Git Product home page Git Product logo

dnnmark's People

Contributors

antonok-edm avatar pyotr777 avatar shidong-ai avatar srikantamd avatar syifan avatar

Stargazers

avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar

Watchers

avatar avatar avatar avatar

Forkers

antonok-edm tituslucretius jaybooth4 s3p02 ashwinvargh syifan mattsinc srikantamd pyotr777 yf817 prdalmia googol-lab tejashah94 sabbaghm alijahan alva007 shiyuw3 christindbose

dnnmark's Issues

Path to CuDNN library not exists

I have a 1060 and a K40 installed. I use Nvidia Digits, and have CuDNN installed.

I changed the setup file to:

CUDNN_PATH=${HOME}/usr/local/cuda

I get the error :

2 GPU CUDA device(s) found

-- GPU device exists
CMake Error at CMakeLists.txt:63 (message):
Path to CuDNN library not exists

-- Configuring incomplete, errors occurred!
See also "/usr/local/DNNMark/build/CMakeFiles/CMakeOutput.log".

How do I resolve this?

/usr/local$ ls | grep cu

Gives me:

cub-1.4.1
cuda
cuda-8.0
cudnn-5.1
cudnn-8.0-linux-x64-v5.1.tgz

Input sizes for DNNMark

How can I increase (or decrease) the input sizes for a given test in DNNMark? For example, if I want to increase the input size for alexnet, fwd_softmax, or bwd_softmax, how do I go about doing that?

Matt

composed_model fails for MIOpen because dnn_params.h has missing case

The composed_model test fails on any AMD hardware I run it on (i.e., with MIOpen) because the code in dnn_params.h does not have an equivalent MIOpen case for CUDNN_CROSS_CORRELATION. I changed this locally to use miopenConvolution, and found that it resolves the issue. However, although this seems like the right solution, since there is no code here I'm wondering if this case was ever tested on AMD HW (maybe I have the wrong branch?)? Currently I'm using the develop branch, if that helps.

Additionally, there is a typo for the miopenTranspose case -- it checks for MIOPEN instead of AMD_MIOPEN.

Here is the diff of my changes:

-#ifdef NVIDIA_CUDNN
       else if (!val.compare("cross_correlation"))
+#ifdef NVIDIA_CUDNN
         conv_param->mode_ = CUDNN_CROSS_CORRELATION;
 #endif
-#ifdef MIOPEN
+#ifdef AMD_MIOPEN
+        conv_param->mode_ = miopenConvolution;
+#endif
+#ifdef AMD_MIOPEN
       else if (!val.compare("transpose"))
         conv_param->mode_ = miopenTranspose;
 #endif
       else
-        LOG(FATAL) << "Invalid conv mode" << std::endl;
+        LOG(FATAL) << "Invalid conv mode: " << val << std::endl;

I'm happy to push a patch if this makes sense, let me know.

Matt

Variables are used in this project, but they are set to NOTFOUND

Hi
Is there any suggestion for fixing this cmake build error?

-- dnnmark_test_composed_model: Benchmark source files: test_composed_model.cc
-- dnnmark_test_alexnet: Benchmark source files: test_alexnet.cc
-- dnnmark_test_VGG: Benchmark source files: test_VGG.cc
CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
GFLAGS_LIBRARY
    linked by target "dnnmark_test_fwd_conv" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_conv
    linked by target "dnnmark_test_bwd_conv" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_conv
    linked by target "dnnmark_test_fwd_pool" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_pool
    linked by target "dnnmark_test_bwd_pool" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_pool
    linked by target "dnnmark_test_fwd_lrn" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_lrn
    linked by target "dnnmark_test_bwd_lrn" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_lrn
    linked by target "dnnmark_test_fwd_activation" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_activation
    linked by target "dnnmark_test_bwd_activation" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_activation
    linked by target "dnnmark_test_fwd_fc" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_fc
    linked by target "dnnmark_test_bwd_fc" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_fc
    linked by target "dnnmark_test_fwd_softmax" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_softmax
    linked by target "dnnmark_test_bwd_softmax" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_softmax
    linked by target "dnnmark_test_fwd_bn" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_bn
    linked by target "dnnmark_test_bwd_bn" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_bn
    linked by target "dnnmark_test_fwd_dropout" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_dropout
    linked by target "dnnmark_test_bwd_dropout" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_dropout
    linked by target "dnnmark_test_fwd_bypass" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_bypass
    linked by target "dnnmark_test_bwd_bypass" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_bypass
    linked by target "dnnmark_test_fwd_composed_model" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_fwd_composed_model
    linked by target "dnnmark_test_bwd_composed_model" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_bwd_composed_model
    linked by target "dnnmark_test_composed_model" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_composed_model
    linked by target "dnnmark_test_alexnet" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_alexnet
    linked by target "dnnmark_test_VGG" in directory /home/mnaderan/suites/DNNMark/benchmarks/test_VGG
GLOG_LIBRARY
    linked by target "dnnmark" in directory /home/mnaderan/suites/DNNMark

