Facebook
From Mathilda, 1 Month ago, written in Plain Text.
Embed
Download Paste or View Raw
Hits: 83
  1. APP_SRCS... main.c BUILD_MODEL_SQ8BIT/networkKernels.c /gap_sdk/libs/gap_lib/img_io/ImgIO.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.c /gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.c /gap_sdk/tools/nntool//autotiler/kernels/copy.c
  2. APP_CFLAGS... -g -O3 -w -mno-memcpy -fno-tree-loop-distribute-patterns -I. -I/gap_sdk/libs/gap_lib/include -I/gap_sdk/tools/autotiler_v3/Emulation -I/gap_sdk/tools/autotiler_v3/Autotiler -I/gap_sdk/tools/autotiler_v3/CNN_Libraries -I/gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8 -I/gap_sdk/tools/nntool//autotiler/kernels -IBUILD_MODEL_SQ8BIT -DAT_MODEL_PREFIX=network -DAT_INPUT_HEIGHT=200 -DAT_INPUT_WIDTH=200 -DAT_INPUT_COLORS=1 -DSTACK_SIZE=6096 -DSLAVE_STACK_SIZE=1024 -DAT_IMAGE=/module/data/images/frame_2.pgm -DPERF -DMODEL_ID= -DFREQ_FC=250 -DFREQ_CL=175 -DAT_CONSTRUCT=networkCNN_Construct -DAT_DESTRUCT=networkCNN_Destruct -DAT_CNN=networkCNN -DAT_L3_ADDR=network_L3_Flash
  3. common/model_rules.mk:28: warning: overriding recipe for target 'BUILD_MODEL_SQ8BIT'
  4. common/model_rules.mk:28: warning: ignoring old recipe for target 'BUILD_MODEL_SQ8BIT'
  5. common/model_rules.mk:31: warning: overriding recipe for target 'BUILD_MODEL_SQ8BIT/network.onnx'
  6. common/model_rules.mk:31: warning: ignoring old recipe for target 'BUILD_MODEL_SQ8BIT/network.onnx'
  7. common/model_rules.mk:35: warning: overriding recipe for target 'BUILD_MODEL_SQ8BIT/network.json'
  8. common/model_rules.mk:35: warning: ignoring old recipe for target 'BUILD_MODEL_SQ8BIT/network.json'
  9. common/model_rules.mk:40: warning: overriding recipe for target 'nntool_model_evaluation'
  10. common/model_rules.mk:40: warning: ignoring old recipe for target 'nntool_model_evaluation'
  11. common/model_rules.mk:48: warning: overriding recipe for target 'nntool_output/networkModel.c'
  12. common/model_rules.mk:48: warning: ignoring old recipe for target 'nntool_output/networkModel.c'
  13. common/model_rules.mk:55: warning: overriding recipe for target 'BUILD_MODEL_SQ8BIT/GenTile'
  14. common/model_rules.mk:55: warning: ignoring old recipe for target 'BUILD_MODEL_SQ8BIT/GenTile'
  15. common/model_rules.mk:62: warning: overriding recipe for target 'BUILD_MODEL_SQ8BIT/networkKernels.c'
  16. common/model_rules.mk:62: warning: ignoring old recipe for target 'BUILD_MODEL_SQ8BIT/networkKernels.c'
  17. common/model_rules.mk:69: warning: overriding recipe for target 'clean_model'
  18. common/model_rules.mk:69: warning: ignoring old recipe for target 'clean_model'
  19. APP_SRCS... main.c BUILD_MODEL_SQ8BIT/networkKernels.c /gap_sdk/libs/gap_lib/img_io/ImgIO.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.c /gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.c /gap_sdk/tools/nntool//autotiler/kernels/copy.c  /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.c /gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.c /gap_sdk/tools/nntool//autotiler/kernels/copy.c   main.c BUILD_MODEL_SQ8BIT/networkKernels.c /gap_sdk/libs/gap_lib/img_io/ImgIO.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.c /gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.c /gap_sdk/tools/nntool//autotiler/kernels/copy.c  /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.c /gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.c /gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.c /gap_sdk/tools/nntool//autotiler/kernels/copy.c
  20. APP_CFLAGS... -g -O3 -w -mno-memcpy -fno-tree-loop-distribute-patterns -I. -I/gap_sdk/libs/gap_lib/include -I/gap_sdk/tools/autotiler_v3/Emulation -I/gap_sdk/tools/autotiler_v3/Autotiler -I/gap_sdk/tools/autotiler_v3/CNN_Libraries -I/gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8 -I/gap_sdk/tools/nntool//autotiler/kernels -IBUILD_MODEL_SQ8BIT -DAT_MODEL_PREFIX=network -DAT_INPUT_HEIGHT=200 -DAT_INPUT_WIDTH=200 -DAT_INPUT_COLORS=1 -DSTACK_SIZE=6096 -DSLAVE_STACK_SIZE=1024 -DAT_IMAGE=/module/data/images/frame_2.pgm -DPERF -DMODEL_ID= -DFREQ_FC=250 -DFREQ_CL=175 -DAT_CONSTRUCT=networkCNN_Construct -DAT_DESTRUCT=networkCNN_Destruct -DAT_CNN=networkCNN -DAT_L3_ADDR=network_L3_Flash -g -O3 -w -mno-memcpy -fno-tree-loop-distribute-patterns -I. -I/gap_sdk/libs/gap_lib/include -I/gap_sdk/tools/autotiler_v3/Emulation -I/gap_sdk/tools/autotiler_v3/Autotiler -I/gap_sdk/tools/autotiler_v3/CNN_Libraries -I/gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8 -I/gap_sdk/tools/nntool//autotiler/kernels -IBUILD_MODEL_SQ8BIT -DAT_MODEL_PREFIX=network -DAT_INPUT_HEIGHT=200 -DAT_INPUT_WIDTH=200 -DAT_INPUT_COLORS=1 -DSTACK_SIZE=6096 -DSLAVE_STACK_SIZE=1024 -DAT_IMAGE=/module/data/images/frame_2.pgm -DPERF -DMODEL_ID= -DFREQ_FC=250 -DFREQ_CL=175 -DAT_CONSTRUCT=networkCNN_Construct -DAT_DESTRUCT=networkCNN_Destruct -DAT_CNN=networkCNN -DAT_L3_ADDR=network_L3_Flash
  21. /gap_sdk/tools/rules/pulp_rules.mk:199: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV'
  22. /gap_sdk/tools/rules/pulp_rules.mk:199: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV'
  23. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/main.o'
  24. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/main.o'
  25. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/BUILD_MODEL_SQ8BIT/networkKernels.o'
  26. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/BUILD_MODEL_SQ8BIT/networkKernels.o'
  27. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/libs/gap_lib/img_io/ImgIO.o'
  28. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/libs/gap_lib/img_io/ImgIO.o'
  29. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.o'
  30. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.o'
  31. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.o'
  32. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.o'
  33. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.o'
  34. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.o'
  35. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.o'
  36. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.o'
  37. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.o'
  38. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.o'
  39. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.o'
  40. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.o'
  41. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.o'
  42. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.o'
  43. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.o'
  44. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.o'
  45. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.o'
  46. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.o'
  47. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.o'
  48. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.o'
  49. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/copy.o'
  50. /gap_sdk/tools/rules/pulp_rules.mk:204: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/copy.o'
  51. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.o' given more than once in the same rule
  52. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.o' given more than once in the same rule
  53. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.o' given more than once in the same rule
  54. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.o' given more than once in the same rule
  55. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.o' given more than once in the same rule
  56. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.o' given more than once in the same rule
  57. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.o' given more than once in the same rule
  58. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.o' given more than once in the same rule
  59. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.o' given more than once in the same rule
  60. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.o' given more than once in the same rule
  61. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/copy.o' given more than once in the same rule
  62. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV/main.o' given more than once in the same rule
  63. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV/BUILD_MODEL_SQ8BIT/networkKernels.o' given more than once in the same rule
  64. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/libs/gap_lib/img_io/ImgIO.o' given more than once in the same rule
  65. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.o' given more than once in the same rule
  66. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.o' given more than once in the same rule
  67. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.o' given more than once in the same rule
  68. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.o' given more than once in the same rule
  69. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.o' given more than once in the same rule
  70. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.o' given more than once in the same rule
  71. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.o' given more than once in the same rule
  72. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.o' given more than once in the same rule
  73. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.o' given more than once in the same rule
  74. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.o' given more than once in the same rule
  75. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/copy.o' given more than once in the same rule
  76. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Activation_SQ8.o' given more than once in the same rule
  77. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Bias_Linear_SQ8.o' given more than once in the same rule
  78. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_SQ8.o' given more than once in the same rule
  79. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Pooling_SQ8.o' given more than once in the same rule
  80. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_Conv_DW_SQ8.o' given more than once in the same rule
  81. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_MatAlgebra_SQ8.o' given more than once in the same rule
  82. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_SoftMax_SQ8.o' given more than once in the same rule
  83. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/CNN_AT_Misc.o' given more than once in the same rule
  84. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8/RNN_SQ8.o' given more than once in the same rule
  85. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/norm_transpose.o' given more than once in the same rule
  86. /gap_sdk/tools/rules/pulp_rules.mk:203: target '/module/data/BUILD/GAP8_V2/GCC_RISCV//gap_sdk/tools/nntool//autotiler/kernels/copy.o' given more than once in the same rule
  87. /gap_sdk/tools/rules/pulp_rules.mk:208: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/pulp-os/conf.o'
  88. /gap_sdk/tools/rules/pulp_rules.mk:208: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/pulp-os/conf.o'
  89. /gap_sdk/tools/rules/pulp_rules.mk:212: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/imagenet'
  90. /gap_sdk/tools/rules/pulp_rules.mk:212: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/imagenet'
  91. /gap_sdk/tools/rules/pulp_rules.mk:229: warning: overriding recipe for target 'flash'
  92. /gap_sdk/tools/rules/pulp_rules.mk:229: warning: ignoring old recipe for target 'flash'
  93. /gap_sdk/tools/rules/pulp_rules.mk:232: warning: overriding recipe for target 'flash_fs'
  94. /gap_sdk/tools/rules/pulp_rules.mk:232: warning: ignoring old recipe for target 'flash_fs'
  95. /gap_sdk/tools/rules/pulp_rules.mk:235: warning: overriding recipe for target 'image'
  96. /gap_sdk/tools/rules/pulp_rules.mk:235: warning: ignoring old recipe for target 'image'
  97. /gap_sdk/tools/rules/pulp_rules.mk:238: warning: overriding recipe for target 'run.prepare'
  98. /gap_sdk/tools/rules/pulp_rules.mk:238: warning: ignoring old recipe for target 'run.prepare'
  99. /gap_sdk/tools/rules/pulp_rules.mk:241: warning: overriding recipe for target 'run.exec'
  100. /gap_sdk/tools/rules/pulp_rules.mk:241: warning: ignoring old recipe for target 'run.exec'
  101. /gap_sdk/tools/rules/pulp_rules.mk:244: warning: overriding recipe for target 'run'
  102. /gap_sdk/tools/rules/pulp_rules.mk:244: warning: ignoring old recipe for target 'run'
  103. /gap_sdk/tools/rules/pulp_rules.mk:248: warning: overriding recipe for target 'profiler'
  104. /gap_sdk/tools/rules/pulp_rules.mk:248: warning: ignoring old recipe for target 'profiler'
  105. /gap_sdk/tools/rules/pulp_rules.mk:255: warning: overriding recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/imagenet.s'
  106. /gap_sdk/tools/rules/pulp_rules.mk:255: warning: ignoring old recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/imagenet.s'
  107. /gap_sdk/tools/rules/pulp_rules.mk:262: warning: overriding recipe for target 'version'
  108. /gap_sdk/tools/rules/pulp_rules.mk:262: warning: ignoring old recipe for target 'version'
  109. rm -f BUILD_MODEL_SQ8BIT/GenTile
  110. rm -f -rf BUILD_MODEL_SQ8BIT
  111. rm -f -rf nntool_output
  112. mkdir BUILD_MODEL_SQ8BIT
  113. cp nntool_input/models_onnx/model_original_himax.onnx BUILD_MODEL_SQ8BIT/network.onnx
  114. echo "GENERATING NNTOOL STATE FILE"
  115. GENERATING NNTOOL STATE FILE
  116. echo BUILD_MODEL_SQ8BIT
  117. BUILD_MODEL_SQ8BIT
  118. nntool -s nntool_input/nntool_scripts/nntool_script_deployment BUILD_MODEL_SQ8BIT/network.onnx -q
  119. -s is not a recognized command, alias, or macro
  120. nntool_input/nntool_scripts/nntool_script_deployment is not a recognized command, alias, or macro
  121. BUILD_MODEL_SQ8BIT/network.onnx is not a recognized command, alias, or macro
  122. -q is not a recognized command, alias, or macro
  123. open - opening graph file BUILD_MODEL_SQ8BIT/network.onnx
  124. onnx - unable to determine batch dimension. if the graph fails to import properly set it to 1 or a variable.
