Hi Marko, On 05/04/2017 03:01 PM, Marko Mäkelä wrote:

Hi Bar,

The purpose of this bit arithmetics (one could say CPU athletics) is to reduce the size of the InnoDB data dictionary cache. If we look at dtype_t and dict_col_t, these are not exactly at any byte boundary at the moment (61 and 84 bits, respectively).

I would simply replace the mbminmaxlen field with two 3-bit fields mbminlen, mbmaxlen. This would use up one more bit and would be faster to extract (no multiplication or division, just simple bit-shifting and masking). Also the code would be simpler, because a number of macros would be removed. I think that a similar patch was done in MySQL, but apparently it was never merged to the main trees.

Thanks for clarification. So it's about in-memory format. Then the things are even simpler. For some reasons I thought that the problem is with disk format, which is not desirable to change.

Marko

On Thu, May 4, 2017 at 1:34 PM, Alexander Barkov <bar@mariadb.org <mailto:bar@mariadb.org>> wrote:

Hello Marko,

InnoDB uses 5 bits to store the combination of mbminlen and mbmaxlen, and its calculated as:

mbmaxlen * 5 + mbminlen

Note, combinations with mbminlen=0 and mbmaxlen=0 are not possible. Also, combinations with mbminlen>mbmaxlen are not possible.

This is a summary:

mbminlen mbmaxlen 0 1 2 3 4 0 (0) (1) (2) (3) (4) 1 (5) 6 (7) (8) (9) 2 (10) 11 12 (13) (14) 3 (15) 16 17 18 (19) 4 (20) 21 22 23 24

I marked impossible numbers with parentheses.

In the range 0..24: - 10 pairs are valid - 15 pairs are not valid

It should be safe to use any impossible number to mean mbmaxlen=5 & mbminlen=1

It seems there is no really a need to extend the header to use separate 3 bits for both mbminlen and mbmaxlen.

-- Marko Mäkelä, Lead Developer InnoDB MariaDB Corporation