已合并
修改Conv3DBackpropInputV2算子aicore error #8090
yejiani创建于 7月30日
修改Conv3DBackpropInputV2算子aicore error #8090
已合并
共 3 个文件变更+17-17
| @@ -1,3 +1,3 @@ | |||
| 1 | -{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 1, "bl0Pbuffer": 1, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 16, "baseK": 168, "baseN": 96, "stepKa": 3, "stepKb": 3, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 28, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 1 | +{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 1, "bl0Pbuffer": 1, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 16, "baseK": 168, "baseN": 96, "stepKa": 3, "stepKb": 3, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 28, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 2 | -{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 1, "bl0Pbuffer": 1, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 16, "baseK": 168, "baseN": 96, "stepKa": 3, "stepKb": 3, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 28, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 2 | +{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 1, "bl0Pbuffer": 1, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 16, "baseK": 168, "baseN": 96, "stepKa": 3, "stepKb": 3, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 28, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 3 | -{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 6, "dedx_cin": 384, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 1, "bl0Pbuffer": 1, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 384, "coutG": 384, "cout1": 48, "cin1": 48, "cout1G": 48, "cin1G": 24, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 48, "baseK": 168, "baseN": 96, "stepKa": 3, "stepKb": 3, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 28, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 3 | +{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 6, "dedx_cin": 384, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 1, "bl0Pbuffer": 1, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 384, "coutG": 384, "cout1": 48, "cin1": 48, "cout1G": 48, "cin1G": 24, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 48, "baseK": 168, "baseN": 96, "stepKa": 3, "stepKb": 3, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 28, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| @@ -1,7 +1,7 @@ | |||
| 1 | -{"id": 0, "info_dict": {"batch_n": 64, "groups": 1, "dedx_d": 1, "dedx_cin": 128, "dedx_h": 32, "dedx_w": 32, "dedy_d": 1, "dedy_cout": 256, "dedy_h": 16, "dedy_w": 16, "kernel_d": 1, "kernel_h": 5, "kernel_w": 5, "stride_d": 1, "stride_h": 2, "stride_w": 2, "pad_h": 0, "pad_t": 0, "pad_u": 1, "pad_d": 2, "pad_l": 1, "pad_r": 2, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 0, "backprop_pad_t": 0, "backprop_pad_u": 3, "backprop_pad_d": 1, "backprop_pad_l": 3, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 0, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 128, "coutG": 256, "cout1": 32, "cin1": 16, "cout1G": 32, "cin1G": 8, "backpropPadTail": 0, "backpropPadUp": 3, "backpropPadDown": 2, "backpropPadLeft": 3, "backpropPadRight": 2, "singleCoreGroup": 1, "singleCoreCout": 256, "singleCoreCin": 128, "singleCoreDin": 1, "baseM": 256, "baseK": 8, "baseN": 128, "stepKa": 45, "stepKb": 45, "singleIterateDk": 1, "singleCoreBatch": 1, "singleCoreM": 256, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 1, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 1, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": true, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 16777216, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 1 | +{"id": 0, "info_dict": {"batch_n": 