  125. nngraph - update graph dimensions
  126. set_aliases - looking for aliased edges
  127. nngraph - calculate liveness
  128. debug - was: False
  129. now: True
  130. adjust_order - adding transposes to correct tensor order for AT kernels
  131. set_aliases - looking for aliased edges
  132. eliminate_transposes - eliminating unnecessary transposes
  133. eliminate_transposes - search for transposes
  134. eliminate_transposes - no transposes to eliminate found
  135. nngraph - update graph dimensions
  136. set_aliases - looking for aliased edges
  137. nngraph - calculate liveness
  138. eliminate_transposes - no further transpose sequences found
  139. set_aliases - looking for aliased edges
  140. nngraph - adjusted order
  141. nngraph - update graph dimensions
  142. set_aliases - looking for aliased edges
  143. nngraph - calculate liveness
  144. matcher - fusions - start remove_relus
  145. matcher - fusions - start remove_noops
  146. matcher - fusions - start fuse_external_bias_sq8
  147. matcher - fusions - start fuse_pad
  148. matcher - fusions - start unused_concats
  149. matcher - fusions - start gather_to_split
  150. gather_to_split - gathers from Gemm_31[0] converted to a split
  151. matcher - fusions - gather_to_split modified graph
  152. nngraph - update graph dimensions
  153. set_aliases - looking for aliased edges
  154. nngraph - calculate liveness
  155. matcher - fusions - start find_missing_quantization
  156. matcher - fusions - start rnn_reverse
  157. nngraph - update graph dimensions
  158. set_aliases - looking for aliased edges
  159. nngraph - calculate liveness
  160. matcher - fusions - start rnn_unpack
  161. matcher - fusions - start match_far_hsigmoid
  162. matcher - fusions - start match_close_hsigmoid
  163. matcher - fusions - start expand_transposes
  164. matcher - fusions - start move_pooling_scale8
  165. matcher - fusions - start move_activations_scale8
  166. move_node_up - Node Sigmoid_36 cannot be moved
  167. matcher - fusions - start fuse_gap_convs
  168. matcher - fusions - start match_conv_active_pool
  169. matcher - fusions - start match_conv_pool_active
  170. matcher - fusions - start match_conv_active
  171. match_gap_conv - fusing nodes Conv_2,Clip_4
  172. match_gap_conv - fusing nodes Conv_5,Clip_7
  173. match_gap_conv - fusing nodes Conv_8,Clip_9
  174. match_gap_conv - fusing nodes Conv_11,Clip_13
  175. match_gap_conv - fusing nodes Conv_14,Clip_16
  176. match_gap_conv - fusing nodes Conv_17,Clip_18
  177. match_gap_conv - fusing nodes Conv_20,Clip_22
  178. match_gap_conv - fusing nodes Conv_23,Clip_25
  179. match_gap_conv - fusing nodes Conv_26,Clip_27
  180. matcher - fusions - match_conv_active modified graph
  181. nngraph - update graph dimensions
  182. set_aliases - looking for aliased edges
  183. nngraph - calculate liveness
  184. matcher - fusions - start match_conv_pool
  185. match_gap_conv - fusing nodes Conv_0,MaxPool_1
  186. matcher - fusions - match_conv_pool modified graph
  187. nngraph - update graph dimensions
  188. set_aliases - looking for aliased edges
  189. nngraph - calculate liveness
  190. matcher - fusions - start fuse_gap_linear
  191. matcher - fusions - start fuse_op_activation_scale8
  192. matcher - fusions - fuse_op_activation_scale8 modified graph
  193. nngraph - update graph dimensions
  194. set_aliases - looking for aliased edges
  195. nngraph - calculate liveness
  196. matcher - fusions - start propagate_softmax_sym_qrec
  197. matcher - fusions - start equalize_sm_concats
  198. matcher - fusions - start filter_bigger_than_input
  199. matcher - fusions - start insert_copies
  200. nngraph - update graph dimensions
  201. set_aliases - looking for aliased edges
  202. nngraph - calculate liveness
  203. matcher - fusions - start propagate_up_rnn_in_qs
  204. nngraph - update graph dimensions
  205. set_aliases - looking for aliased edges
  206. nngraph - calculate liveness
  207. input_norm_func - was: ''
  208. now: 'x:x/255'
  209. aquant - input file ['nntool_input/quantization_files/1479425441232704425.jpg']
  210. graph_executer - execute uncached: quantization mode none
  211. aquant - input file ['nntool_input/quantization_files/1479425443582977575.jpg']
  212. graph_executer - execute uncached: quantization mode none
  213. aquant - input file ['nntool_input/quantization_files/frame_00164.jpg']
  214. graph_executer - execute uncached: quantization mode none
  215. aquant - input file ['nntool_input/quantization_files/1479425448633847925.jpg']
  216. graph_executer - execute uncached: quantization mode none
  217. aquant - input file ['nntool_input/quantization_files/frame_00079.jpg']
  218. graph_executer - execute uncached: quantization mode none
  219. aquant - input file ['nntool_input/quantization_files/frame_00135.jpg']
  220. graph_executer - execute uncached: quantization mode none
  221. aquant - input file ['nntool_input/quantization_files/frame_00064.jpg']
  222. graph_executer - execute uncached: quantization mode none
  223. aquant - input file ['nntool_input/quantization_files/frame_00014.jpg']
  224. graph_executer - execute uncached: quantization mode none
  225. aquant - input file ['nntool_input/quantization_files/frame_00143.jpg']
  226. graph_executer - execute uncached: quantization mode none
  227. aquant - input file ['nntool_input/quantization_files/1479425443382879593.jpg']
  228. graph_executer - execute uncached: quantization mode none
  229. aquant - input file ['nntool_input/quantization_files/1479425443833012205.jpg']
  230. graph_executer - execute uncached: quantization mode none
  231. aquant - input file ['nntool_input/quantization_files/frame_00037.jpg']
  232. graph_executer - execute uncached: quantization mode none
  233. aquant - input file ['nntool_input/quantization_files/frame_00033.jpg']
  234. graph_executer - execute uncached: quantization mode none
  235. aquant - input file ['nntool_input/quantization_files/frame_00191.jpg']
  236. graph_executer - execute uncached: quantization mode none
  237. aquant - input file ['nntool_input/quantization_files/frame_00194.jpg']
  238. graph_executer - execute uncached: quantization mode none
  239. aquant - input file ['nntool_input/quantization_files/1479425443432903742.jpg']
  240. graph_executer - execute uncached: quantization mode none
  241. aquant - input file ['nntool_input/quantization_files/frame_00013.jpg']
  242. graph_executer - execute uncached: quantization mode none
  243. aquant - input file ['nntool_input/quantization_files/frame_00136.jpg']
  244. graph_executer - execute uncached: quantization mode none
  245. aquant - input file ['nntool_input/quantization_files/frame_00078.jpg']
  246. graph_executer - execute uncached: quantization mode none
  247. aquant - input file ['nntool_input/quantization_files/1479425441282730750.jpg']
  248. graph_executer - execute uncached: quantization mode none
  249. aquant - input file ['nntool_input/quantization_files/frame_00192.jpg']
  250. graph_executer - execute uncached: quantization mode none
  251. aquant - input file ['nntool_input/quantization_files/frame_00193.jpg']
  252. graph_executer - execute uncached: quantization mode none
  253. aquant - input file ['nntool_input/quantization_files/1479425443782898808.jpg']
  254. graph_executer - execute uncached: quantization mode none
  255. aquant - input file ['nntool_input/quantization_files/1479425448133784412.jpg']
  256. graph_executer - execute uncached: quantization mode none
  257. aquant - input file ['nntool_input/quantization_files/frame_00139.jpg']
  258. graph_executer - execute uncached: quantization mode none
  259. aquant - input file ['nntool_input/quantization_files/frame_00001.jpg']
  260. graph_executer - execute uncached: quantization mode none
  261. aquant - input file ['nntool_input/quantization_files/1479425451134304083.jpg']
  262. graph_executer - execute uncached: quantization mode none
  263. aquant - input file ['nntool_input/quantization_files/frame_00016.jpg']
  264. graph_executer - execute uncached: quantization mode none
  265. aquant - input file ['nntool_input/quantization_files/frame_00036.jpg']
  266. graph_executer - execute uncached: quantization mode none
  267. aquant - input file ['nntool_input/quantization_files/frame_00151.jpg']
  268. graph_executer - execute uncached: quantization mode none
  269. aquant - input file ['nntool_input/quantization_files/frame_00038.jpg']
  270. graph_executer - execute uncached: quantization mode none
  271. aquant - input file ['nntool_input/quantization_files/1479425444633286162.jpg']
  272. graph_executer - execute uncached: quantization mode none
  273. aquant - input file ['nntool_input/quantization_files/frame_00017.jpg']
  274. graph_executer - execute uncached: quantization mode none
  275. aquant - input file ['nntool_input/quantization_files/1479425447083596868.jpg']
  276. graph_executer - execute uncached: quantization mode none
  277. aquant - input file ['nntool_input/quantization_files/frame_00035.jpg']
  278. graph_executer - execute uncached: quantization mode none
  279. aquant - input file ['nntool_input/quantization_files/1479425447133467484.jpg']
  280. graph_executer - execute uncached: quantization mode none
  281. aquant - input file ['nntool_input/quantization_files/frame_00058.jpg']
  282. graph_executer - execute uncached: quantization mode none
  283. aquant - input file ['nntool_input/quantization_files/frame_00195.jpg']
  284. graph_executer - execute uncached: quantization mode none
  285. aquant - input file ['nntool_input/quantization_files/frame_00121.jpg']
  286. graph_executer - execute uncached: quantization mode none
  287. aquant - input file ['nntool_input/quantization_files/frame_00034.jpg']
  288. graph_executer - execute uncached: quantization mode none
  289. aquant - input file ['nntool_input/quantization_files/frame_00140.jpg']
  290. graph_executer - execute uncached: quantization mode none
  291. aquant - input file ['nntool_input/quantization_files/frame_00015.jpg']
  292. graph_executer - execute uncached: quantization mode none
  293. aquant - input file ['nntool_input/quantization_files/1479425443682982187.jpg']
  294. graph_executer - execute uncached: quantization mode none
  295. aquant - input file ['nntool_input/quantization_files/frame_00131.jpg']
  296. graph_executer - execute uncached: quantization mode none
  297. aquant - input file ['nntool_input/quantization_files/frame_00018.jpg']
  298. graph_executer - execute uncached: quantization mode none
  299. aquant - input file ['nntool_input/quantization_files/frame_00158.jpg']
  300. graph_executer - execute uncached: quantization mode none
  301. aquant - input file ['nntool_input/quantization_files/1479425441182877835.jpg']
  302. graph_executer - execute uncached: quantization mode none
  303. aquant - input file ['nntool_input/quantization_files/1479425445983503068.jpg']
  304. graph_executer - execute uncached: quantization mode none
  305. aquant - input file ['nntool_input/quantization_files/frame_00032.jpg']
  306. graph_executer - execute uncached: quantization mode none
  307. aquant - Quantization set. Use qshow command to see it.
