Just a suggestion to add a "--help" support to benchdnn.
As today benchdnn does nt seem to have a --help support, the doc being only available online on this page:
https://github.com/intel/mkl-dnn/tree/master/tests/benchdnn
the proposal is to add a minimum of "print usage" to the benchdnn CLI tool.

Environ

The docs generated for #431 at https://mratsim.github.io/Arraymancer/pca.html have broken formatting:

• Upstream bug on the type:
  
• Random bolding:
  and
  ![image](https:/

The problem is as follows:

Given an optimization model part which contains elements. Gid is specified as output format. If the elements of the optimization model part are not initialized (which may happen e.g. when no internal analyzer is defined), then the function GiDOutputProcess::ExecuteBeforeSolutionLoop will produce a seg-fault when it calls

`__initialize_results --> gid_io.h line 6

• If you discover a bug or regression in either the code or documentation, please report it. We appreciate your time to make the bug report.

  1. Please provide a small and self-contained program which exposes the bug. The small program should have a main() function and use only functions/classes

It seems like the parameter names used to control AMG and solver settings are pretty different for the MPI vs. non-MPI versions of the runtime AMGCL interface. For example, in examples/mpi/solve_mm_mpi.cpp, we have

prm.put("local.coarsening.type", coarsening);
prm.put("local.relax.type",      relaxation);
prm.put("isolver.type",          iterative_solver);

but these parameters lik

We worked very hard to get everything working in Travis. this demonstrates how to build and run everything, but it would be ideal if we translated this information into documentation for how to build the dependencies.

Description

We just switched to using Github Actions for our CI and it would be nice to add a new step that uses CMake to build the library and tests. The file can be found here:

• https://github.com/libocca/occa/blob/master/.github/workflows/build.yml

The fp_reduction_latency benchmarks were the very first benchmark, optimization and primitive code tested in Laser.

Unfortunately it is currently very confusing.

It should be reorganized:

• 1. Multiple accumulators: https://github.com/numforge/laser/blob/af191c086b4a98c49049ecf18f5519dc6856cc77/benchmarks/fp_reduction_latency/reduction_bench.nim and https://github.com/numforge/laser/blo

The pthreads based multi-threading code should be improved (IMHO). I see two options:

1. keep the internal 'explicit' interface more or less like it is but switch from pthreads to c++11 threads.
2. remove the 'explicit' interface, which should simplify the code significantly and switch to OpenMP where applicable.

I would strongly prefer the second option as is would reduce the complexity of

The Makefiles should define CXXFLAGS etc with ?= instead of = so that it doesn't matter if you set the variable on the left or right and side of the make invocation, i.e.:

VAR=val make
make VAR=val

the description in man/Pagoda2-class.rd is not sufficient (getKnnClusters and getEmbedding are not mentioned)? where can I get a complete description of Pagoda2?