Merge branch 'master' into tj/core/serialize-validation

t-jankowski · Aug 23, 2024 · 8d3bb34 · 8d3bb34
2 parents 1fcccae + fa2d87a
commit 8d3bb34
Show file tree

Hide file tree

Showing 28 changed files with 885 additions and 219 deletions.
diff --git a/docs/articles_en/about-openvino/performance-benchmarks.rst b/docs/articles_en/about-openvino/performance-benchmarks.rst
@@ -206,12 +206,11 @@ You can also test performance for your system yourself, following the guide on
 
    <h2>Disclaimers</h2>
 
-
 * Intel® Distribution of OpenVINO™ toolkit performance results are based on release
   2024.3, as of July 31, 2024.
 
 * OpenVINO Model Server performance results are based on release
-  2024.2, as of June 28, 2024.
+  2024.3, as of Aug. 19, 2024.
 
 The results may not reflect all publicly available updates. Intel technologies' features and
 benefits depend on system configuration and may require enabled hardware, software, or service

diff --git a/...vino-ir-format/operation-sets/operation-specs/bitwise/bitwise-left-shift-15.rst b/...vino-ir-format/operation-sets/operation-specs/bitwise/bitwise-left-shift-15.rst
@@ -25,7 +25,10 @@ After broadcasting input tensors *a* and *b*, *BitwiseLeftShift* performs a bitw
 .. note::
 
     If the number of shifts is negative, or if it equals or exceeds the total number of bits in the type **T**, the behavior can be undefined or implementation-defined (depends on the hardware).
+
+    Unsigned integer shift is always performed modulo 2^n where n is the number of bits in the type **T**.
 
+    When signed integer shift operation overflows (the result does not fit in the result type), the behavior is undefined.
 
 **Attributes**:
 

diff --git a/...ino-ir-format/operation-sets/operation-specs/bitwise/bitwise-right-shift-15.rst b/...ino-ir-format/operation-sets/operation-specs/bitwise/bitwise-right-shift-15.rst
@@ -26,6 +26,9 @@ After broadcasting input tensors *a* and *b*, *BitwiseRightShift* performs a bit
 
     If the number of shifts is negative, or if it equals or exceeds the total number of bits in the type **T**, the behavior can be undefined or implementation-defined (depends on the hardware).
 
+    Unsigned integer shift is always performed modulo 2^n where n is the number of bits in the type **T**.
+
+    When signed integer shift operation overflows (the result does not fit in the result type), the behavior is undefined.
 
