Fixed rand_gmres.exe example

[shakedregev]

Description

./examples/rand_gmres.exe -i -m ~/ReSolve_dir/testcases/ACTIVSg200_AC/matrix_ACTIVSg200_AC_00.mtx -r ~/ReSolve_dir/testcases/ACTIVSg200_AC/rhs_ACTIVSg200_AC_00.mtx


Running randomized GMRES solver on CPU ...
[ERROR] Csr::syncData is trying to sync host, but host already up to date!
rand_gmres.exe: /home/szb/ReSolve_dir/ReSolve/resolve/matrix/Csr.cpp:336: virtual int ReSolve::matrix::Csr::syncData(ReSolve::memory::MemorySpace): Assertion `!h_data_updated_' failed.
Aborted

Closes #285

Mentions @(user)

Proposed changes

I eliminated a sloppy data sync that happens even when the code is running on CPU

Checklist

Put an x in the boxes that apply. You can also fill these out after creating
the PR. If you're unsure about any of them, don't hesitate to ask. We're here
to help! This is simply a reminder of what we are going to look for before
merging your code.

• All tests pass. Code tested on
  • CPU backend
  • CUDA backend
  • HIP backend
• Code compiles cleanly with flags -Wall -Wpedantic -Wconversion -Wextra.
• The new code follows Re::Solve style guidelines.
• There are unit tests for the new code.
• The new code is documented.
• The feature branch is rebased with respect to the target branch.

Further comments

#287 should be merged first to prevent the spurious error message. (Example runs but complains about sloppy synching.)