dispatchSize = submitSize + directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false)) - directSubmission.getSizeNewResourceHandler();
 
     EXPECT_EQ(dispatchSize, directSubmission.ringCommandStream.getUsed());
     EXPECT_TRUE(directSubmission.ringStart);
@@ -1827,7 +1826,7 @@ HWTEST_F(DirectSubmissionRelaxedOrderingTests, givenNotEnoughSpaceForTaskStoreSe
     directSubmission.ringCommandStream.getUsed();
 
     auto sizeToConsume = directSubmission.ringCommandStream.getAvailableSpace() -
-                         (directSubmission.getSizeDispatch(false, false) + directSubmission.getSizeEnd(false) + directSubmission.getSizeSwitchRingBufferSection());
+                         (directSubmission.getSizeDispatch(false, false, directSubmission.dispatchMonitorFenceRequired(false)) + directSubmission.getSizeEnd(false) + directSubmission.getSizeSwitchRingBufferSection());
 
     directSubmission.ringCommandStream.getSpace(sizeToConsume);
 
@@ -2266,7 +2265,7 @@ HWTEST2_F(DirectSubmissionRelaxedOrderingTests, whenDispatchingWorkloadSectionTh
     auto originalBbStart = *reinterpret_cast<MI_BATCH_BUFFER_START *>(batchBuffer.endCmdPtr);
 
     batchBuffer.hasRelaxedOrderingDependencies = true;
-    directSubmission.dispatchWorkloadSection(batchBuffer);
+    directSubmission.dispatchWorkloadSection(batchBuffer, directSubmission.dispatchMonitorFenceRequired(batchBuffer.hasStallingCmds));
 
     uint64_t returnPtr = directSubmission.ringCommandStream.getGpuBase() + offset + (4 * sizeof(MI_LOAD_REGISTER_IMM)) + directSubmission.getSizeStartSection();
 
@@ -2290,7 +2289,7 @@ HWTEST2_F(DirectSubmissionRelaxedOrderingTests, givenBbWithoutRelaxedOrderingDep
     auto offset = directSubmission.ringCommandStream.getUsed();
 
     batchBuffer.hasRelaxedOrderingDependencies = false;
-    directSubmission.dispatchWorkloadSection(batchBuffer);
+    directSubmission.dispatchWorkloadSection(batchBuffer, directSubmission.dispatchMonitorFenceRequired(batchBuffer.hasStallingCmds));
 
     auto lriCmd = genCmdCast<MI_LOAD_REGISTER_IMM *>(ptrOffset(directSubmission.ringCommandStream.getCpuBase(), offset));
     EXPECT_EQ(nullptr, lriCmd);
@@ -2696,8 +2695,8 @@ HWTEST2_F(DirectSubmissionRelaxedOrderingTests, givenSchedulerRequiredWhenDispat
 HWTEST2_F(DirectSubmissionRelaxedOrderingTests, givenReturnPtrsRequiredWhenAskingForDispatchSizeTheAddMmioSizes, IsAtLeastXeHpcCore) {
     MockDirectSubmissionHw<FamilyType, RenderDispatcher<FamilyType>> directSubmission(*pDevice->getDefaultEngine().commandStreamReceiver);
 
-    size_t baseSize = directSubmission.getSizeDispatch(true, false);
-    size_t sizeWitfRetPtr = directSubmission.getSizeDispatch(true, true);
+    size_t baseSize = directSubmission.getSizeDispatch(true, false, directSubmission.dispatchMonitorFenceRequired(false));
+    size_t sizeWitfRetPtr = directSubmission.getSizeDispatch(true, true, directSubmission.dispatchMonitorFenceRequired(false));
 
     EXPECT_EQ(baseSize + RelaxedOrderingHelper::getSizeReturnPtrRegs<FamilyType>(), sizeWitfRetPtr);
 }
diff --git a/shared/test/unit_test/mocks/windows/mock_wddm_direct_submission.h b/shared/test/unit_test/mocks/windows/mock_wddm_direct_submission.h
index 68ae369a3d..2329dc3a9e 100644
--- a/shared/test/unit_test/mocks/windows/mock_wddm_direct_submission.h
+++ b/shared/test/unit_test/mocks/windows/mock_wddm_direct_submission.h
@@ -20,6 +20,7 @@ struct MockWddmDirectSubmission : public WddmDirectSubmission<GfxFamily, Dispatc
     using BaseClass::completionFenceAllocation;
     using BaseClass::currentRingBuffer;
     using BaseClass::disableMonitorFence;
+    using BaseClass::dispatchMonitorFenceRequired;
     using BaseClass::getSizeDispatch;
     using BaseClass::getSizeNewResourceHandler;
     using BaseClass::getSizeSemaphoreSection;
diff --git a/shared/test/unit_test/os_interface/windows/device_command_stream_tests.cpp b/shared/test/unit_test/os_interface/windows/device_command_stream_tests.cpp
index 70864aa71e..4182669e4e 100644
--- a/shared/test/unit_test/os_interface/windows/device_command_stream_tests.cpp
+++ b/shared/test/unit_test/os_interface/windows/device_command_stream_tests.cpp
@@ -1238,7 +1238,7 @@ HWTEST_TEMPLATED_F(WddmCommandStreamMockGdiTest, givenDirectSubmissionEnabledOnR
     size_t actualDispatchSize = directSubmission->ringCommandStream.getUsed();
     size_t expectedSize = directSubmission->getSizeSemaphoreSection(false) +
                           Dispatcher::getSizePreemption() +
-                          directSubmission->getSizeDispatch(false, false);
+                          directSubmission->getSizeDispatch(false, false, directSubmission->dispatchMonitorFenceRequired(false));
 
     if (directSubmission->miMemFenceRequired) {
         expectedSize += directSubmission->getSizeSystemMemoryFenceAddress();
@@ -1286,7 +1286,7 @@ HWTEST_TEMPLATED_F(WddmCommandStreamMockGdiTest, givenDirectSubmissionEnabledOnB
     size_t actualDispatchSize = directSubmission->ringCommandStream.getUsed();
     size_t expectedSize = directSubmission->getSizeSemaphoreSection(false) +
                           Dispatcher::getSizePreemption() +
-                          directSubmission->getSizeDispatch(false, false);
+                          directSubmission->getSizeDispatch(false, false, directSubmission->dispatchMonitorFenceRequired(false));
 
     if (directSubmission->miMemFenceRequired) {
         expectedSize += directSubmission->getSizeSystemMemoryFenceAddress();