  308. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  309. | Step |      Name      |        In        |       Out        |     Weights      |   Bias   | Mulbias  | Calc  |  Acc  |
  310. +======+================+==================+==================+==================+==========+==========+=======+=======+
  311. |  0   | input_1        |                  | -1.01<i8*0.00787 |                  |          |          |       |       |
  312. |      |                |                  | 402<1.00         |                  |          |          |       |       |
  313. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  314. |  1   | Conv_0         | -1.01<i8*0.00787 | -4.27<i8*0.03333 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  315. |      |                | 402<1.00         | 554<4.23         | n                |          |          |       |       |
  316. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  317. |  1   | MaxPool_1      | -4.27<i8*0.03333 | -4.27<i8*0.03333 |                  |          |          |       |       |
  318. |      |                | 554<4.23         | 554<4.23         |                  |          |          |       |       |
  319. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  320. |  1   | Conv_0_fusion  | -1.01<i8*0.00787 | -4.27<i8*0.03333 |                  |          |          |       |       |
  321. |      |                | 402<1.00         | 554<4.23         |                  |          |          |       |       |
  322. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  323. |  2   | Conv_2         | -4.27<i8*0.03333 | -6.05<i8*0.04724 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  324. |      |                | 554<4.23         | 409<6.00         | n                |          |          |       |       |
  325. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  326. |  2   | Clip_4         | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  327. |      |                | 409<6.00         | 409<6.00         |                  |          |          |       |       |
  328. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  329. |  2   | Conv_2_fusion  | -4.27<i8*0.03333 | -6.05<i8*0.04724 |                  |          |          |       |       |
  330. |      |                | 554<4.23         | 409<6.00         |                  |          |          |       |       |
  331. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  332. |  3   | Conv_5         | -6.05<i8*0.04724 | -6.05<i8*0.04724 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  333. |      |                | 409<6.00         | 409<6.00         | n                |          |          |       |       |
  334. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  335. |  3   | Clip_7         | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  336. |      |                | 409<6.00         | 409<6.00         |                  |          |          |       |       |
  337. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  338. |  3   | Conv_5_fusion  | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  339. |      |                | 409<6.00         | 409<6.00         |                  |          |          |       |       |
  340. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  341. |  4   | Conv_8         | -4.27<i8*0.03333 | -4.89<i8*0.03819 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  342. |      |                | 554<4.23         | 656<4.85         | n                |          |          |       |       |
  343. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  344. |  4   | Clip_9         | -4.89<i8*0.03819 | -4.89<i8*0.03819 |                  |          |          |       |       |
  345. |      |                | 656<4.85         | 656<4.85         |                  |          |          |       |       |
  346. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  347. |  4   | Conv_8_fusion  | -4.27<i8*0.03333 | -4.89<i8*0.03819 |                  |          |          |       |       |
  348. |      |                | 554<4.23         | 656<4.85         |                  |          |          |       |       |
  349. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  350. |  5   | Add_10         | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  351. |      |                | 409<6.00,-4.89<i | 409<6.00         |                  |          |          |       |       |
  352. |      |                | 8*0.03819656<4.8 |                  |                  |          |          |       |       |
  353. |      |                | 5                |                  |                  |          |          |       |       |
  354. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  355. |  6   | Conv_11        | -6.05<i8*0.04724 | -6.05<i8*0.04724 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  356. |      |                | 409<6.00         | 409<6.00         | n                |          |          |       |       |
  357. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  358. |  6   | Clip_13        | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  359. |      |                | 409<6.00         | 409<6.00         |                  |          |          |       |       |
  360. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  361. |  6   | Conv_11_fusion | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  362. |      |                | 409<6.00         | 409<6.00         |                  |          |          |       |       |
  363. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  364. |  7   | Conv_14        | -6.05<i8*0.04724 | -6.05<i8*0.04724 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  365. |      |                | 409<6.00         | 409<6.00         | n                |          |          |       |       |
  366. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  367. |  7   | Clip_16        | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  368. |      |                | 409<6.00         | 409<6.00         |                  |          |          |       |       |
  369. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  370. |  7   | Conv_14_fusion | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  371. |      |                | 409<6.00         | 409<6.00         |                  |          |          |       |       |
  372. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  373. |  8   | Conv_17        | -6.05<i8*0.04724 | -4.91<i8*0.03837 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  374. |      |                | 409<6.00         | 075<4.87         | n                |          |          |       |       |
  375. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  376. |  8   | Clip_18        | -4.91<i8*0.03837 | -4.91<i8*0.03837 |                  |          |          |       |       |
  377. |      |                | 075<4.87         | 075<4.87         |                  |          |          |       |       |
  378. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  379. |  8   | Conv_17_fusion | -6.05<i8*0.04724 | -4.91<i8*0.03837 |                  |          |          |       |       |
  380. |      |                | 409<6.00         | 075<4.87         |                  |          |          |       |       |
  381. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  382. |  9   | Add_19         | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  383. |      |                | 409<6.00,-4.91<i | 409<6.00         |                  |          |          |       |       |
  384. |      |                | 8*0.03837075<4.8 |                  |                  |          |          |       |       |
  385. |      |                | 7                |                  |                  |          |          |       |       |
  386. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  387. |  10  | Conv_20        | -6.05<i8*0.04724 | -6.05<i8*0.04724 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  388. |      |                | 409<6.00         | 409<6.00         | n                |          |          |       |       |
  389. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  390. |  10  | Clip_22        | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  391. |      |                | 409<6.00         | 409<6.00         |                  |          |          |       |       |
  392. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  393. |  10  | Conv_20_fusion | -6.05<i8*0.04724 | -6.05<i8*0.04724 |                  |          |          |       |       |
  394. |      |                | 409<6.00         | 409<6.00         |                  |          |          |       |       |
  395. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  396. |  11  | Conv_23        | -6.05<i8*0.04724 | -1.14<i8*0.00890 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  397. |      |                | 409<6.00         | 456<1.13         | n                |          |          |       |       |
  398. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  399. |  11  | Clip_25        | -1.14<i8*0.00890 | -1.14<i8*0.00890 |                  |          |          |       |       |
  400. |      |                | 456<1.13         | 456<1.13         |                  |          |          |       |       |
  401. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  402. |  11  | Conv_23_fusion | -6.05<i8*0.04724 | -1.14<i8*0.00890 |                  |          |          |       |       |
  403. |      |                | 409<6.00         | 456<1.13         |                  |          |          |       |       |
  404. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  405. |  12  | Conv_26        | -6.05<i8*0.04724 | -0.05<i8*0.00040 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  406. |      |                | 409<6.00         | 486<0.05         | n                |          |          |       |       |