 **Attributes**:
 

diff --git a/docs/sphinx_setup/_static/benchmarks_files/OV-2024.3-Performance-Data.xlsx b/docs/sphinx_setup/_static/benchmarks_files/OV-2024.3-Performance-Data.xlsx
diff --git a/docs/sphinx_setup/_static/benchmarks_files/OV-2024.3-platform_list.pdf b/docs/sphinx_setup/_static/benchmarks_files/OV-2024.3-platform_list.pdf
diff --git a/docs/sphinx_setup/_static/benchmarks_files/OV-2024.3-system-info-detailed.xlsx b/docs/sphinx_setup/_static/benchmarks_files/OV-2024.3-system-info-detailed.xlsx
diff --git a/docs/sphinx_setup/_static/benchmarks_files/OVMS-benchmark-data.csv b/docs/sphinx_setup/_static/benchmarks_files/OVMS-benchmark-data.csv
@@ -1,78 +1,78 @@
-Network model,Release,IE-Type,Platform name,Throughput-OVMS-INT8,Throughput-OV-INT8,Throughput-OVMS-FP32,Throughput-OV-FP32,
-begin_rec,,,,,,,,
-bert-base-cased,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,484.765,486.962,181.829,179.94
-bert-base-cased,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,430.151,434.276,157.835,158.277
-bert-base-cased,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,101.044,102.838,35.727,36.57
-bert-base-cased,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,100.741,103.322,35.046,36.607
-bert-base-cased,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,26.124,26.329,17.155,17.387
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-bert-large-uncased,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,41.302,41.994,14.937,14.482
-bert-large-uncased,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,36.595,37.1,13.114,13.03
-bert-large-uncased,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,10.076,10.145,3.267,3.246
-bert-large-uncased,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,10.161,10.203,3.287,3.26
-bert-large-uncased,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,2.422,2.424,1.447,1.427
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-Efficientdet-D0,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,433.166,479.055,285.65,287.547
-Efficientdet-D0,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,367.395,407,251.401,257.516
-Efficientdet-D0,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,132.153,149.424,57.682,61.811
-Efficientdet-D0,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,124.984,142.514,50.265,53.089
-Efficientdet-D0,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,47.048,50.328,30.026,31.473
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-mask_rcnn_resnet50_atrous_coco,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,6.323,6.488,1.978,1.868
-mask_rcnn_resnet50_atrous_coco,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,5.637,5.742,1.715,1.633
-mask_rcnn_resnet50_atrous_coco,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,1.302,1.276,0.396,0.373
-mask_rcnn_resnet50_atrous_coco,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,1.307,1.28,0.374,0.36
-mask_rcnn_resnet50_atrous_coco,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,0.381,0.35,0.181,0.15
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-Mobilenet-V2 ,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,7578.115,12346.3,3354.203,3938.523
-Mobilenet-V2 ,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,7513.034,10367.947,2915.906,3349.306
-Mobilenet-V2 ,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,2152.015,2740.691,745.81,882.839
-Mobilenet-V2 ,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,2093.311,2822.613,667.391,795.616
-Mobilenet-V2 ,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,615.392,719.715,381.686,454.574
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-Resnet-50,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,2354.173,2482.832,639.577,645.443
-Resnet-50,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,2070.726,2177.751,571.252,575.778
-Resnet-50,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,440.402,458.622,114.169,116.577
-Resnet-50,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,448.464,470.586,111.785,114.628
-Resnet-50,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,114.267,118.502,57.29,58.233
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-SSD-Resnet34-1200 ,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,44.587,47.293,12.111,12.248
-SSD-Resnet34-1200 ,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,38.784,40.602,10.521,10.613
-SSD-Resnet34-1200 ,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,7.736,7.821,2.034,2.011
-SSD-Resnet34-1200 ,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,7.953,8.033,2.083,2.058
-SSD-Resnet34-1200 ,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,1.951,1.936,1.04,1.014
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-SSD_Mobilenet_V1_Coco,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,4753.674,4933.241,1370.423,1379.026
-SSD_Mobilenet_V1_Coco,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,4165.318,4276.949,1197.151,1222.112
-SSD_Mobilenet_V1_Coco,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,920.957,1001.56,270.597,281.293
-SSD_Mobilenet_V1_Coco,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,941.323,1030.464,256.649,266.715
-SSD_Mobilenet_V1_Coco,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,256.291,266.245,129.84,135.453
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-Unet-Camvid--0001 ,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,73.871,78.016,18.23,18.374
-Unet-Camvid--0001 ,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,64.573,67.713,15.815,16.023
-Unet-Camvid--0001 ,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,12.572,12.669,3.28,3.254
-Unet-Camvid--0001 ,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,12.779,12.894,3.315,3.299
-Unet-Camvid--0001 ,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,2.99,2.971,1.549,1.542
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-Yolo_V3_Tiny,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,1863.229,2344.128,775.001,786.09
-Yolo_V3_Tiny,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,1669.35,2066.437,675.625,703.203
-Yolo_V3_Tiny,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,430.441,505.532,146.284,151.09
-Yolo_V3_Tiny,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,419.347,513.112,142.009,148.117
-Yolo_V3_Tiny,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,117.889,133.312,63.598,69.377
-end_rec,,,,,,,,
-begin_rec,,,,,,,,
-Yolo_V8n,OV-2024.2,xeon,Intel®  Xeon® 8260M CPU-only,,705.714,845.484,316.319,388.763
-Yolo_V8n,OV-2024.2,xeon,Intel®  Xeon® Gold 6238M CPU-only,,641.815,746.965,278.824,338.806
-Yolo_V8n,OV-2024.2,core,Intel® Core™  i9-11900K CPU-only,,154.54,205.294,67.744,75.634
-Yolo_V8n,OV-2024.2,core,Intel® Core™  i7-11700K CPU-only,,149.289,199.997,65.243,72.677
-Yolo_V8n,OV-2024.2,core,Intel® Core™  i3-10100 CPU-only,,55.999,68.155,34.454,38.544
-end_rec,,,,,,,,
+Network model,Release,IE-Type,Platform name,Throughput-OVMS-INT8,Throughput-OV-INT8,Throughput-OVMS-FP32,Throughput-OV-FP32,UOM_T