-- Configuring incomplete, errors occurred!

Invalid filter channel number - conv on AMD

Hello,

I'm trying to use the DNNMark benchmark to characterize my AMD GPU.
My setup is;

  • CentOS 7.5
  • ROCm 3.0 (with all the included libraries rocmblas and miopen)
  • AMD Vega 10 Frontier Edition

I'm able to successfully run all the layer tests, besides the convolution and dropout, with the provided files in config_examples.

When running the convolution (both forward and backward), I'm getting an error related to "Invalid filter channel number"

The convolution terminal output is the following:
I0428 13:03:25.984831 23359 test_bwd_conv.cc:11] DNNMark suites: Start... I0428 13:03:26.016577 23359 dnnmark.cc:226] Search and parse general DNNMark configuration I0428 13:03:26.016685 23359 dnnmark.cc:286] Search and parse layer configuration I0428 13:03:26.016721 23359 dnnmark.cc:300] Add [Convolution] layer I0428 13:03:26.016850 23359 dnn_layer.h:84] Layer name: conv1 I0428 13:03:26.016899 23359 dnn_layer.h:89] Previous layer: null I0428 13:03:26.016937 23359 dnnmark.cc:355] DNNMark: Initialize... I0428 13:03:26.016944 23359 dnnmark.cc:356] Running mode: 1 I0428 13:03:26.016950 23359 dnnmark.cc:357] Number of Layers: 1 I0428 13:03:26.016963 23359 dnnmark.cc:359] Layer type: 1 I0428 13:03:26.016968 23359 dnnmark.cc:361] DNNMark: Setup parameters of Convolution layer I0428 13:03:26.016974 23359 dnn_layer.h:110] Bottom dimension: N: 126 C: 3 H: 256 W: 256 I0428 13:03:26.017006 23359 data_manager.h:44] Create Data chunk of size 24772608 I0428 13:03:26.017105 23359 data_manager.h:101] Create data with ID: 0 I0428 13:03:26.017154 23359 data_manager.h:44] Create Data chunk of size 24772608 I0428 13:03:26.017248 23359 data_manager.h:101] Create data with ID: 1 I0428 13:03:26.017269 23359 data_manager.h:44] Create Data chunk of size 264241152 I0428 13:03:26.017419 23359 data_manager.h:101] Create data with ID: 2 I0428 13:03:26.017431 23359 data_manager.h:44] Create Data chunk of size 264241152 I0428 13:03:26.017577 23359 data_manager.h:101] Create data with ID: 3 I0428 13:03:26.017590 23359 data_manager.h:44] Create Data chunk of size 2400 I0428 13:03:26.017650 23359 data_manager.h:101] Create data with ID: 4 I0428 13:03:26.017660 23359 data_manager.h:44] Create Data chunk of size 2400 I0428 13:03:26.017670 23359 data_manager.h:101] Create data with ID: 5 I0428 13:03:26.046165 23359 data_manager.h:44] Create Data chunk of size 19660800 I0428 13:03:26.046329 23359 data_manager.h:101] Create data with ID: 6 I0428 13:03:26.046348 23359 conv_layer.h:220] Setting Bwd Filter Algo to I0428 13:03:26.048985 23359 data_manager.h:44] Create Data chunk of size 19660800 I0428 13:03:26.049124 23359 data_manager.h:101] Create data with ID: 7 I0428 13:03:26.050288 23359 data_manager.h:44] Create Data chunk of size 19660800 I0428 13:03:26.050415 23359 data_manager.h:101] Create data with ID: 8 I0428 13:03:26.050449 23359 dnnmark.cc:526] DNNMark: Running convolution backward: STARTED MIOpen Error: /root/driver/MLOpen/src/ocl/convolutionocl.cpp:122: Invalid filter channel number I0428 13:03:36.680456 23359 data_manager.h:53] Free Data chunk of size 19660800 I0428 13:03:36.680564 23359 data_manager.h:53] Free Data chunk of size 19660800 I0428 13:03:36.680598 23359 data_manager.h:53] Free Data chunk of size 19660800 I0428 13:03:36.680630 23359 data_manager.h:53] Free Data chunk of size 2400 I0428 13:03:36.680649 23359 data_manager.h:53] Free Data chunk of size 2400 I0428 13:03:36.680656 23359 data_manager.h:53] Free Data chunk of size 264241152 I0428 13:03:36.680737 23359 data_manager.h:53] Free Data chunk of size 264241152 I0428 13:03:36.680815 23359 data_manager.h:53] Free Data chunk of size 24772608 I0428 13:03:36.680847 23359 data_manager.h:53] Free Data chunk of size 24772608 MIOpen Error: 3 at /homelocal/fmendeslocal/DNNMark/core/include/dnn_utility.h983