  407. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  408. |  12  | Clip_27        | -0.05<i8*0.00040 | -0.05<i8*0.00040 |                  |          |          |       |       |
  409. |      |                | 486<0.05         | 486<0.05         |                  |          |          |       |       |
  410. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  411. |  12  | Conv_26_fusion | -6.05<i8*0.04724 | -0.05<i8*0.00040 |                  |          |          |       |       |
  412. |      |                | 409<6.00         | 486<0.05         |                  |          |          |       |       |
  413. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  414. |  13  | Add_28fusion   | -1.14<i8*0.00890 | -2.28<i8*0.01780 |                  |          |          |       |       |
  415. |      |                | 456<1.13,-0.05<i | 912<2.26         |                  |          |          |       |       |
  416. |      |                | 8*0.00040486<0.0 |                  |                  |          |          |       |       |
  417. |      |                | 5                |                  |                  |          |          |       |       |
  418. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  419. |  14  | Flatten_30     | -2.28<i8*0.01780 | -2.28<i8*0.01780 |                  |          |          |       |       |
  420. |      |                | 912<2.26         | 912<2.26         |                  |          |          |       |       |
  421. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  422. |  15  | Gemm_31        | -2.28<i8*0.01780 | -40.88<i8*0.3193 | chan<i8*chan<cha | i32*chan | 8b>>chan | Q32.0 | Q32.0 |
  423. |      |                | 912<2.26         | 5582<40.56       | n                |          |          |       |       |
  424. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  425. |  16  | Gemm_31_split  | -40.88<i8*0.3193 | -40.88<i8*0.3193 |                  |          |          |       |       |
  426. |      |                | 5582<40.56       | 5582<40.56,-40.8 |                  |          |          |       |       |
  427. |      |                |                  | 8<i8*0.31935582< |                  |          |          |       |       |
  428. |      |                |                  | 40.56            |                  |          |          |       |       |
  429. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  430. |  17  | output_1       | -40.88<i8*0.3193 | -40.88<i8*0.3193 |                  |          |          |       |       |
  431. |      |                | 5582<40.56       | 5582<40.56       |                  |          |          |       |       |
  432. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  433. |  18  | Sigmoid_36     | -40.88<i8*0.3193 | -1.01<i8*0.00787 |                  |          |          |       |       |
  434. |      |                | 5582<40.56       | 402<1.00         |                  |          |          |       |       |
  435. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  436. |  19  | output_2       | -1.01<i8*0.00787 | -1.01<i8*0.00787 |                  |          |          |       |       |
  437. |      |                | 402<1.00         | 402<1.00         |                  |          |          |       |       |
  438. +------+----------------+------------------+------------------+------------------+----------+----------+-------+-------+
  439. nngraph - update graph dimensions
  440. set_aliases - looking for aliased edges
  441. nngraph - calculate liveness
  442. l3_ram_ext_managed - was: False
  443. now: True
  444. default_input_exec_location - was: 'AT_MEM_L2'
  445. now: 'AT_MEM_L3_HRAM'
  446. graph_produce_node_names - was: False
  447. now: True
  448. graph_reorder_constant_in - was: True
  449. now: True
  450. graph_produce_operinfos - was: False
  451. now: True
  452. graph_monitor_cycles - was: False
  453. now: True
  454. new_param_state - saving parameters for step input_1 input_1
  455. new_param_state - saving parameters for step input_1_formatter input_1_formatter
  456. new_param_state - saving parameters for step Conv_0_fusion Conv_0
  457. new_param_state - saving parameters for step Conv_0_fusion MaxPool_1
  458. new_param_state - saving parameters for step Conv_0_fusion Conv_0_fusion
  459. new_param_state - saving parameters for step Conv_2_fusion Conv_2
  460. new_param_state - saving parameters for step Conv_2_fusion Clip_4
  461. new_param_state - saving parameters for step Conv_2_fusion Conv_2_fusion
  462. new_param_state - saving parameters for step Conv_5_fusion Conv_5
  463. new_param_state - saving parameters for step Conv_5_fusion Clip_7
  464. new_param_state - saving parameters for step Conv_5_fusion Conv_5_fusion
  465. new_param_state - saving parameters for step Conv_8_fusion Conv_8
  466. new_param_state - saving parameters for step Conv_8_fusion Clip_9
  467. new_param_state - saving parameters for step Conv_8_fusion Conv_8_fusion
  468. new_param_state - saving parameters for step Add_10 Add_10
  469. new_param_state - saving parameters for step Conv_11_fusion Conv_11
  470. new_param_state - saving parameters for step Conv_11_fusion Clip_13
  471. new_param_state - saving parameters for step Conv_11_fusion Conv_11_fusion
  472. new_param_state - saving parameters for step Conv_14_fusion Conv_14
  473. new_param_state - saving parameters for step Conv_14_fusion Clip_16
  474. new_param_state - saving parameters for step Conv_14_fusion Conv_14_fusion
  475. new_param_state - saving parameters for step Conv_17_fusion Conv_17
  476. new_param_state - saving parameters for step Conv_17_fusion Clip_18
  477. new_param_state - saving parameters for step Conv_17_fusion Conv_17_fusion
  478. new_param_state - saving parameters for step Add_19 Add_19
  479. new_param_state - saving parameters for step Conv_20_fusion Conv_20
  480. new_param_state - saving parameters for step Conv_20_fusion Clip_22
  481. new_param_state - saving parameters for step Conv_20_fusion Conv_20_fusion
  482. new_param_state - saving parameters for step Conv_23_fusion Conv_23
  483. new_param_state - saving parameters for step Conv_23_fusion Clip_25
  484. new_param_state - saving parameters for step Conv_23_fusion Conv_23_fusion
  485. new_param_state - saving parameters for step Conv_26_fusion Conv_26
  486. new_param_state - saving parameters for step Conv_26_fusion Clip_27
  487. new_param_state - saving parameters for step Conv_26_fusion Conv_26_fusion
  488. new_param_state - saving parameters for step Add_28fusion Add_28fusion
  489. new_param_state - saving parameters for step Flatten_30 Flatten_30
  490. new_param_state - saving parameters for step Gemm_31 Gemm_31
  491. new_param_state - saving parameters for step Gemm_31_split Gemm_31_split
  492. new_param_state - saving parameters for step output_1 output_1
  493. new_param_state - saving parameters for step Sigmoid_36 Sigmoid_36
  494. new_param_state - saving parameters for step output_2 output_2
  495. new_param_state - dumping graph state to /module/data/BUILD_MODEL_SQ8BIT/network.json
  496. new_param_state - dumping tensors to /module/data/BUILD_MODEL_SQ8BIT/network.nnparam
  497. echo "GENERATING AUTOTILER MODEL"
  498. GENERATING AUTOTILER MODEL
  499. nntool -g -M nntool_output -m networkModel.c -T nntool_output/tensors BUILD_MODEL_SQ8BIT/network.json
  500. onnx - unable to determine batch dimension. if the graph fails to import properly set it to 1 or a variable.
  501. echo "COMPILING AUTOTILER MODEL"
  502. COMPILING AUTOTILER MODEL
  503. gcc -g -o BUILD_MODEL_SQ8BIT/GenTile -I. -I/gap_sdk/tools/autotiler_v3/Autotiler -I/gap_sdk/tools/autotiler_v3/Emulation -I/gap_sdk/tools/autotiler_v3/CNN_Generators -I/gap_sdk/tools/autotiler_v3/CNN_Generators_SQ8 -I/gap_sdk/tools/nntool//autotiler/generators -I/gap_sdk/tools/autotiler_v3/CNN_Libraries -I/gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8 -I/gap_sdk/tools/nntool//autotiler/kernels /gap_sdk/tools/autotiler_v3/CNN_Generators/CNN_Generator_Util.c /gap_sdk/tools/autotiler_v3/CNN_Generators_SQ8/CNN_Generators_SQ8.c /gap_sdk/tools/autotiler_v3/CNN_Generators_SQ8/RNN_Generators_SQ8.c /gap_sdk/tools/nntool//autotiler/generators/nntool_extra_generators.c nntool_output/networkModel.c /gap_sdk/tools/autotiler_v3/Autotiler/LibTile.a
  504. echo "RUNNING AUTOTILER MODEL"
  505. RUNNING AUTOTILER MODEL
  506. BUILD_MODEL_SQ8BIT/GenTile -o BUILD_MODEL_SQ8BIT -c BUILD_MODEL_SQ8BIT --L1 46736 --L2 200000 --L3 8000000 --L1 46736 --L2 200000 --L3 8000000
  507. CNN_NormBW: S1_Op_input_1_formatter
  508. In  => Feat: 1 W:  200, H:    1
  509. Out => Feat: 1, W:  200, H:    1
  510.              KerName: CNN_NormBW_shift_fps
  511. Nb Oper : 200
  512.  
  513. ==== Process Tiling For User Kernel:              S1_Op_input_1_formatter =======================
  514. S1_Op_input_1_formatter Partition[0] Size =    808 (Min:      0, Max:    816), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  515. Kernel: S1_Op_input_1_formatter, Total Raw Memory: 400 fits into L1 memory 46736. Promoting all kernel arguments to initialized buffers.
  516. S1_Op_input_1_formatter For Iter Space: 0 Iteration count:   1, Given L1 Memory:  46736, Used L1 Memory:    400, Reusable Memory: 46336, Used L2 Memory: 0
  517. =================================================================================================
  518.  
  519. InFeat: 1, OutFeat: 32
  520. Conv => W:  200, Pad:[2,1] PadT:[2,1] => Wc: 100, Filter:[5,5]
  521.      => H:  200, Pad:[2,1] PadT:[2,1] => Hc: 100
  522. Pool => Wc: 100, Pad:[0,0] => Wo: 50, Filter:[2,2]
  523.      => Hc: 100, Pad:[0,0] => Ho: 50
  524. OverlapC: 3
  525. OverlapP: 0
  526. TileCons: 2
  527. UsedIn  : [200 x 200]
  528. UsedC   : [100 x 100]
  529.       SetBiasKerName: KerParSetBiasB32_SQ8
  530.          ConvKerName: KerParConv5x5Stride2_SQ8
  531.   DPReductionKerName: KerParReductIO_CC_SQ8
  532.          PoolKerName: KerParPool2x2Stride2_SQ8
  533. Nb Oper : 8320000
  534.  