+begin_rec,,,,,,,,
+bert-base-cased,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,479.649,482.878,180.7,179.541,FPS
+bert-base-cased,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,428.173,430.397,156.73,159.276,FPS
+bert-base-cased,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,100.783,101.983,35.711,36.35,FPS
+bert-base-cased,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,98.441,102.62,34.303,36.096,FPS
+bert-base-cased,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,26.185,26.436,17.108,17.395,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+bert-large-uncased,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,41.872,42.401,14.949,14.473,FPS
+bert-large-uncased,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,37.05,37.864,13.075,13.031,FPS
+bert-large-uncased,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,10.047,10.111,3.259,3.237,FPS
+bert-large-uncased,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,9.961,10.167,3.236,3.224,FPS
+bert-large-uncased,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,2.43,2.427,1.447,1.428,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+Efficientdet-D0,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,439.435,485.287,274.772,272.856,FPS
+Efficientdet-D0,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,376.1,415.275,253.829,259.188,FPS
+Efficientdet-D0,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,131.735,148.558,57.036,59.907,FPS
+Efficientdet-D0,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,119.798,140.129,,,FPS
+Efficientdet-D0,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,47.382,50.573,30.226,31.492,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+mask_rcnn_resnet50_atrous_coco,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,6.306,6.364,1.96,1.868,FPS
+mask_rcnn_resnet50_atrous_coco,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,5.652,5.771,1.714,1.639,FPS
+mask_rcnn_resnet50_atrous_coco,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,1.309,1.267,0.396,0.371,FPS
+mask_rcnn_resnet50_atrous_coco,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,1.293,1.271,0.355,0.346,FPS
+mask_rcnn_resnet50_atrous_coco,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,0.38,0.352,0.182,0.151,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+Mobilenet-V2 ,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,7563.199,12406.597,3336.015,3972.673,FPS
+Mobilenet-V2 ,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,7475.62,10373.146,2934.976,3381.725,FPS
+Mobilenet-V2 ,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,2158.818,2742.363,740.988,874.037,FPS
+Mobilenet-V2 ,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,2042.633,2809.471,631.59,759.984,FPS
+Mobilenet-V2 ,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,614.174,718.416,381.882,455.793,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+Resnet-50,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,2356.238,2483.3,628.616,635.411,FPS
+Resnet-50,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,2071.836,2202.317,568.945,575.057,FPS
+Resnet-50,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,440.533,458.665,113.442,116.116,FPS
+Resnet-50,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,441.7,469.848,107.395,113.605,FPS
+Resnet-50,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,114.045,118.024,57.165,58.366,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+SSD-Resnet34-1200 ,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,44.499,47.251,12.074,12.167,FPS
+SSD-Resnet34-1200 ,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,38.714,40.662,10.504,10.653,FPS
+SSD-Resnet34-1200 ,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,7.756,7.818,2.029,2.005,FPS
+SSD-Resnet34-1200 ,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,7.929,8.032,2.072,2.054,FPS
+SSD-Resnet34-1200 ,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,1.947,1.937,1.037,1.008,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+SSD_Mobilenet_V1_Coco,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,4732.691,4875.291,1362.268,1375.237,FPS
+SSD_Mobilenet_V1_Coco,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,4168.575,4279.825,1199.883,1226.189,FPS
+SSD_Mobilenet_V1_Coco,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,921.041,1001.672,268.066,280.987,FPS
+SSD_Mobilenet_V1_Coco,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,915.4,1028.233,244.534,260.822,FPS
+SSD_Mobilenet_V1_Coco,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,256.018,266.401,129.917,135.312,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+Unet-Camvid--0001 ,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,73.429,77.693,18.104,17.938,FPS
+Unet-Camvid--0001 ,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,64.29,67.517,15.777,15.927,FPS
+Unet-Camvid--0001 ,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,12.574,12.628,3.267,3.253,FPS
+Unet-Camvid--0001 ,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,12.718,12.881,3.272,3.297,FPS
+Unet-Camvid--0001 ,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,2.995,2.976,1.555,1.53,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+Yolo_V3_Tiny,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,1842.129,2317.052,755.451,777.681,FPS
+Yolo_V3_Tiny,OV-2024.3,xeon,Intel®  Xeon® Gold 6238M CPU-only,1667.812,2056.27,675.447,704.412,FPS
+Yolo_V3_Tiny,OV-2024.3,core,Intel® Core™  i9-11900K CPU-only,431.387,504.093,145.92,151.499,FPS
+Yolo_V3_Tiny,OV-2024.3,core,Intel® Core™  i7-11700K CPU-only,409.268,516.794,139.903,147.235,FPS
+Yolo_V3_Tiny,OV-2024.3,core,Intel® Core™  i3-10100 CPU-only,117.276,133.578,65.341,69.29,FPS
+end_rec,,,,,,,,
+begin_rec,,,,,,,,
+Yolo_V8n,OV-2024.3,xeon,Intel®  Xeon® 8260M CPU-only,,,314.652,386.299,FPS
+Yolo_V8n,OV-2024.34,xeon,Intel®  Xeon® Gold 6238M CPU-only,,,282.302,340.845,FPS
+Yolo_V8n,OV-2024.65,core,Intel® Core™  i9-11900K CPU-only,153.817,204.691,67.421,74.996,FPS
+Yolo_V8n,OV-2024.96,core,Intel® Core™  i7-11700K CPU-only,143.19,197.409,62.948,70.913,FPS
+Yolo_V8n,OV-2024.127,core,Intel® Core™  i3-10100 CPU-only,56.244,67.968,34.396,38.576,FPS
+end_rec,,,,,,,,
diff --git a/src/plugins/intel_cpu/src/cpu_types.cpp b/src/plugins/intel_cpu/src/cpu_types.cpp
@@ -94,6 +94,8 @@ static const TypeToNameMap& get_type_to_name_tbl() {
         {"BitwiseNot", Type::Eltwise},
         {"BitwiseOr", Type::Eltwise},
         {"BitwiseXor", Type::Eltwise},
+        {"BitwiseLeftShift", Type::Eltwise},
+        {"BitwiseRightShift", Type::Eltwise},
         {"Reshape", Type::Reshape},
         {"Squeeze", Type::Reshape},
         {"Unsqueeze", Type::Reshape},
@@ -445,6 +447,8 @@ std::string algToString(const Algorithm alg) {
         CASE(EltwiseBitwiseNot);
         CASE(EltwiseBitwiseOr);
         CASE(EltwiseBitwiseXor);
+        CASE(EltwiseBitwiseLeftShift);
+        CASE(EltwiseBitwiseRightShift);
         CASE(FQCommon);
         CASE(FQQuantization);
         CASE(FQBinarization);

