hgyhungry / ge-spmm Goto Github PK

Hello, I am running your code for benchmarking. I found conv.c mentioned in data/download_SNAP.sh is missing in this repo. Can you put it in the repo or explain what it is for? Thanks a lot!

python3 gcn_custom.py --n-hidden=32
Traceback (most recent call last):
File "/home/jinyaox/ge-spmm/pytorch-custom/gcn_custom.py", line 6, in
from torch_geometric.datasets import Planetoid
File "/home/jinyaox/.local/lib/python3.10/site-packages/torch_geometric-1.4.3-py3.10.egg/torch_geometric/init.py", line 2, in
File "/home/jinyaox/.local/lib/python3.10/site-packages/torch_geometric-1.4.3-py3.10.egg/torch_geometric/nn/init.py", line 2, in
File "/home/jinyaox/.local/lib/python3.10/site-packages/torch_geometric-1.4.3-py3.10.egg/torch_geometric/nn/data_parallel.py", line 5, in
File "/home/jinyaox/.local/lib/python3.10/site-packages/torch_geometric-1.4.3-py3.10.egg/torch_geometric/data/init.py", line 5, in
File "/home/jinyaox/.local/lib/python3.10/site-packages/torch_geometric-1.4.3-py3.10.egg/torch_geometric/data/dataloader.py", line 5, in
ImportError: cannot import name 'container_abcs' from 'torch._six' (/home/jinyaox/.local/lib/python3.10/site-packages/torch/_six.py)

I followed all steps

Hi,

I followed the steps in the Readme file to try to complete your code. But I met the compilation error. The error message is below.

(base) ygong07@mira0:~/ge-spmm$ source compile.sh
spmm_test.cu(660): error: identifier "cusparseScsrmm2" is undefined

Could you tell me what I need to do to avoid these error and compile the code? Thank you.

Hello, I am running ge-spmm according to README. After "cd gunrock-test/build", there is no "spmm" when running the "make spmm -j8" instruction. How can I reproduce it?

I've been thinking about it(rid = blockDim.yblockIdx.x+threadIdx.y; cid = (blockIdx.y<<5)+threadIdx.x;)for a long time,But I still can't understand why this is.
As I have learned,usually,rid = blockDim.xblockIdx.x+threadIdx.x; cid = blockIdx.y*blockDim.y+threadIdx.y;
I hope you can explain why this is the case.
Thank you very much.

git clone --recursive https://github.com/hgyhungry/ge-spmm.git
For the following command, it ends with a fatal error
fatal: Failed to recurse into submodule path 'gunrock-test/gunrock'

The gunrock-test building process is not working.

Problem:

• makefile cannot be generated due to inconsistency between the cmakelist I provide and the one pulled from the gunrock repo
• users report that running test script cause 'invalid device function' exception. It is probably due to incorrect gencode flag during the building process.

"RuntimeError: /pytorch/torch/csrc/autograd/profiler_cuda.cpp:23: out of memory" happened when I ran "python gcn_dgl.py --gpu=0 --dataset=pubmed --n-hidden=128 --n-layers=1" .
The first run is successful and the rerun the second time has this problem.I seek for many solutions,but it has not yet been resolved.Highlight, I haven't changed this code.

If I wanted to run SpMM with a different aggregation/combiner operator, how can I do that? That seems to be the whole point of the paper but I can't see how the code enables that to the end user.

When I run "python3.8 gcn_custom_2layer.py --n-hidden=32", the following situation occurred:

Using /tmp/torch_extensions as PyTorch extensions root...
Detected CUDA files, patching ldflags
Emitting ninja build file /tmp/torch_extensions/spmm/build.ninja...
Building extension module spmm...
[1/3] c++ -MMD -MF spmm.o.d -DTORCH_EXTENSION_NAME=spmm -DTORCH_API_INCLUDE_EXTENSION_H -isystem /home/wjy/.local/lib/python3.8/site-packages/torch/include -isystem /home/wjy/.local/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -isystem /home/wjy/.local/lib/python3.8/site-packages/torch/include/TH -isystem /home/wjy/.local/lib/python3.8/site-packages/torch/include/THC -isystem /usr/local/cuda-10.1/include -isystem /usr/include/python3.8 -D_GLIBCXX_USE_CXX11_ABI=0 -fPIC -std=c++11 -c /home/wjy/Documents/ge-spmm-master/pytorch-custom/spmm.cpp -o spmm.o
[2/3] /usr/local/cuda-10.1/bin/nvcc -DTORCH_EXTENSION_NAME=spmm -DTORCH_API_INCLUDE_EXTENSION_H -isystem /home/wjy/.local/lib/python3.8/site-packages/torch/include -isystem /home/wjy/.local/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -isystem /home/wjy/.local/lib/python3.8/site-packages/torch/include/TH -isystem /home/wjy/.local/lib/python3.8/site-packages/torch/include/THC -isystem /usr/local/cuda-10.1/include -isystem /usr/include/python3.8 -D_GLIBCXX_USE_CXX11_ABI=0 -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr -gencode=arch=compute_61,code=sm_61 --compiler-options '-fPIC' -std=c++11 -c /home/wjy/Documents/ge-spmm-master/pytorch-custom/spmm_kernel.cu -o spmm_kernel.cuda.o
[3/3] c++ spmm.o spmm_kernel.cuda.o -shared -L/usr/local/cuda-10.1/lib64 -lcudart -o spmm.so
Loading extension module spmm...
Traceback (most recent call last):
  File "gcn_custom_2layer.py", line 9, in <module>
    from op import GCNConv
  File "/home/wjy/Documents/ge-spmm-master/pytorch-custom/op.py", line 6, in <module>
    spmm = load(name='spmm', sources=['spmm.cpp', 'spmm_kernel.cu'], verbose=True)
  File "/home/wjy/.local/lib/python3.8/site-packages/torch/utils/cpp_extension.py", line 670, in load
    return _jit_compile(
  File "/home/wjy/.local/lib/python3.8/site-packages/torch/utils/cpp_extension.py", line 877, in _jit_compile
    return _import_module_from_library(name, build_directory, is_python_module)
  File "/home/wjy/.local/lib/python3.8/site-packages/torch/utils/cpp_extension.py", line 1088, in _import_module_from_library
    return imp.load_module(module_name, file, path, description)
  File "/usr/lib/python3.8/imp.py", line 242, in load_module
    return load_dynamic(name, filename, file)
  File "/usr/lib/python3.8/imp.py", line 342, in load_dynamic
    return _load(spec)
ImportError: /tmp/torch_extensions/spmm/spmm.so: undefined symbol: cusparseCsr2cscEx2

Please help me how to run through next.

After running source compile.sh the fllowing error occured:

spmm_test.cu(660): error: identifier "cusparseScsrmm2" is undefined

But in the paper Efficient Sparse Matrix Kernels based on Adaptive Workload-Balancing and Parallel-Reduction the result is compared with cuda11.
Please update the code to fit cuda11.

Thankyou~