When running the dropout (both forward and backward), the program is also trying to allocate an abnormal quantity of memory, that results in an error.

Dropout terminal output:
I0428 13:03:37.052639 23368 test_bwd_dropout.cc:14] DNNMark suites: Start... I0428 13:03:37.084635 23368 dnnmark.cc:226] Search and parse general DNNMark configuration I0428 13:03:37.084750 23368 dnnmark.cc:286] Search and parse layer configuration I0428 13:03:37.084781 23368 dnnmark.cc:300] Add [Dropout] layer I0428 13:03:37.084897 23368 dnn_layer.h:84] Layer name: dropout I0428 13:03:37.084964 23368 dnnmark.cc:355] DNNMark: Initialize... I0428 13:03:37.084971 23368 dnnmark.cc:356] Running mode: 1 I0428 13:03:37.084977 23368 dnnmark.cc:357] Number of Layers: 1 I0428 13:03:37.084990 23368 dnnmark.cc:359] Layer type: 8 I0428 13:03:37.084995 23368 dnnmark.cc:389] DNNMark: Setup parameters of Dropout layer I0428 13:03:37.085001 23368 dnn_layer.h:110] Bottom dimension: N: 100 C: 1000 H: 1 W: 1 I0428 13:03:37.085032 23368 data_manager.h:44] Create Data chunk of size 100000 I0428 13:03:37.085119 23368 data_manager.h:101] Create data with ID: 0 I0428 13:03:37.085168 23368 data_manager.h:44] Create Data chunk of size 100000 I0428 13:03:37.085182 23368 data_manager.h:101] Create data with ID: 1 I0428 13:03:37.085192 23368 data_manager.h:44] Create Data chunk of size 6065577211415516756 HIP Error at /homelocal/fmendeslocal/DNNMark/core/include/data_manager.h49 hipErrorMemoryAllocation I0428 13:03:37.085266 23368 data_manager.h:53] Free Data chunk of size 100000 I0428 13:03:37.085304 23368 data_manager.h:53] Free Data chunk of size 100000

Can somebody help me to understand what is wrong and how to fix it?

Thank you so much!

error: ‘gflags’ has not been declared

I've encountered an error when 'make'.