  535. ==== Process Tiling For User Kernel:       S2_Conv2d_32x1x5x5_MaxPool_2x2 =======================
  536. S2_Conv2d_32x1x5x5_MaxPool_2x2 Partition[0] Size = 1290425 (Min:   2000, Max: 1523449), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  537. S2_Conv2d_32x1x5x5_MaxPool_2x2 Full buffering on Arg: Bias, was using 256 bytes will require 128 bytes buffer
  538. S2_Conv2d_32x1x5x5_MaxPool_2x2 Full buffering on Arg: Scale, was using 64 bytes will require 32 bytes buffer
  539. S2_Conv2d_32x1x5x5_MaxPool_2x2 Full buffering on Arg: ScaleN, was using 64 bytes will require 32 bytes buffer
  540. S2_Conv2d_32x1x5x5_MaxPool_2x2, TiledSpace: Tile0 Iteration Count: 50 Parametric Space: [D1, M0=32] Parametric Space: [D0, M1=1]
  541.               In : Ratio: 4.000000, FixDim:    200, VarDim:      7 [   200], Size:   2800, Total:    2800, Move:      69400 (Decl x 1.735000) L2
  542. *           Bias : Ratio: 0.000000,                                          Size:    128, Total:    2928, Move:        128 (Decl x 1.000000) L2
  543. *          Scale : Ratio: 0.000000,                                          Size:     32, Total:    2960, Move:         32 (Decl x 1.000000) L2
  544. *         ScaleN : Ratio: 0.000000,                                          Size:     32, Total:    2992, Move:         32 (Decl x 1.000000) L2
  545. @         Filter : Ratio: 0.000000,                                          Size:    800, Total:    3792, Move:        800 (Decl x 1.000000) L2
  546.              Out : Ratio: 1.000000, FixDim:     50, VarDim:      1 [    50], Size:   3200, Total:    6992, Move:      80000 (Decl x 1.000000) L2
  547. *        ConvOut : Ratio: 2.000000, FixDim:    100, VarDim:      2 [   100], Size:  25600, Total:   32592, Move:          0 (Decl x 0.000000) L2
  548. *          Infos : Ratio: 0.000000,                                          Size:     12, Total:   32604, Move:          9 (Decl x 1.000000) L2
  549. S2_Conv2d_32x1x5x5_MaxPool_2x2 - IterSpace: Tile0 - L1 Memory:  32604, L2Move: 150401, L3Move: 0, Tiling Overhead: 1.242973
  550. S2_Conv2d_32x1x5x5_MaxPool_2x2 Partial buffering on Arg: Filter, From: D0 To: D1. Current is (Par) 1 x [W:1, H:1] x 25 => Partial buffer size is 1600 bytes
  551. S2_Conv2d_32x1x5x5_MaxPool_2x2 Found Parametric value for space D1 (Initial: 32, Div: 8) = 32 [32*1 + 0] and space D0 (Initial: 1, Div: 4) = 1 [1*1 + 0], Iteration for Tiled Space: 50
  552. S2_Conv2d_32x1x5x5_MaxPool_2x2 For Iter Space: 0 Iteration count:  50, Given L1 Memory:  46736, Used L1 Memory:  33404, Reusable Memory: 13332, Used L2 Memory: 0
  553. =================================================================================================
  554.  
  555. InFeat: 32, OutFeat: 32
  556. Conv => W:  50, Pad:[1,0] PadT:[1,0] => Wc: 25, Filter:[3,3]
  557.      => H:  50, Pad:[1,0] PadT:[1,0] => Hc: 25
  558. Pool => Wc: 25, Pad:[0,0] => Wo: 25, Filter:[1,1]
  559.      => Hc: 25, Pad:[0,0] => Ho: 25
  560. OverlapC: 1
  561. OverlapP: 0
  562. TileCons: 2
  563. UsedIn  : [50 x 50]
  564. UsedC   : [25 x 25]
  565.       SetBiasKerName: KerParSetBiasB32_SQ8
  566.          ConvKerName: KerParConv3x3Stride2_SQ8
  567.   DPReductionKerName: KerParReduct_CC_ReLU_SQ8
  568. Nb Oper : 5780000
  569.  
  570. ==== Process Tiling For User Kernel:            S3_Conv2d_32x32x3x3_Relu6 =======================
  571. S3_Conv2d_32x32x3x3_Relu6 Partition[0] Size =  93801 (Min:    300, Max: 284425), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  572. S3_Conv2d_32x32x3x3_Relu6 Full buffering on Arg: Bias, was using 256 bytes will require 128 bytes buffer
  573. S3_Conv2d_32x32x3x3_Relu6 Full buffering on Arg: Scale, was using 64 bytes will require 32 bytes buffer
  574. S3_Conv2d_32x32x3x3_Relu6 Full buffering on Arg: ScaleN, was using 64 bytes will require 32 bytes buffer
  575. S3_Conv2d_32x32x3x3_Relu6, TiledSpace: Tile0 Iteration Count: 7 Parametric Space: [D1, M0=32] Parametric Space: [D0, M1=8]
  576.               In : Ratio: 2.000000, FixDim:     50, VarDim:      9 [    50], Size:   7200, Total:    7200, Move:      89600 (Decl x 1.120000) L2
  577. *           Bias : Ratio: 0.000000,                                          Size:    128, Total:    7328, Move:        128 (Decl x 1.000000) L2
  578. *          Scale : Ratio: 0.000000,                                          Size:     32, Total:    7360, Move:         32 (Decl x 1.000000) L2
  579. *         ScaleN : Ratio: 0.000000,                                          Size:     32, Total:    7392, Move:         32 (Decl x 1.000000) L2
  580. @         Filter : Ratio: 0.000000,                                          Size:   9216, Total:   16608, Move:       9216 (Decl x 1.000000) L2
  581.              Out : Ratio: 1.000000, FixDim:     25, VarDim:      4 [    25], Size:   6400, Total:   23008, Move:      20000 (Decl x 1.000000) L2
  582. *        ConvOut : Ratio: 1.000000, FixDim:     25, VarDim:      4 [    25], Size:  12800, Total:   35808, Move:          0 (Decl x 0.000000) L2
  583. *          Infos : Ratio: 0.000000,                                          Size:     12, Total:   35820, Move:          9 (Decl x 1.000000) L2
  584. S3_Conv2d_32x32x3x3_Relu6 - IterSpace: Tile0 - L1 Memory:  35820, L2Move: 119017, L3Move: 0, Tiling Overhead: 1.087738
  585. S3_Conv2d_32x32x3x3_Relu6 Partial buffering on Arg: Filter, From: D0 To: D1. Current is (Par) 32 x [W:1, H:1] x 9 => Partial buffer size is 18432 bytes
  586. S3_Conv2d_32x32x3x3_Relu6 Found Parametric value for space D1 (Initial: 32, Div: 8) = 32 [32*1 + 0] and space D0 (Initial: 32, Div: 4) = 8 [8*4 + 0], Iteration for Tiled Space: 7
  587. S3_Conv2d_32x32x3x3_Relu6 For Iter Space: 0 Iteration count:   7 (Last one is truncated), Given L1 Memory:  46736, Used L1 Memory:  45036, Reusable Memory: 1700, Used L2 Memory: 0
  588. =================================================================================================
  589.  
  590. InFeat: 32, OutFeat: 32
  591. Conv => W:  25, Pad:[1,1] PadT:[1,1] => Wc: 25, Filter:[3,3]
  592.      => H:  25, Pad:[1,1] PadT:[1,1] => Hc: 25
  593. Pool => Wc: 25, Pad:[0,0] => Wo: 25, Filter:[1,1]
  594.      => Hc: 25, Pad:[0,0] => Ho: 25
  595. OverlapC: 2
  596. OverlapP: 0
  597. TileCons: 1
  598. UsedIn  : [25 x 25]
  599. UsedC   : [25 x 25]
  600.       SetBiasKerName: KerParSetBiasB32_SQ8
  601.          ConvKerName: KerParConv3x3Stride1_SQ8
  602.   DPReductionKerName: KerParReduct_CC_ReLU_SQ8
  603. Nb Oper : 5780000
  604.  
  605. ==== Process Tiling For User Kernel:            S4_Conv2d_32x32x3x3_Relu6 =======================
  606. S4_Conv2d_32x32x3x3_Relu6 Partition[0] Size =  87401 (Min:    150, Max: 164425), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  607. S4_Conv2d_32x32x3x3_Relu6 Full buffering on Arg: Bias, was using 256 bytes will require 128 bytes buffer
  608. S4_Conv2d_32x32x3x3_Relu6 Full buffering on Arg: Scale, was using 64 bytes will require 32 bytes buffer
  609. S4_Conv2d_32x32x3x3_Relu6 Full buffering on Arg: ScaleN, was using 64 bytes will require 32 bytes buffer
  610. S4_Conv2d_32x32x3x3_Relu6, TiledSpace: Tile0 Iteration Count: 5 Parametric Space: [D1, M0=32] Parametric Space: [D0, M1=8]
  611.               In : Ratio: 1.000000, FixDim:     25, VarDim:      7 [    25], Size:   2800, Total:    2800, Move:      26400 (Decl x 1.320000) L2
  612. *           Bias : Ratio: 0.000000,                                          Size:    128, Total:    2928, Move:        128 (Decl x 1.000000) L2
  613. *          Scale : Ratio: 0.000000,                                          Size:     32, Total:    2960, Move:         32 (Decl x 1.000000) L2
  614. *         ScaleN : Ratio: 0.000000,                                          Size:     32, Total:    2992, Move:         32 (Decl x 1.000000) L2
  615. @         Filter : Ratio: 0.000000,                                          Size:   9216, Total:   12208, Move:       9216 (Decl x 1.000000) L2
  616.              Out : Ratio: 1.000000, FixDim:     25, VarDim:      5 [    25], Size:   8000, Total:   20208, Move:      20000 (Decl x 1.000000) L2
  617. *        ConvOut : Ratio: 1.000000, FixDim:     25, VarDim:      5 [    25], Size:  16000, Total:   36208, Move:          0 (Decl x 0.000000) L2
  618. *          Infos : Ratio: 0.000000,                                          Size:     12, Total:   36220, Move:          9 (Decl x 1.000000) L2
  619. S4_Conv2d_32x32x3x3_Relu6 - IterSpace: Tile0 - L1 Memory:  36220, L2Move: 55817, L3Move: 0, Tiling Overhead: 1.129510
  620. S4_Conv2d_32x32x3x3_Relu6 Partial buffering on Arg: Filter, From: D0 To: D1. Current is (Par) 32 x [W:1, H:1] x 9 => Partial buffer size is 18432 bytes
  621. S4_Conv2d_32x32x3x3_Relu6 Found Parametric value for space D1 (Initial: 32, Div: 8) = 32 [32*1 + 0] and space D0 (Initial: 32, Div: 4) = 8 [8*4 + 0], Iteration for Tiled Space: 5
  622. S4_Conv2d_32x32x3x3_Relu6 For Iter Space: 0 Iteration count:   5, Given L1 Memory:  46736, Used L1 Memory:  45436, Reusable Memory: 1300, Used L2 Memory: 0
  623. =================================================================================================
  624.  
  625. InFeat: 32, OutFeat: 32
  626. Conv => W:  50, Pad:[0,0] PadT:[0,0] => Wc: 25, Filter:[1,1]
  627.      => H:  50, Pad:[0,0] PadT:[0,0] => Hc: 25
  628. Pool => Wc: 25, Pad:[0,0] => Wo: 25, Filter:[1,1]
  629.      => Hc: 25, Pad:[0,0] => Ho: 25
  630. OverlapC: -1
  631. OverlapP: 0
  632. TileCons: 2
  633. UsedIn  : [49 x 49]
  634. UsedC   : [25 x 25]
  635.       SetBiasKerName: KerParSetBiasB32_SQ8
  636.          ConvKerName: KerParConv1x1Stride2_SQ8
  637.   DPReductionKerName: KerParReduct_CC_ReLU_SQ8
  638. Nb Oper : 660000
  639. Mapping this convolution to matrix multiplication with small first operand
  640. CNN_MatMulSmallM1_SQ8: S5_Conv2d_32x32x1x1_Relu6
  641. In1  => W:   32, H:   32
  642. In2  => W: 2500, H:   32, w:   50, h:   50, Sx: 2, Sy: 2, TileCons: 100
  643. Out  => W:  625, H:   32
  644.        MatMulKerName: KerParMatMulB32_ReLU_SF_SQ8
  645.      MatTransKerName: CNN_TransposeSxSy_fps
  646. Act: ReLU
  647. Nb Oper : 640000
  648.  