64, "groups": 1, "dedx_d": 1, "dedx_cin": 128, "dedx_h": 32, "dedx_w": 32, "dedy_d": 1, "dedy_cout": 256, "dedy_h": 16, "dedy_w": 16, "kernel_d": 1, "kernel_h": 5, "kernel_w": 5, "stride_d": 1, "stride_h": 2, "stride_w": 2, "pad_h": 0, "pad_t": 0, "pad_u": 1, "pad_d": 2, "pad_l": 1, "pad_r": 2, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 0, "backprop_pad_t": 0, "backprop_pad_u": 3, "backprop_pad_d": 1, "backprop_pad_l": 3, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 0, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 128, "coutG": 256, "cout1": 32, "cin1": 16, "cout1G": 32, "cin1G": 8, "backpropPadTail": 0, "backpropPadUp": 3, "backpropPadDown": 2, "backpropPadLeft": 3, "backpropPadRight": 2, "singleCoreGroup": 1, "singleCoreCout": 256, "singleCoreCin": 128, "singleCoreDin": 1, "baseM": 256, "baseK": 8, "baseN": 128, "stepKa": 45, "stepKb": 45, "singleIterateDk": 1, "singleCoreBatch": 1, "singleCoreM": 256, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 1, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 1, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": true, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 16777216, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 2 | -{"id": 0, "info_dict": {"batch_n": 64, "groups": 1, "dedx_d": 1, "dedx_cin": 32, "dedx_h": 64, "dedx_w": 64, "dedy_d": 1, "dedy_cout": 128, "dedy_h": 32, "dedy_w": 32, "kernel_d": 1, "kernel_h": 5, "kernel_w": 5, "stride_d": 1, "stride_h": 2, "stride_w": 2, "pad_h": 0, "pad_t": 0, "pad_u": 1, "pad_d": 2, "pad_l": 1, "pad_r": 2, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 0, "backprop_pad_t": 0, "backprop_pad_u": 3, "backprop_pad_d": 1, "backprop_pad_l": 3, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 0, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 32, "coutG": 128, "cout1": 16, "cin1": 4, "cout1G": 16, "cin1G": 2, "backpropPadTail": 0, "backpropPadUp": 3, "backpropPadDown": 2, "backpropPadLeft": 3, "backpropPadRight": 2, "singleCoreGroup": 1, "singleCoreCout": 128, "singleCoreCin": 32, "singleCoreDin": 1, "baseM": 1024, "baseK": 8, "baseN": 32, "stepKa": 54, "stepKb": 54, "singleIterateDk": 1, "singleCoreBatch": 1, "singleCoreM": 1024, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 1, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 1, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": true, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 16777216, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 2 | +{"id": 0, "info_dict": {"batch_n": 64, "groups": 1, "dedx_d": 1, "dedx_cin": 32, "dedx_h": 64, "dedx_w": 64, "dedy_d": 1, "dedy_cout": 128, "dedy_h": 32, "dedy_w": 32, "kernel_d": 1, "kernel_h": 5, "kernel_w": 5, "stride_d": 1, "stride_h": 2, "stride_w": 2, "pad_h": 0, "pad_t": 0, "pad_u": 1, "pad_d": 2, "pad_l": 1, "pad_r": 2, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 0, "backprop_pad_t": 0, "backprop_pad_u": 3, "backprop_pad_d": 1, "backprop_pad_l": 3, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 0, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 32, "coutG": 128, "cout1": 16, "cin1": 4, "cout1G": 16, "cin1G": 2, "backpropPadTail": 0, "backpropPadUp": 3, "backpropPadDown": 2, "backpropPadLeft": 3, "backpropPadRight": 2, "singleCoreGroup": 1, "singleCoreCout": 128, "singleCoreCin": 32, "singleCoreDin": 1, "baseM": 1024, "baseK": 8, "baseN": 32, "stepKa": 54, "stepKb": 54, "singleIterateDk": 1, "singleCoreBatch": 1, "singleCoreM": 1024, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 1, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 1, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": true, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 