Scanning dependencies of target dnnmark
[  2%] Building CXX object CMakeFiles/dnnmark.dir/core/src/common.cc.o
[  4%] Building CXX object CMakeFiles/dnnmark.dir/core/src/dnn_config_keywords.cc.o
[  6%] Building CXX object CMakeFiles/dnnmark.dir/core/src/dnn_utility.cc.o
[  8%] Building CXX object CMakeFiles/dnnmark.dir/core/src/dnnmark.cc.o
[ 10%] Building CXX object CMakeFiles/dnnmark.dir/core/src/gpu_utility.cc.o
[ 13%] Building CXX object CMakeFiles/dnnmark.dir/core/src/utility.cc.o
Linking CXX shared library libdnnmark.so
[ 13%] Built target dnnmark
Scanning dependencies of target dnnmark_test_fwd_conv
[ 15%] Building CXX object benchmarks/test_fwd_conv/CMakeFiles/dnnmark_test_fwd_conv.dir/__/usage.cc.o
[ 17%] Building CXX object benchmarks/test_fwd_conv/CMakeFiles/dnnmark_test_fwd_conv.dir/test_fwd_conv.cc.o
In file included from /users/student/mr104/m104061597/DNNMark/DNNMark/benchmarks/test_fwd_conv/test_fwd_conv.cc:4:0:
/users/student/mr104/m104061597/DNNMark/DNNMark/benchmarks/test_fwd_conv/test_fwd_conv.cc: In function ‘int main(int, char**)’:
/users/student/mr104/m104061597/DNNMark/DNNMark/benchmarks/usage.h:32:1: error: ‘gflags’ has not been declared
 gflags::SetUsageMessage(\
 ^
/users/student/mr104/m104061597/DNNMark/DNNMark/benchmarks/test_fwd_conv/test_fwd_conv.cc:9:3: note: in expansion of macro ‘INIT_FLAGS’
   INIT_FLAGS(argc, argv);
   ^
make[2]: *** [benchmarks/test_fwd_conv/CMakeFiles/dnnmark_test_fwd_conv.dir/test_fwd_conv.cc.o] Error 1
make[1]: *** [benchmarks/test_fwd_conv/CMakeFiles/dnnmark_test_fwd_conv.dir/all] Error 2
make: *** [all] Error 2

Following is the output of apt search gflags:

Sorting... Done
Full Text Search... Done
libgflags-dev/trusty,now 2.0-1.1ubuntu1 amd64 [installed]
  commandline flags module for C++ (development files)

libgflags-doc/trusty 2.0-1.1ubuntu1 all
  documentation of gflags

libgflags2/trusty,now 2.0-1.1ubuntu1 amd64 [installed]
  commandline flags module for C++ (shared library)

python-gflags/trusty 1.5.1-1build1 all
  Python implementation of the Google command line flags module

python-google-apputils/trusty 0.4.0-1 all
  Google Application Utilities for Python

Issues with CMakeLists.txt when NVCC and HCC installed

I have a computer that has both NVCC and HCC installed on it, but uses an AMD GPU to run applications on. When I try to run setup.sh, it takes the CUDA path because CUDA_FOUND is true (and the if/else checks CUDA_FOUND before HCC_FOUND). This is a problem because it subsequently does not find any NVIDIA GPUs on my machine, and it never tries to install the AMD version subsequently.

For me, the fix was to change "if (CUDA_FOUND)" to "if(CUDA_FOUND AND NOT(HCC_FOUND))." I'm not sure if you all would want a more robust solution about what to do in general though. I'm happy to generate a pull request for this fix if that's easiest.