  649. ==== Process Tiling For User Kernel:            S5_Conv2d_32x32x1x1_Relu6 =======================
  650. S5_Conv2d_32x32x1x1_Relu6 Partition[0] Size =  34057 (Min:   6400, Max: 221481), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  651. S5_Conv2d_32x32x1x1_Relu6, TiledSpace: Tile0 Iteration Count: 5
  652. *            In1 : Ratio: 0.000000,                                          Size:   1024, Total:    1024, Move:       1024 (Decl x 1.000000) L2
  653.              In2 : Ratio: 4.000000, FixDim:     32, VarDim:    500 [  2500], Size:  32000, Total:   33024, Move:      80000 (Decl x 1.000000) L2
  654. *       TransIn2 : Ratio: 1.000000, FixDim:     32, VarDim:    125 [   625], Size:   4000, Total:   37024, Move:          0 (Decl x 0.000000) L2
  655. *           Bias : Ratio: 0.000000,                                          Size:    128, Total:   37152, Move:        128 (Decl x 1.000000) L2
  656.              Out : Ratio: 1.000000, FixDim:     32, VarDim:    125 [   625], Size:   8000, Total:   45152, Move:      20000 (Decl x 1.000000) L2
  657. *          Scale : Ratio: 0.000000,                                          Size:     32, Total:   45184, Move:         32 (Decl x 1.000000) L2
  658. *         ScaleN : Ratio: 0.000000,                                          Size:     32, Total:   45216, Move:         32 (Decl x 1.000000) L2
  659. *          Infos : Ratio: 0.000000,                                          Size:     12, Total:   45228, Move:          9 (Decl x 1.000000) L2
  660. S5_Conv2d_32x32x1x1_Relu6 - IterSpace: Tile0 - L1 Memory:  45228, L2Move: 101225, L3Move: 0, Tiling Overhead: 1.000000
  661. S5_Conv2d_32x32x1x1_Relu6 Iteration for Tiled Space: 5
  662. S5_Conv2d_32x32x1x1_Relu6 For Iter Space: 0 Iteration count:   5, Given L1 Memory:  46736, Used L1 Memory:  45228, Reusable Memory: 1508, Used L2 Memory: 0
  663. =================================================================================================
  664.  
  665.  
  666. ==== Process Tiling For User Kernel:                   S6_MatAdd_32x25x25 =======================
  667.   S6_MatAdd_32x25x25 Partition[0] Size =   4825 (Min:      0, Max: 120073), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  668.   S6_MatAdd_32x25x25, TiledSpace: Tile0 Iteration Count: 3 Parametric Space: [D0, M0=32]
  669.              In1 : Ratio: 1.000000, FixDim:     25, VarDim:      9 [    25], Size:  14400, Total:   14400, Move:      20000 (Decl x 1.000000) L2
  670.              In2 : Ratio: 1.000000, FixDim:     25, VarDim:      9 [    25], Size:  14400, Total:   28800, Move:      20000 (Decl x 1.000000) L2
  671.              Out : Ratio: 1.000000, FixDim:     25, VarDim:      9 [    25], Size:  14400, Total:   43200, Move:      20000 (Decl x 1.000000) L2
  672. *          Infos : Ratio: 0.000000,                                          Size:     12, Total:   43212, Move:          9 (Decl x 1.000000) L2
  673.   S6_MatAdd_32x25x25 - IterSpace: Tile0 - L1 Memory:  43212, L2Move: 60009, L3Move: 0, Tiling Overhead: 1.000000
  674.   S6_MatAdd_32x25x25 Found Parametric value for space D0 (Initial: 32, Div: 8) = 32 [32*1 + 0], Iteration for Tiled Space: 3
  675.   S6_MatAdd_32x25x25 For Iter Space: 0 Iteration count:   3 (Last one is truncated), Given L1 Memory:  46736, Used L1 Memory:  43212, Reusable Memory: 3524, Used L2 Memory: 0
  676. =================================================================================================
  677.  
  678. InFeat: 32, OutFeat: 64
  679. Conv => W:  25, Pad:[1,1] PadT:[1,1] => Wc: 13, Filter:[3,3]
  680.      => H:  25, Pad:[1,1] PadT:[1,1] => Hc: 13
  681. Pool => Wc: 13, Pad:[0,0] => Wo: 13, Filter:[1,1]
  682.      => Hc: 13, Pad:[0,0] => Ho: 13
  683. OverlapC: 1
  684. OverlapP: 0
  685. TileCons: 2
  686. UsedIn  : [25 x 25]
  687. UsedC   : [13 x 13]
  688.       SetBiasKerName: KerParSetBiasB32_SQ8
  689.          ConvKerName: KerParConv3x3Stride2_SQ8
  690.   DPReductionKerName: KerParReduct_CC_ReLU_SQ8
  691. Nb Oper : 3125824
  692.  
  693. ==== Process Tiling For User Kernel:            S7_Conv2d_64x32x3x3_Relu6 =======================
  694. S7_Conv2d_64x32x3x3_Relu6 Partition[0] Size =  53353 (Min:    150, Max: 110281), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  695. S7_Conv2d_64x32x3x3_Relu6 Full buffering on Arg: Bias, was using 512 bytes will require 256 bytes buffer
  696. S7_Conv2d_64x32x3x3_Relu6 Full buffering on Arg: Scale, was using 128 bytes will require 64 bytes buffer
  697. S7_Conv2d_64x32x3x3_Relu6 Full buffering on Arg: ScaleN, was using 128 bytes will require 64 bytes buffer
  698. S7_Conv2d_64x32x3x3_Relu6, TiledSpace: Tile0 Iteration Count: 13 Parametric Space: [D1, M0=64] Parametric Space: [D0, M1=24]
  699.               In : Ratio: 2.000000, FixDim:     25, VarDim:      3 [    25], Size:   3600, Total:    3600, Move:      29600 (Decl x 1.480000) L2
  700. *           Bias : Ratio: 0.000000,                                          Size:    256, Total:    3856, Move:        256 (Decl x 1.000000) L2
  701. *          Scale : Ratio: 0.000000,                                          Size:     64, Total:    3920, Move:         64 (Decl x 1.000000) L2
  702. *         ScaleN : Ratio: 0.000000,                                          Size:     64, Total:    3984, Move:         64 (Decl x 1.000000) L2
  703. @         Filter : Ratio: 0.000000,                                          Size:  18432, Total:   22416, Move:      18432 (Decl x 1.000000) L2
  704.              Out : Ratio: 1.000000, FixDim:     13, VarDim:      1 [    13], Size:   1664, Total:   24080, Move:      10816 (Decl x 1.000000) L2
  705. *        ConvOut : Ratio: 1.000000, FixDim:     13, VarDim:      1 [    13], Size:   3328, Total:   27408, Move:          0 (Decl x 0.000000) L2
  706. *          Infos : Ratio: 0.000000,                                          Size:     12, Total:   27420, Move:          9 (Decl x 1.000000) L2
  707. S7_Conv2d_64x32x3x3_Relu6 - IterSpace: Tile0 - L1 Memory:  27420, L2Move: 59241, L3Move: 0, Tiling Overhead: 1.193389
  708. S7_Conv2d_64x32x3x3_Relu6 Partial buffering on Arg: Filter, From: D0 To: D1. Current is (Par) 32 x [W:1, H:1] x 9 => Partial buffer size is 36864 bytes
  709. S7_Conv2d_64x32x3x3_Relu6 Found Parametric value for space D1 (Initial: 64, Div: 8) = 64 [64*1 + 0] and space D0 (Initial: 32, Div: 4) = 24 [24*1 + 8], Iteration for Tiled Space: 13
  710. S7_Conv2d_64x32x3x3_Relu6 For Iter Space: 0 Iteration count:  13, Given L1 Memory:  46736, Used L1 Memory:  45852, Reusable Memory: 884, Used L2 Memory: 0
  711. =================================================================================================
  712.  
  713. InFeat: 64, OutFeat: 64
  714. Conv => W:  13, Pad:[1,1] PadT:[1,1] => Wc: 13, Filter:[3,3]
  715.      => H:  13, Pad:[1,1] PadT:[1,1] => Hc: 13
  716. Pool => Wc: 13, Pad:[0,0] => Wo: 13, Filter:[1,1]
  717.      => Hc: 13, Pad:[0,0] => Ho: 13
  718. OverlapC: 2
  719. OverlapP: 0
  720. TileCons: 1
  721. UsedIn  : [13 x 13]
  722. UsedC   : [13 x 13]
  723.       SetBiasKerName: KerParSetBiasB32_SQ8
  724.          ConvKerName: KerParConv3x3Stride1_SQ8
  725.   DPReductionKerName: KerParReduct_CC_ReLU_SQ8
  726. Nb Oper : 6240832
  727.  
