Make sure $(batch) is set to zero on CPU

[neworderofjamie]

Code using $(batch) previously didn't work on the single-threaded CPU backend. This PR fixes this by substituting zero for $(batch) in all CPU kernels

Fixes #513

[codecov]

Codecov Report

Merging #523 (a94754d) into master (93a50ca) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #523   +/-   ##
=======================================
  Coverage   86.96%   86.96%           
=======================================
  Files          84       84           
  Lines       18089    18091    +2     
=======================================
+ Hits        15731    15733    +2     
  Misses       2358     2358           

Impacted Files	Coverage Δ
src/genn/backends/cuda/backend.cc	82.75% <ø> (-0.02%)	⬇️
src/genn/backends/single_threaded_cpu/backend.cc	56.28% <100.00%> (+0.09%)	⬆️

---

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 93a50ca...a94754d. Read the comment docs.

[tnowotny]

... does this imply that models with model.batch_size = 1 will now run on the CPU backend (or where batch_size is not set by the user)?

[neworderofjamie]

If the reason they didn't run before was because you referenced $(batch) in your code and it didn't compile, yes they will!

[tnowotny]

Yes, this is very sensible. I will test if it allows to run my eventprop code on CPU-only by simply setting batch=1 once you've merged.