Sorawee Porncharoenwase cefcb428db fix: make parser significantly more efficient ...

Prior this commit, `read-r1cs` copied bytes in the file over and over
again in a hot loop, causing the time complexity to be quadratic.
This commit switches to use index-based access in the hot loop instead,
resulting in a large performance improvement.
The benchmark file that accompanies this fix took 30 minutes to
successfully parse (according to @shankarapailoor).
This commit reduces the parsing time to 2 seconds.

It should be noted that not all bytes copying is avoided,
since bytes copying outside the hot loop, although not ideal,
does not really impact the performance.

2023-08-28 16:23:53 -07:00

..

TreeHasher.r1cs

fix: make parser significantly more efficient

2023-08-28 16:23:53 -07:00