16777216, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 3 | -{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 6, "dedx_cin": 768, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 384, "cout1": 48, "cin1": 96, "cout1G": 48, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 96, "baseK": 72, "baseN": 96, "stepKa": 24, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 3 | +{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 6, "dedx_cin": 768, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 384, "cout1": 48, "cin1": 96, "cout1G": 48, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 96, "baseK": 72, "baseN": 96, "stepKa": 24, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 4 | -{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 32, "baseK": 72, "baseN": 96, "stepKa": 48, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 4 | +{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 32, "baseK": 72, "baseN": 96, "stepKa": 48, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 5 | -{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 32, "baseK": 72, "baseN": 96, "stepKa": 48, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 5 | +{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 32, "baseK": 72, "baseN": 96, "stepKa": 48, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 6 | -{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 6, "dedx_cin": 768, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 384, "cout1": 48, "cin1": 96, "cout1G": 48, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 64, "singleCoreDin": 1, "baseM": 48, "baseK": 120, "baseN": 64, "stepKa": 6, "stepKb": 6, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 6 | +{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 6, "dedx_cin": 768, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 384, "cout1": 48, "cin1": 96, "cout1G": 48, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 64, "singleCoreDin": 1, "baseM": 48, "baseK": 120, "baseN": 64, "stepKa": 6, "stepKb": 6, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 7 | -{"id":0,"info_dict":{"batch_n":1,"groups":1,"dedx_d":1,"dedx_cin":1,"dedx_h":131073,"dedx_w":1,"dedy_d":1,"dedy_cout":257,"dedy_h":131083,"dedy_w":11,"kernel_d":1,"kernel_h":1,"kernel_w":1,"stride_d":1,"stride_h":1,"stride_w":1,"pad_h":0,"pad_t":0,"pad_u":5,"pad_d":5,"pad_l":5,"pad_r":5,"dilation_d":1,"dilation_h":1,"dilation_w":1,"backprop_pad_h":0,"backprop_pad_t":0,"backprop_pad_u":-5,"backprop_pad_d":-5,"backprop_pad_l":-5,"backprop_pad_r":-5,"hf32_flag":0,"a_dtype_bytes":2,"b_dtype_bytes":2,"c_dtype_bytes":2,"outBackpropFormat":30,"filterFormat":30,"yFormat":30},"knowledge":{"al0Pbuffer":2,"bl0Pbuffer":2,"cl0Pbuffer":2,"al1Pbuffer":2,"bl1Pbuffer":2,"iterateOrder":1,"c0":16,"c0BitsA":4,"c0BitsB":4,"enlarge":1,"initOutputFlag":0,"isBiasFullLoad":1,"enableVecTrans":0,"enableFullLoad":0,"quantMode":0,"cinG":1,"coutG":257,"cout1":17,"cin1":1,"cout1G":17,"cin1G":1,"backpropPadTail":0,"backpropPadUp":-5,"backpropPadDown":-5,"backpropPadLeft":-5,"backpropPadRight":-5,"singleCoreGroup":1,"singleCoreCout":257,"singleCoreCin":16,"singleCoreDin":1,"baseM":512,"baseK":16,"baseN":16,"stepKa":1,"stepKb":9,"singleIterateDk":1,"singleCoreBatch":1,"singleCoreM":512,"enRelu":0,"coreNum":32,"kSCoutFullLoad":0,"kSUseWorkSpace":0,"loadB2Condition":3,"loadB1Condition":0,"kernelSplitMode":0,"enableC04Flag":false,"enableFullLoadTiling":false,"enableVecTransFlag":false,"enableSplitKernelFlag":false,"tilingHkWkMode":0,"usrSpaceSizeForKernelSplit":0,"kSegment":1,"kSegmentTail":1,"kValueSegment":16,"enableSplitK":true,"useUbAccumForSplitK":true,"reservedParam1":0,"reservedParam2":0,"reservedParam3":0,"reservedParam4":0,"reservedParam5":0,"reservedParam6":0,"reservedParam7":0,"reservedParam8":0},"op":"Conv3DBackpropInputV2","version":0} | 