  728. ==== Process Tiling For User Kernel:            S8_Conv2d_64x64x3x3_Relu6 =======================
  729. S8_Conv2d_64x64x3x3_Relu6 Partition[0] Size =  51881 (Min:     78, Max:  92041), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  730. S8_Conv2d_64x64x3x3_Relu6 Full buffering on Arg: Bias, was using 320 bytes will require 256 bytes buffer
  731. S8_Conv2d_64x64x3x3_Relu6 Full buffering on Arg: Scale, was using 80 bytes will require 64 bytes buffer
  732. S8_Conv2d_64x64x3x3_Relu6 Full buffering on Arg: ScaleN, was using 80 bytes will require 64 bytes buffer
  733. S8_Conv2d_64x64x3x3_Relu6, TiledSpace: Tile0 Iteration Count: 1 Parametric Space: [D1, M0=40] Parametric Space: [D0, M1=4]
  734.               In : Ratio: 1.000000, FixDim:     13, VarDim:     13 [    13], Size:   1352, Total:    1352, Move:      21632 (Decl x 2.000000) L2
  735. *           Bias : Ratio: 0.000000,                                          Size:    256, Total:    1608, Move:        256 (Decl x 1.000000) L2
  736. *          Scale : Ratio: 0.000000,                                          Size:     64, Total:    1672, Move:         64 (Decl x 1.000000) L2
  737. *         ScaleN : Ratio: 0.000000,                                          Size:     64, Total:    1736, Move:         64 (Decl x 1.000000) L2
  738.           Filter : Ratio: 0.000000,                                          Size:   2880, Total:    4616, Move:      36864 (Decl x 1.000000) L2
  739.              Out : Ratio: 1.000000, FixDim:     13, VarDim:     13 [    13], Size:  13520, Total:   18136, Move:      10816 (Decl x 1.000000) L2
  740. *        ConvOut : Ratio: 1.000000, FixDim:     13, VarDim:     13 [    13], Size:  27040, Total:   45176, Move:          0 (Decl x 0.000000) L2
  741. *          Infos : Ratio: 0.000000,                                          Size:     12, Total:   45188, Move:          9 (Decl x 1.000000) L2
  742. S8_Conv2d_64x64x3x3_Relu6 - IterSpace: Tile0 - L1 Memory:  45188, L2Move: 69705, L3Move: 0, Tiling Overhead: 1.183668
  743. S8_Conv2d_64x64x3x3_Relu6 Found Parametric value for space D1 (Initial: 64, Div: 8) = 40 [40*1 + 24] and space D0 (Initial: 64, Div: 4) = 4 [4*16 + 0], Iteration for Tiled Space: 1
  744. S8_Conv2d_64x64x3x3_Relu6 For Iter Space: 0 Iteration count:   1, Given L1 Memory:  46736, Used L1 Memory:  45188, Reusable Memory: 1548, Used L2 Memory: 0
  745. =================================================================================================
  746.  
  747. InFeat: 32, OutFeat: 64
  748. Conv => W:  25, Pad:[0,0] PadT:[0,0] => Wc: 13, Filter:[1,1]
  749.      => H:  25, Pad:[0,0] PadT:[0,0] => Hc: 13
  750. Pool => Wc: 13, Pad:[0,0] => Wo: 13, Filter:[1,1]
  751.      => Hc: 13, Pad:[0,0] => Ho: 13
  752. OverlapC: -1
  753. OverlapP: 0
  754. TileCons: 2
  755. UsedIn  : [25 x 25]
  756. UsedC   : [13 x 13]
  757.       SetBiasKerName: KerParSetBiasB32_SQ8
  758.          ConvKerName: KerParConv1x1Stride2_SQ8
  759.   DPReductionKerName: KerParReduct_CC_ReLU_SQ8
  760. Nb Oper : 356928
  761. Mapping this convolution to matrix multiplication with small first operand
  762. CNN_MatMulSmallM1_SQ8: S9_Conv2d_64x32x1x1_Relu6
  763. In1  => W:   32, H:   64
  764. In2  => W:  625, H:   32, w:   25, h:   25, Sx: 2, Sy: 2, TileCons: 50
  765. Out  => W:  169, H:   64
  766.        MatMulKerName: KerParMatMulB32_ReLU_SF_SQ8
  767.      MatTransKerName: CNN_TransposeSxSy_fps
  768. Act: ReLU
  769. Nb Oper : 346112
  770.  
  771. ==== Process Tiling For User Kernel:            S9_Conv2d_64x32x1x1_Relu6 =======================
  772. S9_Conv2d_64x32x1x1_Relu6 Partition[0] Size =  17481 (Min:   3200, Max:  69737), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  773. Kernel: S9_Conv2d_64x32x1x1_Relu6, Total Raw Memory: 38668 fits into L1 memory 46736. Promoting all kernel arguments to initialized buffers.
  774. S9_Conv2d_64x32x1x1_Relu6 For Iter Space: 0 Iteration count:   1, Given L1 Memory:  46736, Used L1 Memory:  38668, Reusable Memory: 8068, Used L2 Memory: 0
  775. =================================================================================================
  776.  
  777.  
  778. ==== Process Tiling For User Kernel:                  S10_MatAdd_64x13x13 =======================
  779.  S10_MatAdd_64x13x13 Partition[0] Size =   5017 (Min:      0, Max:  64969), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  780. Kernel:  S10_MatAdd_64x13x13, Total Raw Memory: 32460 fits into L1 memory 46736. Promoting all kernel arguments to initialized buffers.
  781.  S10_MatAdd_64x13x13 For Iter Space: 0 Iteration count:   1, Given L1 Memory:  46736, Used L1 Memory:  32460, Reusable Memory: 14276, Used L2 Memory: 0
  782. =================================================================================================
  783.  
  784. InFeat: 64, OutFeat: 128
  785. Conv => W:  13, Pad:[1,1] PadT:[1,1] => Wc: 7, Filter:[3,3]
  786.      => H:  13, Pad:[1,1] PadT:[1,1] => Hc: 7
  787. Pool => Wc: 7, Pad:[0,0] => Wo: 7, Filter:[1,1]
  788.      => Hc: 7, Pad:[0,0] => Ho: 7
  789. OverlapC: 1
  790. OverlapP: 0
  791. TileCons: 2
  792. UsedIn  : [13 x 13]
  793. UsedC   : [7 x 7]
  794.       SetBiasKerName: KerParSetBiasB32_SQ8
  795.          ConvKerName: KerParConv3x3Stride2_SQ8
  796.   DPReductionKerName: KerParReduct_CC_ReLU_SQ8
  797. Nb Oper : 3618944
  798.  
  799. ==== Process Tiling For User Kernel:          S11_Conv2d_128x64x3x3_Relu6 =======================
  800. S11_Conv2d_128x64x3x3_Relu6 Partition[0] Size =  37545 (Min:     78, Max:  66697), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  801. S11_Conv2d_128x64x3x3_Relu6 Full buffering on Arg: Bias, was using 896 bytes will require 512 bytes buffer
  802. S11_Conv2d_128x64x3x3_Relu6 Full buffering on Arg: Scale, was using 224 bytes will require 128 bytes buffer
  803. S11_Conv2d_128x64x3x3_Relu6 Full buffering on Arg: ScaleN, was using 224 bytes will require 128 bytes buffer
  804. S11_Conv2d_128x64x3x3_Relu6, TiledSpace: Tile0 Iteration Count: 1 Parametric Space: [D1, M0=112] Parametric Space: [D0, M1=4]
  805.               In : Ratio: 2.000000, FixDim:     13, VarDim:     13 [    13], Size:   1352, Total:    1352, Move:      21632 (Decl x 2.000000) L2
  806. *           Bias : Ratio: 0.000000,                                          Size:    512, Total:    1864, Move:        512 (Decl x 1.000000) L2
  807. *          Scale : Ratio: 0.000000,                                          Size:    128, Total:    1992, Move:        128 (Decl x 1.000000) L2
  808. *         ScaleN : Ratio: 0.000000,                                          Size:    128, Total:    2120, Move:        128 (Decl x 1.000000) L2
  809.           Filter : Ratio: 0.000000,                                          Size:   8064, Total:   10184, Move:      73728 (Decl x 1.000000) L2
  810.              Out : Ratio: 1.000000, FixDim:      7, VarDim:      7 [     7], Size:  10976, Total:   21160, Move:       6272 (Decl x 1.000000) L2
  811. *        ConvOut : Ratio: 1.000000, FixDim:      7, VarDim:      7 [     7], Size:  21952, Total:   43112, Move:          0 (Decl x 0.000000) L2
  812. *          Infos : Ratio: 0.000000,                                          Size:     12, Total:   43124, Move:          9 (Decl x 1.000000) L2
  813. S11_Conv2d_128x64x3x3_Relu6 - IterSpace: Tile0 - L1 Memory:  43124, L2Move: 102409, L3Move: 0, Tiling Overhead: 1.118088
  814. S11_Conv2d_128x64x3x3_Relu6 Found Parametric value for space D1 (Initial: 128, Div: 8) = 112 [112*1 + 16] and space D0 (Initial: 64, Div: 4) = 4 [4*16 + 0], Iteration for Tiled Space: 1
  815. S11_Conv2d_128x64x3x3_Relu6 For Iter Space: 0 Iteration count:   1, Given L1 Memory:  46736, Used L1 Memory:  43124, Reusable Memory: 3612, Used L2 Memory: 0
  816. =================================================================================================
  817.  
  818. InFeat: 128, OutFeat: 128
  819. Conv => W:  7, Pad:[1,1] PadT:[1,1] => Wc: 7, Filter:[3,3]
  820.      => H:  7, Pad:[1,1] PadT:[1,1] => Hc: 7
  821. Pool => Wc: 7, Pad:[0,0] => Wo: 7, Filter:[1,1]
  822.      => Hc: 7, Pad:[0,0] => Ho: 7
  823. OverlapC: 2
  824. OverlapP: 0
  825. TileCons: 1
  826. UsedIn  : [7 x 7]
  827. UsedC   : [7 x 7]
  828.       SetBiasKerName: KerParSetBiasB32_SQ8
  829.          ConvKerName: KerParConv3x3Stride1_SQ8
  830.   DPReductionKerName: KerParReduct_CC_ReLU_SQ8
  831. Nb Oper : 7231616
  832.  
  833. ==== Process Tiling For User Kernel:         S12_Conv2d_128x128x3x3_Relu6 =======================
  834. S12_Conv2d_128x128x3x3_Relu6 Partition[0] Size =  36137 (Min:     42, Max:  57865), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  835. S12_Conv2d_128x128x3x3_Relu6 Full buffering on Arg: Bias, was using 960 bytes will require 512 bytes buffer
  836. S12_Conv2d_128x128x3x3_Relu6 Full buffering on Arg: Scale, was using 240 bytes will require 128 bytes buffer
  837. S12_Conv2d_128x128x3x3_Relu6 Full buffering on Arg: ScaleN, was using 240 bytes will require 128 bytes buffer
  838. S12_Conv2d_128x128x3x3_Relu6, TiledSpace: Tile0 Iteration Count: 1 Parametric Space: [D1, M0=120] Parametric Space: [D0, M1=4]
  839.               In : Ratio: 1.000000, FixDim:      7, VarDim:      7 [     7], Size:    392, Total:     392, Move:      12544 (Decl x 2.000000) L2
  840. *           Bias : Ratio: 0.000000,                                          Size:    512, Total:     904, Move:        512 (Decl x 1.000000) L2
  841. *          Scale : Ratio: 0.000000,                                          Size:    128, Total:    1032, Move:        128 (Decl x 1.000000) L2
  842. *         ScaleN : Ratio: 0.000000,                                          Size:    128, Total:    1160, Move:        128 (Decl x 1.000000) L2
  843.           Filter : Ratio: 0.000000,                                          Size:   8640, Total:    9800, Move:     147456 (Decl x 1.000000) L2
  844.              Out : Ratio: 1.000000, FixDim:      7, VarDim:      7 [     7], Size:  11760, Total:   21560, Move:       6272 (Decl x 1.000000) L2
  845. *        ConvOut : Ratio: 1.000000, FixDim:      7, VarDim:      7 [     7], Size:  23520, Total:   45080, Move:          0 (Decl x 0.000000) L2
  846. *          Infos : Ratio: 0.000000,                                          Size:     12, Total:   45092, Move:          9 (Decl x 1.000000) L2
  847. S12_Conv2d_128x128x3x3_Relu6 - IterSpace: Tile0 - L1 Memory:  45092, L2Move: 167049, L3Move: 0, Tiling Overhead: 1.039011
  848. S12_Conv2d_128x128x3x3_Relu6 Found Parametric value for space D1 (Initial: 128, Div: 8) = 120 [120*1 + 8] and space D0 (Initial: 128, Div: 4) = 4 [4*32 + 0], Iteration for Tiled Space: 1
  849. S12_Conv2d_128x128x3x3_Relu6 For Iter Space: 0 Iteration count:   1, Given L1 Memory:  46736, Used L1 Memory:  45092, Reusable Memory: 1644, Used L2 Memory: 0
  850. =================================================================================================
  851.  