Matt

Segmentation fault

I want to run a benchmark with the following configuration file:

    [DNNMark]
    run_mode=composed

    [Convolution]
    name=conv1
    n=90
    c=3
    h=32
    w=32
    previous_layer=null
    conv_mode=convolution
    num_output=512
    kernel_size=3
    pad=1
    stride=1
    conv_fwd_pref=fastest
    conv_bwd_filter_pref=fastest
    conv_bwd_data_pref=fastest

    [BatchNorm]
    name=batchnorm1
    previous_layer=conv1
    batchnorm_mode=per_activation
    save_intermediates=true
    exp_avg_factor=0.5
    epsilon=1e-5

    [Activation]
    name=relu1
    previous_layer=batchnorm1
    activation_mode=relu

    [FullyConnected]
    name=fc7
    previous_layer=relu1
    num_output=10

I tried running it with two executables: benchmarks/test_bwd_conv/dnnmark_test_bwd_conv and test_bwd_composed_model/dnnmark_test_bwd_composed_model, but I always get segmentation fault:

~/DNNMark$ ./build/benchmarks/test_bwd_conv/dnnmark_test_bwd_conv -config conf_multiconv.dnnmark  -debuginfo 1
I0730 15:31:23.570518  1827 test_bwd_conv.cc:11] DNNMark suites: Start...
I0730 15:31:24.145496  1827 dnnmark.cc:226] Search and parse general DNNMark configuration
I0730 15:31:24.145601  1827 dnnmark.cc:286] Search and parse layer configuration
I0730 15:31:24.145622  1827 dnnmark.cc:300] Add [Convolution] layer
I0730 15:31:24.145772  1827 dnnmark.cc:300] Add [BatchNorm] layer
I0730 15:31:24.145834  1827 dnnmark.cc:300] Add [Activation] layer
I0730 15:31:24.145869  1827 dnnmark.cc:300] Add [FullyConnected] layer
I0730 15:31:24.145903  1827 dnnmark.cc:355] DNNMark: Initialize...
I0730 15:31:24.145906  1827 dnnmark.cc:356] Running mode: 2
I0730 15:31:24.145910  1827 dnnmark.cc:357] Number of Layers: 4
I0730 15:31:24.145915  1827 dnnmark.cc:359] Layer type: 1
I0730 15:31:24.145918  1827 dnnmark.cc:361] DNNMark: Setup parameters of Convolution layer
I0730 15:31:24.145921  1827 dnn_layer.h:106] Bottom dimension: N: 90 C: 3 H: 32 W: 32
I0730 15:31:24.145934  1827 data_manager.h:44] Create Data chunk of size 276480
I0730 15:31:24.146524  1827 data_manager.h:101] Create data with ID: 0
I0730 15:31:24.146550  1827 data_manager.h:44] Create Data chunk of size 276480
I0730 15:31:24.146695  1827 data_manager.h:101] Create data with ID: 1
I0730 15:31:24.146708  1827 data_manager.h:44] Create Data chunk of size 47185920
I0730 15:31:24.147017  1827 data_manager.h:101] Create data with ID: 2
I0730 15:31:24.147040  1827 data_manager.h:44] Create Data chunk of size 47185920
I0730 15:31:24.147356  1827 data_manager.h:101] Create data with ID: 3
I0730 15:31:24.147379  1827 data_manager.h:44] Create Data chunk of size 13824
I0730 15:31:24.147390  1827 data_manager.h:101] Create data with ID: 4
I0730 15:31:24.147394  1827 data_manager.h:44] Create Data chunk of size 13824
I0730 15:31:24.147404  1827 data_manager.h:101] Create data with ID: 5
I0730 15:31:24.159068  1827 dnnmark.cc:359] Layer type: 7
I0730 15:31:24.159081  1827 dnnmark.cc:385] DNNMark: Setup parameters of Batch Normalization layer
I0730 15:31:24.159093  1827 dnn_layer.h:157] Bottom dimension: N: 90 C: 512 H: 32 W: 32
I0730 15:31:24.159104  1827 data_manager.h:44] Create Data chunk of size 524288
I0730 15:31:24.160066  1827 data_manager.h:101] Create data with ID: 6
I0730 15:31:24.160074  1827 data_manager.h:44] Create Data chunk of size 524288
I0730 15:31:24.160274  1827 data_manager.h:101] Create data with ID: 7
