Try to workaround a padding issue in Apple M1

[tkf]

Not sure if this is enough for ARM or Apple M1, but seems like an easy workaround.

close #35

[codecov]

Codecov Report

Merging #37 (9ae42e4) into master (4dcb9eb) will decrease coverage by 1.18%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #37      +/-   ##
==========================================
- Coverage   81.97%   80.79%   -1.19%     
==========================================
  Files          12       12              
  Lines        1337     1338       +1     
==========================================
- Hits         1096     1081      -15     
- Misses        241      257      +16     

Flag	Coverage Δ
unittests	80.79% <100.00%> (-1.19%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/dlcrq.jl	74.06% <ø> (-1.16%)	⬇️
src/lcrq.jl	76.16% <ø> (-4.07%)	⬇️
src/utils.jl	72.32% <100.00%> (+0.24%)	⬆️
src/stack.jl	89.28% <0.00%> (-7.15%)	⬇️
src/dict.jl	85.25% <0.00%> (-1.29%)	⬇️
src/msqueue.jl	85.41% <0.00%> (+6.25%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4dcb9eb...9ae42e4. Read the comment docs.

[anandijain]

thanks for investing so much time into this. I feel a bit bad because I'm mainly just testing it out for curiousity/fun.

tests pass locally!

[tkf]

tests pass locally!

That's good to hear! ARM is actually an interesting target since it has weaker memory model than x86. So I was curious if it even works. There's some "sloppy" parts (not fully C++20 memory model-compliant) in this package since I'm pushing things a bit beyond what current Julia atomics provide.