7 | +{"id":0,"info_dict":{"batch_n":1,"groups":1,"dedx_d":1,"dedx_cin":1,"dedx_h":131073,"dedx_w":1,"dedy_d":1,"dedy_cout":257,"dedy_h":131083,"dedy_w":11,"kernel_d":1,"kernel_h":1,"kernel_w":1,"stride_d":1,"stride_h":1,"stride_w":1,"pad_h":0,"pad_t":0,"pad_u":5,"pad_d":5,"pad_l":5,"pad_r":5,"dilation_d":1,"dilation_h":1,"dilation_w":1,"backprop_pad_h":0,"backprop_pad_t":0,"backprop_pad_u":-5,"backprop_pad_d":-5,"backprop_pad_l":-5,"backprop_pad_r":-5,"hf32_flag":0,"a_dtype_bytes":2,"b_dtype_bytes":2,"c_dtype_bytes":2,"outBackpropFormat":30,"filterFormat":30,"yFormat":30},"knowledge":{"al0Pbuffer":2,"bl0Pbuffer":2,"cl0Pbuffer":2,"al1Pbuffer":2,"bl1Pbuffer":2,"iterateOrder":1,"c0":16,"c0BitsA":4,"c0BitsB":4,"enlarge":1,"initOutputFlag":0,"isBiasFullLoad":1,"enableVecTrans":0,"enableFullLoad":0,"quantMode":0,"cinG":1,"coutG":257,"cout1":17,"cin1":1,"cout1G":17,"cin1G":1,"backpropPadTail":0,"backpropPadUp":-5,"backpropPadDown":-5,"backpropPadLeft":-5,"backpropPadRight":-5,"singleCoreGroup":1,"singleCoreCout":257,"singleCoreCin":16,"singleCoreDin":1,"baseM":512,"baseK":16,"baseN":16,"stepKa":1,"stepKb":9,"singleIterateDk":1,"singleCoreBatch":1,"singleCoreM":512,"enRelu":0,"coreNum":32,"kSCoutFullLoad":0,"kSUseWorkSpace":0,"loadB2Condition":3,"loadB1Condition":0,"kernelSplitMode":0,"enableC04Flag":false,"enableFullLoadTiling":false,"enableVecTransFlag":false,"enableSplitKernelFlag":false,"tilingHkWkMode":0,"usrSpaceSizeForKernelSplit":0,"kSegment":1,"kSegmentTail":1,"kValueSegment":16,"enableSplitK":true,"useUbAccumForSplitK":true,"reservedParam1":0,"reservedParam2":0,"reservedParam3":0,"reservedParam4":0,"reservedParam5":0,"reservedParam6":0,"reservedParam7":0,"reservedParam8":0},"op":"Conv3DBackpropInputV2","version":0} |
| @@ -1,7 +1,7 @@ | |||
| 1 | -{"id": 0, "info_dict": {"batch_n": 64, "groups": 1, "dedx_d": 1, "dedx_cin": 128, "dedx_h": 32, "dedx_w": 32, "dedy_d": 1, "dedy_cout": 256, "dedy_h": 16, "dedy_w": 16, "kernel_d": 1, "kernel_h": 5, "kernel_w": 5, "stride_d": 1, "stride_h": 2, "stride_w": 2, "pad_h": 0, "pad_t": 0, "pad_u": 1, "pad_d": 2, "pad_l": 1, "pad_r": 2, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 0, "backprop_pad_t": 0, "backprop_pad_u": 3, "backprop_pad_d": 1, "backprop_pad_l": 3, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 0, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 128, "coutG": 256, "cout1": 32, "cin1": 16, "cout1G": 32, "cin1G": 8, "backpropPadTail": 0, "backpropPadUp": 3, "backpropPadDown": 2, "backpropPadLeft": 3, "backpropPadRight": 2, "singleCoreGroup": 1, "singleCoreCout": 256, "singleCoreCin": 128, "singleCoreDin": 1, "baseM": 256, "baseK": 8, "baseN": 128, "stepKa": 45, "stepKb": 45, "singleIterateDk": 1, "singleCoreBatch": 1, "singleCoreM": 256, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 1, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 1, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": true, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 16777216, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 1 | +{"id": 0, "info_dict": {"batch_n": 64, "groups": 1, "dedx_d": 1, "dedx_cin": 128, "dedx_h": 32, "dedx_w": 32, "dedy_d": 1, "dedy_cout": 256, "dedy_h": 16, "dedy_w": 16, "kernel_d": 1, "kernel_h": 5, "kernel_w": 5, "stride_d": 1, "stride_h": 2, "stride_w": 2, "pad_h": 0, "pad_t": 0, "pad_u": 1, "pad_d": 2, "pad_l": 1, "pad_r": 2, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 0, "backprop_pad_t": 0, "backprop_pad_u": 