thunderx: simplified code of 052591 PMC emulation for clarity #11759
Conversation
…nding - removed unneeded loop counters - removed unneeded multiplication/division operations - matched variable names in both functions
|
I have applied the changes requested |
| Y0 = m_pmcram[0]; | ||
| Y0 = (Y0 << 8) + m_pmcram[1]; | ||
| Y0 = (Y0 - 15) / 5; | ||
| Y1 = (m_pmcram[2] - 15) / 5; | ||
| const int e0 = (m_pmcram[0]<<8) | m_pmcram[1]; | ||
| const int e1 = m_pmcram[2]; |
There was a problem hiding this comment.
Note that this actually does change behaviour – previously it divided the Y1 value by 5 and then multiplied by 5 again later. If the actual value was not a multiple of 5, the remainder would be truncated. That’s no longer the case. It now uses the actual value from RAM directly.
There was a problem hiding this comment.
Yes, but data is accessed in groups of 5 later on. The new code makes more sense and is more parsimonius. I have not been able to find any problem in the gameplay after the change.
I do not know why the original author ended up writting those divisions and multiplications. They might be some leftover from all the guess work needed to emulate the chip or maybe he was trying to make sense of the program bytes written by the CPU.
Played the game for 20 minutes with no issues. Collision detection works the same.