  852. InFeat: 64, OutFeat: 128
  853. Conv => W:  13, Pad:[0,0] PadT:[0,0] => Wc: 7, Filter:[1,1]
  854.      => H:  13, Pad:[0,0] PadT:[0,0] => Hc: 7
  855. Pool => Wc: 7, Pad:[0,0] => Wo: 7, Filter:[1,1]
  856.      => Hc: 7, Pad:[0,0] => Ho: 7
  857. OverlapC: -1
  858. OverlapP: 0
  859. TileCons: 2
  860. UsedIn  : [13 x 13]
  861. UsedC   : [7 x 7]
  862.       SetBiasKerName: KerParSetBiasB32_SQ8
  863.          ConvKerName: KerParConv1x1Stride2_SQ8
  864.   DPReductionKerName: KerParReduct_CC_ReLU_SQ8
  865. Nb Oper : 407680
  866. Mapping this convolution to matrix multiplication
  867. CNN_MatMul_SQ8: S13_Conv2d_128x64x1x1_Relu6
  868. In1  => W:   64, H:  128
  869. In2  => W:  169, H:   64, w:   13, h:   13, Sx: 2, Sy: 2
  870. Out  => W:   49, H:  128 => Column first
  871.        MatMulKerName: KerParMatMulSxSyB32_ReLU_SQ8
  872.  
  873. ==== Process Tiling For User Kernel:          S13_Conv2d_128x64x1x1_Relu6 =======================
  874. S13_Conv2d_128x64x1x1_Relu6 Partition[0] Size =   1824 (Min:   1024, Max:  17280), Fraction:       0.15, Giving:  12358 bytes out of  46736 bytes
  875. S13_Conv2d_128x64x1x1_Relu6 Partition[1] Size =  10089 (Min:   3328, Max:  34377), Fraction:       0.85, Giving:  34377 bytes out of  46736 bytes
  876. Kernel: S13_Conv2d_128x64x1x1_Relu6, Total Raw Memory: 17164 fits into L1 memory 34377. Promoting all kernel arguments to initialized buffers.
  877. S13_Conv2d_128x64x1x1_Relu6 For Iter Space: 1 Iteration count:   1, Given L1 Memory:  34377, Used L1 Memory:  17164, Reusable Memory: 17212, Used L2 Memory: 0
  878. Kernel: S13_Conv2d_128x64x1x1_Relu6, Total Raw Memory: 8960 fits into L1 memory 12358. Promoting all kernel arguments to initialized buffers.
  879. S13_Conv2d_128x64x1x1_Relu6 For Iter Space: 0 Iteration count:   1, Given L1 Memory:  12358, Used L1 Memory:   8960, Reusable Memory: 3396, Used L2 Memory: 0
  880. =================================================================================================
  881.  
  882.  
  883. ==== Process Tiling For User Kernel:             S14_MatAdd_128x7x7_Relu6 =======================
  884. S14_MatAdd_128x7x7_Relu6 Partition[0] Size =   5401 (Min:      0, Max:  37705), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  885. Kernel: S14_MatAdd_128x7x7_Relu6, Total Raw Memory: 18828 fits into L1 memory 46736. Promoting all kernel arguments to initialized buffers.
  886. S14_MatAdd_128x7x7_Relu6 For Iter Space: 0 Iteration count:   1, Given L1 Memory:  46736, Used L1 Memory:  18828, Reusable Memory: 27908, Used L2 Memory: 0
  887. =================================================================================================
  888.  
  889. Linear Layer S16_Linear_2x6272, Linear: InDim: 6272, OutDim: 2, Activation: None
  890. Linear Kernel: KerParLinearLayerFullFeatB32_SQ8
  891.  
  892. ==== Process Tiling For User Kernel:                    S16_Linear_2x6272 =======================
  893.    S16_Linear_2x6272 Partition[0] Size =  31425 (Min:      0, Max:  31593), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  894. Kernel:    S16_Linear_2x6272, Total Raw Memory: 18848 fits into L1 memory 46736. Promoting all kernel arguments to initialized buffers.
  895.    S16_Linear_2x6272 For Iter Space: 0 Iteration count:   1, Given L1 Memory:  46736, Used L1 Memory:  18848, Reusable Memory: 27888, Used L2 Memory: 0
  896. =================================================================================================
  897.  
  898. Act  => W: 1, Wo: 1
  899.      => H: 1, Ho: 1
  900.           ActKerName: KerPar_HSigmoid_SQ8
  901. Nb Oper : 1
  902.  
  903. ==== Process Tiling For User Kernel:                     S19_Act_Hsigmoid =======================
  904.     S19_Act_Hsigmoid Partition[0] Size =     25 (Min:      0, Max:     49), Fraction:       1.00, Giving:  46736 bytes out of  46736 bytes
  905. Kernel:     S19_Act_Hsigmoid, Total Raw Memory: 20 fits into L1 memory 46736. Promoting all kernel arguments to initialized buffers.
  906.     S19_Act_Hsigmoid For Iter Space: 0 Iteration count:   1, Given L1 Memory:  46736, Used L1 Memory:     20, Reusable Memory: 46716, Used L2 Memory: 0
  907. =================================================================================================
  908.  
  909. Error: Graph stacked tensor S16_Output, input tensor Output_1 is also defined as a graph local or graph argument
  910. Execution aborted
  911. riscv32-unknown-elf-gcc -g -O3 -w -mno-memcpy -fno-tree-loop-distribute-patterns -I. -I/gap_sdk/libs/gap_lib/include -I/gap_sdk/tools/autotiler_v3/Emulation -I/gap_sdk/tools/autotiler_v3/Autotiler -I/gap_sdk/tools/autotiler_v3/CNN_Libraries -I/gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8 -I/gap_sdk/tools/nntool//autotiler/kernels -IBUILD_MODEL_SQ8BIT -DAT_MODEL_PREFIX=network -DAT_INPUT_HEIGHT=200 -DAT_INPUT_WIDTH=200 -DAT_INPUT_COLORS=1 -DSTACK_SIZE=6096 -DSLAVE_STACK_SIZE=1024 -DAT_IMAGE=/module/data/images/frame_2.pgm -DPERF -DMODEL_ID= -DFREQ_FC=250 -DFREQ_CL=175 -DAT_CONSTRUCT=networkCNN_Construct -DAT_DESTRUCT=networkCNN_Destruct -DAT_CNN=networkCNN -DAT_L3_ADDR=network_L3_Flash -g -O3 -w -mno-memcpy -fno-tree-loop-distribute-patterns -I. -I/gap_sdk/libs/gap_lib/include -I/gap_sdk/tools/autotiler_v3/Emulation -I/gap_sdk/tools/autotiler_v3/Autotiler -I/gap_sdk/tools/autotiler_v3/CNN_Libraries -I/gap_sdk/tools/autotiler_v3/CNN_Libraries_SQ8 -I/gap_sdk/tools/nntool//autotiler/kernels -IBUILD_MODEL_SQ8BIT -DAT_MODEL_PREFIX=network -DAT_INPUT_HEIGHT=200 -DAT_INPUT_WIDTH=200 -DAT_INPUT_COLORS=1 -DSTACK_SIZE=6096 -DSLAVE_STACK_SIZE=1024 -DAT_IMAGE=/module/data/images/frame_2.pgm -DPERF -DMODEL_ID= -DFREQ_FC=250 -DFREQ_CL=175 -DAT_CONSTRUCT=networkCNN_Construct -DAT_DESTRUCT=networkCNN_Destruct -DAT_CNN=networkCNN -DAT_L3_ADDR=network_L3_Flash -DCONFIG_AI_DECK -DCONFIG_AI_DECK -DRT_FC_STACK_SIZE=2048 -D__PLATFORM_GVSOC__ -fno-jump-tables -fno-tree-loop-distribute-patterns -fdata-sections -ffunction-sections -mchip=gap8 -mPE=8 -mFC=1 -D__riscv__ -D__GAP__ -D__GAP8__ -DCHIP_VERSION=2 -mnativeomp -mnativeomp -D__pulp__ -DCONFIG_GAP -D__PULP_OS__ -D__pulp__ -DCONFIG_GAP -D__PULP_OS__ -MMD -MP -c main.c  -I/gap_sdk/install/GAP8_V2/include  -I/gap_sdk/install/GAP8_V2/include/io  -I/gap_sdk/install/workstation/include  -I/gap_sdk/tools/autotiler_v3/Emulation  -I/gap_sdk/tools/autotiler_v3/Emulation  -I/gap_sdk/tools/autotiler_v3/Emulation  -I/gap_sdk/tools/autotiler_v3/Emulation  -include /gap_sdk/install/GAP8_V2/include/rt/chips/gap_rev1/config.h -MD -MF /module/data/BUILD/GAP8_V2/GCC_RISCV/main.d -o /module/data/BUILD/GAP8_V2/GCC_RISCV/main.o
  912. main.c:39:10: fatal error: networkKernels.h: No such file or directory
  913.  #include "networkKernels.h"
  914.           ^~~~~~~~~~~~~~~~~~
  915. compilation terminated.
  916. /gap_sdk/tools/rules/pulp_rules.mk:204: recipe for target '/module/data/BUILD/GAP8_V2/GCC_RISCV/main.o' failed
  917. make: *** [/module/data/BUILD/GAP8_V2/GCC_RISCV/main.o] Error 1
  918. ​​​
  919. Shift + Enter 换行
  920.