3, "backprop_pad_d": 1, "backprop_pad_l": 3, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 0, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 128, "coutG": 256, "cout1": 32, "cin1": 16, "cout1G": 32, "cin1G": 8, "backpropPadTail": 0, "backpropPadUp": 3, "backpropPadDown": 2, "backpropPadLeft": 3, "backpropPadRight": 2, "singleCoreGroup": 1, "singleCoreCout": 256, "singleCoreCin": 128, "singleCoreDin": 1, "baseM": 256, "baseK": 8, "baseN": 128, "stepKa": 45, "stepKb": 45, "singleIterateDk": 1, "singleCoreBatch": 1, "singleCoreM": 256, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 1, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 1, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": true, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 16777216, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 2 | -{"id": 0, "info_dict": {"batch_n": 64, "groups": 1, "dedx_d": 1, "dedx_cin": 32, "dedx_h": 64, "dedx_w": 64, "dedy_d": 1, "dedy_cout": 128, "dedy_h": 32, "dedy_w": 32, "kernel_d": 1, "kernel_h": 5, "kernel_w": 5, "stride_d": 1, "stride_h": 2, "stride_w": 2, "pad_h": 0, "pad_t": 0, "pad_u": 1, "pad_d": 2, "pad_l": 1, "pad_r": 2, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 0, "backprop_pad_t": 0, "backprop_pad_u": 3, "backprop_pad_d": 1, "backprop_pad_l": 3, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 0, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 32, "coutG": 128, "cout1": 16, "cin1": 4, "cout1G": 16, "cin1G": 2, "backpropPadTail": 0, "backpropPadUp": 3, "backpropPadDown": 2, "backpropPadLeft": 3, "backpropPadRight": 2, "singleCoreGroup": 1, "singleCoreCout": 128, "singleCoreCin": 32, "singleCoreDin": 1, "baseM": 1024, "baseK": 8, "baseN": 32, "stepKa": 54, "stepKb": 54, "singleIterateDk": 1, "singleCoreBatch": 1, "singleCoreM": 1024, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 1, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 1, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": true, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 16777216, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 2 | +{"id": 0, "info_dict": {"batch_n": 64, "groups": 1, "dedx_d": 1, "dedx_cin": 32, "dedx_h": 64, "dedx_w": 64, "dedy_d": 1, "dedy_cout": 128, "dedy_h": 32, "dedy_w": 32, "kernel_d": 1, "kernel_h": 5, "kernel_w": 5, "stride_d": 1, "stride_h": 2, "stride_w": 2, "pad_h": 0, "pad_t": 0, "pad_u": 1, "pad_d": 2, "pad_l": 1, "pad_r": 2, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 0, "backprop_pad_t": 0, "backprop_pad_u": 3, "backprop_pad_d": 1, "backprop_pad_l": 3, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 0, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 32, "coutG": 128, "cout1": 16, "cin1": 4, "cout1G": 16, "cin1G": 2, "backpropPadTail": 0, "backpropPadUp": 3, "backpropPadDown": 2, "backpropPadLeft": 3, "backpropPadRight": 2, "singleCoreGroup": 1, "singleCoreCout": 128, "singleCoreCin": 32, "singleCoreDin": 1, "baseM": 1024, "baseK": 8, "baseN": 32, "stepKa": 54, "stepKb": 54, "singleIterateDk": 1, "singleCoreBatch": 1, "singleCoreM": 1024, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 1, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 1, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": true, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 16777216, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 3 | -{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 6, "dedx_cin": 768, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 384, "cout1": 48, "cin1": 96, "cout1G": 48, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 96, "baseK": 72, "baseN": 96, "stepKa": 24, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 3 | +{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 6, "dedx_cin": 768, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 384, "cout1": 48, "cin1": 96, "cout1G": 48, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 96, "baseK": 72, "baseN": 96, "stepKa": 24, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 4 | -{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 32, "baseK": 72, "baseN": 96, "stepKa": 48, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 4 | +{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 32, "baseK": 72, "baseN": 96, "stepKa": 48, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 5 | -{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 32, "baseK": 72, "baseN": 96, "stepKa": 48, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 5 | +{"id": 0, "info_dict": {"batch_n": 4, "groups": 1, "dedx_d": 3, "dedx_cin": 768, "dedx_h": 3, "dedx_w": 3, "dedy_d": 3, "dedy_cout": 768, "dedy_h": 3, "dedy_w": 3, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 27, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 0, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 768, "cout1": 96, "cin1": 96, "cout1G": 96, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 768, "singleCoreCin": 96, "singleCoreDin": 1, "baseM": 32, "baseK": 72, "baseN": 96, "stepKa": 48, "stepKb": 4, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 15, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": false, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 6 | -{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 6, "dedx_cin": 768, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 384, "cout1": 48, "cin1": 96, "cout1G": 48, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 64, "singleCoreDin": 1, "baseM": 48, "baseK": 120, "baseN": 64, "stepKa": 6, "stepKb": 6, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} | 6 | +{"id": 0, "info_dict": {"batch_n": 8, "groups": 1, "dedx_d": 6, "dedx_cin": 768, "dedx_h": 6, "dedx_w": 6, "dedy_d": 6, "dedy_cout": 384, "dedy_h": 6, "dedy_w": 6, "kernel_d": 3, "kernel_h": 3, "kernel_w": 3, "stride_d": 1, "stride_h": 1, "stride_w": 1, "pad_h": 1, "pad_t": 1, "pad_u": 1, "pad_d": 1, "pad_l": 1, "pad_r": 1, "dilation_d": 1, "dilation_h": 1, "dilation_w": 1, "backprop_pad_h": 1, "backprop_pad_t": 1, "backprop_pad_u": 1, "backprop_pad_d": 1, "backprop_pad_l": 1, "backprop_pad_r": 1, "hf32_flag": 1, "a_dtype_bytes": 4, "b_dtype_bytes": 4, "c_dtype_bytes": 4, "outBackpropFormat": 30, "filterFormat": 30, "yFormat": 30}, "knowledge": {"al0Pbuffer": 2, "bl0Pbuffer": 2, "cl0Pbuffer": 2, "al1Pbuffer": 2, "bl1Pbuffer": 2, "iterateOrder": 1, "c0": 8, "c0BitsA": 3, "c0BitsB": 3, "enlarge": 1, "initOutputFlag": 0, "isBiasFullLoad": 1, "enableVecTrans": 1, "enableFullLoad": 0, "quantMode": 0, "cinG": 768, "coutG": 384, "cout1": 48, "cin1": 96, "cout1G": 48, "cin1G": 48, "backpropPadTail": 1, "backpropPadUp": 1, "backpropPadDown": 1, "backpropPadLeft": 1, "backpropPadRight": 1, "singleCoreGroup": 1, "singleCoreCout": 384, "singleCoreCin": 64, "singleCoreDin": 1, "baseM": 48, "baseK": 120, "baseN": 64, "stepKa": 6, "stepKb": 6, "singleIterateDk": 3, "singleCoreBatch": 1, "singleCoreM": 48, "enRelu": 0, "coreNum": 32, "kSCoutFullLoad": 0, "kSUseWorkSpace": 0, "loadB2Condition": 2, "loadB1Condition": 0, "kernelSplitMode": 0, "enableC04Flag": false, "enableFullLoadTiling": false, "enableVecTransFlag": true, "enableSplitKernelFlag": false, "tilingHkWkMode": 0, "usrSpaceSizeForKernelSplit": 0, "kSegment":0,"kSegmentTail":0,"kValueSegment":0,"enableSplitK":false,"useUbAccumForSplitK":false,"reservedParam1": 0, "reservedParam2": 0, "reservedParam3": 0, "reservedParam4": 0, "reservedParam5": 0, "reservedParam6": 0, "reservedParam7": 0, "reservedParam8": 0}, "op": "Conv3DBackpropInputV2", "version": 0} |
