How would you find the non-zero base in this situation?
pk267
@mario: the first element you come across that is not compressible with zero base becomes your non-zero base
ask
In the worst case, each element becomes a separate base and additionally, the offsets of 0 will also have to be added, thus increasing the size instead of compressing. Are there heuristics which just retain the original form in case the compression algorithm bloats the size for certain kinds of data?
firebb
Since there are multiple bases here, how can the hardware know which base to use for each value?
hzxa21
@firebb I think some meta bits are packed into the compressed data
How would you find the non-zero base in this situation?
@mario: the first element you come across that is not compressible with zero base becomes your non-zero base
In the worst case, each element becomes a separate base and additionally, the offsets of 0 will also have to be added, thus increasing the size instead of compressing. Are there heuristics which just retain the original form in case the compression algorithm bloats the size for certain kinds of data?
Since there are multiple bases here, how can the hardware know which base to use for each value?
@firebb I think some meta bits are packed into the compressed data