diff --git a/src/plugins/intel_cpu/src/cpu_types.h b/src/plugins/intel_cpu/src/cpu_types.h
@@ -201,6 +201,8 @@ enum class Algorithm {
     EltwiseBitwiseNot,
     EltwiseBitwiseOr,
     EltwiseBitwiseXor,
+    EltwiseBitwiseLeftShift,
+    EltwiseBitwiseRightShift,
 
     // FakeQuantize algorithms
     FQCommon,

diff --git a/src/plugins/intel_cpu/src/dnnl_extension_utils.cpp b/src/plugins/intel_cpu/src/dnnl_extension_utils.cpp
@@ -266,8 +266,8 @@ bool DnnlExtensionUtils::isUnarySupportedAsPostOp(Algorithm alg) {
 #endif
 }
 
-std::string DnnlExtensionUtils::computeWeightsStringHash(const std::shared_ptr<const IMemory> memory,
-                                                         const std::shared_ptr<DnnlMemoryDesc> dstDesc) {
+std::string DnnlExtensionUtils::computeWeightsStringHash(const std::shared_ptr<const IMemory>& memory,
+                                                         const std::shared_ptr<DnnlMemoryDesc>& dstDesc) {
     const auto desc_hash = dnnl::impl::primitive_hashing::get_md_hash(*dstDesc->getDnnlDesc().get());
     return std::to_string(desc_hash) + "_" + std::to_string(reinterpret_cast<uint64_t>(memory->getData()));
 }

diff --git a/src/plugins/intel_cpu/src/dnnl_extension_utils.h b/src/plugins/intel_cpu/src/dnnl_extension_utils.h
@@ -102,13 +102,15 @@ class DnnlExtensionUtils {
     static dnnl_memory_desc_t clone_desc(const_dnnl_memory_desc_t cdesc);
     static const char* query_pd_info(const_dnnl_primitive_desc_t pd);
     static bool isUnarySupportedAsPostOp(Algorithm alg);
+
     /**
      * @brief Computes weights string hash based on weights memory and requested descriptor
      * @param memory Weights memory pointer
      * @param dstDesc descriptor defining weights representation after repacking
      * @return string hash
      */
-    static std::string computeWeightsStringHash(const std::shared_ptr<const IMemory> memory, const std::shared_ptr<DnnlMemoryDesc> dstDesc);
+    static std::string computeWeightsStringHash(const std::shared_ptr<const IMemory>& memory,
+                                                const std::shared_ptr<DnnlMemoryDesc>& dstDesc);
 };
 