| 7 | -{"id":0,"info_dict":{"batch_n":1,"groups":1,"dedx_d":1,"dedx_cin":1,"dedx_h":131073,"dedx_w":1,"dedy_d":1,"dedy_cout":257,"dedy_h":131083,"dedy_w":11,"kernel_d":1,"kernel_h":1,"kernel_w":1,"stride_d":1,"stride_h":1,"stride_w":1,"pad_h":0,"pad_t":0,"pad_u":5,"pad_d":5,"pad_l":5,"pad_r":5,"dilation_d":1,"dilation_h":1,"dilation_w":1,"backprop_pad_h":0,"backprop_pad_t":0,"backprop_pad_u":-5,"backprop_pad_d":-5,"backprop_pad_l":-5,"backprop_pad_r":-5,"hf32_flag":0,"a_dtype_bytes":2,"b_dtype_bytes":2,"c_dtype_bytes":2,"outBackpropFormat":30,"filterFormat":30,"yFormat":30},"knowledge":{"al0Pbuffer":2,"bl0Pbuffer":2,"cl0Pbuffer":2,"al1Pbuffer":2,"bl1Pbuffer":2,"iterateOrder":1,"c0":16,"c0BitsA":4,"c0BitsB":4,"enlarge":1,"initOutputFlag":0,"isBiasFullLoad":1,"enableVecTrans":0,"enableFullLoad":0,"quantMode":0,"cinG":1,"coutG":257,"cout1":17,"cin1":1,"cout1G":17,"cin1G":1,"backpropPadTail":0,"backpropPadUp":-5,"backpropPadDown":-5,"backpropPadLeft":-5,"backpropPadRight":-5,"singleCoreGroup":1,"singleCoreCout":257,"singleCoreCin":16,"singleCoreDin":1,"baseM":512,"baseK":16,"baseN":16,"stepKa":1,"stepKb":9,"singleIterateDk":1,"singleCoreBatch":1,"singleCoreM":512,"enRelu":0,"coreNum":32,"kSCoutFullLoad":0,"kSUseWorkSpace":0,"loadB2Condition":3,"loadB1Condition":0,"kernelSplitMode":0,"enableC04Flag":false,"enableFullLoadTiling":false,"enableVecTransFlag":false,"enableSplitKernelFlag":false,"tilingHkWkMode":0,"usrSpaceSizeForKernelSplit":0,"kSegment":1,"kSegmentTail":1,"kValueSegment":16,"enableSplitK":true,"useUbAccumForSplitK":true,"reservedParam1":0,"reservedParam2":0,"reservedParam3":0,"reservedParam4":0,"reservedParam5":0,"reservedParam6":0,"reservedParam7":0,"reservedParam8":0},"op":"Conv3DBackpropInputV2","version":0} | 7 | +{"id":0,"info_dict":{"batch_n":1,"groups":1,"dedx_d":1,"dedx_cin":1,"dedx_h":131073,"dedx_w":1,"dedy_d":1,"dedy_cout":257,"dedy_h":131083,"dedy_w":11,"kernel_d":1,"kernel_h":1,"kernel_w":1,"stride_d":1,"stride_h":1,"stride_w":1,"pad_h":0,"pad_t":0,"pad_u":5,"pad_d":5,"pad_l":5,"pad_r":5,"dilation_d":1,"dilation_h":1,"dilation_w":1,"backprop_pad_h":0,"backprop_pad_t":0,"backprop_pad_u":-5,"backprop_pad_d":-5,"backprop_pad_l":-5,"backprop_pad_r":-5,"hf32_flag":0,"a_dtype_bytes":2,"b_dtype_bytes":2,"c_dtype_bytes":2,"outBackpropFormat":30,"filterFormat":30,"yFormat":30},"knowledge":{"al0Pbuffer":2,"bl0Pbuffer":2,"cl0Pbuffer":2,"al1Pbuffer":2,"bl1Pbuffer":2,"iterateOrder":1,"c0":16,"c0BitsA":4,"c0BitsB":4,"enlarge":1,"initOutputFlag":0,"isBiasFullLoad":1,"enableVecTrans":0,"enableFullLoad":0,"quantMode":0,"cinG":1,"coutG":257,"cout1":17,"cin1":1,"cout1G":17,"cin1G":1,"backpropPadTail":0,"backpropPadUp":-5,"backpropPadDown":-5,"backpropPadLeft":-5,"backpropPadRight":-5,"singleCoreGroup":1,"singleCoreCout":257,"singleCoreCin":16,"singleCoreDin":1,"baseM":512,"baseK":16,"baseN":16,"stepKa":1,"stepKb":9,"singleIterateDk":1,"singleCoreBatch":1,"singleCoreM":512,"enRelu":0,"coreNum":32,"kSCoutFullLoad":0,"kSUseWorkSpace":0,"loadB2Condition":3,"loadB1Condition":0,"kernelSplitMode":0,"enableC04Flag":false,"enableFullLoadTiling":false,"enableVecTransFlag":false,"enableSplitKernelFlag":false,"tilingHkWkMode":0,"usrSpaceSizeForKernelSplit":0,"kSegment":1,"kSegmentTail":1,"kValueSegment":16,"enableSplitK":true,"useUbAccumForSplitK":true,"reservedParam1":0,"reservedParam2":0,"reservedParam3":0,"reservedParam4":0,"reservedParam5":0,"reservedParam6":0,"reservedParam7":0,"reservedParam8":0},"op":"Conv3DBackpropInputV2","version":0} |