I0730 15:31:24.160281  1827 data_manager.h:44] Create Data chunk of size 524288
I0730 15:31:24.160487  1827 data_manager.h:101] Create data with ID: 8
I0730 15:31:24.160495  1827 data_manager.h:44] Create Data chunk of size 524288
I0730 15:31:24.160661  1827 data_manager.h:101] Create data with ID: 9
I0730 15:31:24.160668  1827 data_manager.h:44] Create Data chunk of size 524288
I0730 15:31:24.160801  1827 data_manager.h:101] Create data with ID: 10
I0730 15:31:24.160809  1827 data_manager.h:44] Create Data chunk of size 524288
I0730 15:31:24.160948  1827 data_manager.h:101] Create data with ID: 11
I0730 15:31:24.160971  1827 data_manager.h:44] Create Data chunk of size 524288
I0730 15:31:24.161108  1827 data_manager.h:101] Create data with ID: 12
I0730 15:31:24.161116  1827 data_manager.h:44] Create Data chunk of size 524288
I0730 15:31:24.161255  1827 data_manager.h:101] Create data with ID: 13
I0730 15:31:24.161273  1827 data_manager.h:44] Create Data chunk of size 47185920
I0730 15:31:24.161540  1827 data_manager.h:101] Create data with ID: 14
I0730 15:31:24.161550  1827 data_manager.h:44] Create Data chunk of size 47185920
I0730 15:31:24.161805  1827 data_manager.h:101] Create data with ID: 15
I0730 15:31:24.161815  1827 dnnmark.cc:359] Layer type: 4
I0730 15:31:24.161818  1827 dnnmark.cc:373] DNNMark: Setup parameters of Activation layer
I0730 15:31:24.161823  1827 dnn_layer.h:157] Bottom dimension: N: 90 C: 512 H: 32 W: 32
I0730 15:31:24.161831  1827 data_manager.h:44] Create Data chunk of size 47185920
I0730 15:31:24.162084  1827 data_manager.h:101] Create data with ID: 16
I0730 15:31:24.162094  1827 data_manager.h:44] Create Data chunk of size 47185920
I0730 15:31:24.162348  1827 data_manager.h:101] Create data with ID: 17
I0730 15:31:24.162358  1827 dnnmark.cc:359] Layer type: 5
I0730 15:31:24.162360  1827 dnnmark.cc:377] DNNMark: Setup parameters of Fully Connected layer
I0730 15:31:24.162365  1827 dnn_layer.h:157] Bottom dimension: N: 90 C: 512 H: 32 W: 32
I0730 15:31:24.162371  1827 data_manager.h:44] Create Data chunk of size 900
I0730 15:31:24.162385  1827 data_manager.h:101] Create data with ID: 18
I0730 15:31:24.162390  1827 data_manager.h:44] Create Data chunk of size 900
I0730 15:31:24.162405  1827 data_manager.h:101] Create data with ID: 19
I0730 15:31:24.162410  1827 data_manager.h:44] Create Data chunk of size 5242880
I0730 15:31:24.162564  1827 data_manager.h:101] Create data with ID: 20
I0730 15:31:24.162572  1827 data_manager.h:44] Create Data chunk of size 5242880
I0730 15:31:24.162721  1827 data_manager.h:101] Create data with ID: 21
I0730 15:31:24.162744  1827 dnnmark.cc:550] DNNMark: Running FullyConnected backward: STARTED
Segmentation fault (core dumped)

My environment:
OS: Ubuntu 16.04.3 LTS
GPU: NVIDIA Quadro P2000
CUDA: 9.0
cuDNN: 7.0.5.15-1

Regards,
Peter

bwd_bypass kernel is empty on AMD HW

I've been looking at the disassembly of the bwd_bypass kernel when run with MIOpen on AMD HW, and I'm noticing that the Bwd kernel for this run is empty:

(literally the entire program)
MIOpenNeuronBwd:
s_endpgm

I've checked the resultant code on multiple generations of AMD HW and multiple versions of MIOpen, and all of them are generating similar kernels. Thus, I was wondering if this is intentional? Is it really expected that the Bwd pass does nothing?

Thanks,
Matt

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.