 }   // namespace intel_cpu

diff --git a/src/plugins/intel_cpu/src/node.h b/src/plugins/intel_cpu/src/node.h
@@ -737,7 +737,7 @@ class Node {
                               impl_desc_type implType);
 
     void prepareMemory(const std::vector<DnnlMemoryDescPtr>& intDescs);
-    void prepareMemory(const DnnlMemoryDescPtr& intDesc, size_t indx);
+    virtual void prepareMemory(const DnnlMemoryDescPtr& intDesc, size_t indx);
     void prepareMemory(dnnl::primitive_desc_iterator& itpd);
 
     MemoryPtr prepareWeightMemory(DnnlMemoryDescPtr dstWeightDesc, DnnlMemoryDescPtr srcWeightDesc = nullptr);

diff --git a/src/plugins/intel_cpu/src/nodes/batch_to_space.cpp b/src/plugins/intel_cpu/src/nodes/batch_to_space.cpp
@@ -183,11 +183,11 @@ void BatchToSpace::batchToSpaceKernel() {
             begin[1] = (blockShape[1] - 1 - oAdd[1]) / blockShape[1] / blockSize;
             finish[1] = (outShape5D[1] - 1 - oAdd[1]) / blockShape[1] / blockSize;
             begin[2] = (blockShape[2] - 1 - oAdd[2]) / blockShape[2];
-            finish[2] = (outShape5D[2] - 1 - oAdd[2]) / blockShape[2];
+            finish[2] = (outShape5D[2] + blockShape[2] - 1 - oAdd[2]) / blockShape[2];
             begin[3] = (blockShape[3] - 1 - oAdd[3]) / blockShape[3];
-            finish[3] = (outShape5D[3] - 1 - oAdd[3]) / blockShape[3];
+            finish[3] = (outShape5D[3] + blockShape[3] - 1 - oAdd[3]) / blockShape[3];
             begin[4] = (blockShape[4] - 1 - oAdd[4]) / blockShape[4];
-            finish[4] = (outShape5D[4] - 1 - oAdd[4]) / blockShape[4];
+            finish[4] = (outShape5D[4] + blockShape[4] - 1 - oAdd[4]) / blockShape[4];
             const int64_t addTmpOC = blocked ? 0lu : oAdd[1];
             const int64_t addTmpOc = blocked ? oAdd[1] : 0lu;
             indxStart[1] = begin[1] > indxStart[1] ? begin[1] : indxStart[1];
@@ -198,15 +198,15 @@ void BatchToSpace::batchToSpaceKernel() {
                 const size_t srcIdx1 = srcIdx0 + indxStart[1] * inSpatialStep * blockSize;
                 const size_t dstIdx1 = dstIdx0 + tmpOC * outSpatialStep * blockSize;
                 const size_t itEnd = blocked ? ((block - 1) * blockShape[1] + oAdd[1]) / blockSize : 0lu;
-                for (size_t i2 = begin[2]; i2 < finish[2] + 1; ++i2) {
+                for (size_t i2 = begin[2]; i2 < finish[2]; ++i2) {
                     const int64_t tmpOd = i2 * blockShape[2] + oAdd[2];
                     const size_t srcIdx2 = srcIdx1 + i2 * inShape5D[3] * inShape5D[4] * blockSize;
                     const size_t dstIdx2 = dstIdx1 + tmpOd * outShape5D[3] * outShape5D[4] * blockSize;
-                    for (size_t i3 = begin[3]; i3 < finish[3] + 1; ++i3) {
+                    for (size_t i3 = begin[3]; i3 < finish[3]; ++i3) {
                         const int64_t tmpOh = i3 * blockShape[3] + oAdd[3];
                         const size_t srcIdx3 = srcIdx2 + i3 * inShape5D[4] * blockSize;
                         const size_t dstIdx3 = dstIdx2 + tmpOh * outShape5D[4] * blockSize;
-                        for (size_t i4 = begin[4]; i4 < finish[4] + 1; ++i4) {
+                        for (size_t i4 = begin[4]; i4 < finish[4]; ++i4) {
                             const int64_t tmpOw = i4 * blockShape[4] + oAdd[4];
                             const size_t srcIdx4 = srcIdx3 + i4 * blockSize;
                             const size_t dstIdx4 = dstIdx3 + tmpOw * blockSize;

diff --git a/src/plugins/intel_cpu/src/nodes/batch_to_space.h b/src/plugins/intel_cpu/src/nodes/batch_to_space.h
@@ -16,6 +16,12 @@ class BatchToSpace : public Node {
 
     void getSupportedDescriptors() override {};
     void initSupportedPrimitiveDescriptors() override;
+
+    // output shape can potentially be empty
+    bool isExecutable() const override {
+        return !hasEmptyInputTensors() && !hasEmptyOutputTensors();
+    }
+
     void execute(dnnl::stream strm) override;
     bool created() const override;