[Libre-soc-bugs] [Bug 1155] O(n^2) multiplication REMAP mode(s)

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=1155

--- Comment #38 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
svshape4 (daaamn combined with svshape3 that's a lot of opcode space)

modulo with offset priority

 945 | 0:5|6:9 |10|11:15  |16:20  | 21:24  | 25 | 26:31 |  Form      |
 946 | -- |----|--| ---   | ----- | ------ | -- | ------| --------   |
 947 | PO |offs|yx| XXXX  | SVd   | YYYY   | sk | XO    | SVM2-Form  |
 948 
 949 * svshape4 offs,yx,XXXX,SVd,sk,YYYY

by using MAXVL to indirectly compute one of the dimensions
it is possible although inconvenient to have quite a few bits
spare and cover all dimensions. perhaps even enough to set
z=1-4 as well... yes. maybe even some spare.

svshape3:

 814 | 0:5|6:10 |11:15  |16:20 | 21:25       | 26:31 |  Form    |
 815 | -- | --  | ---   | ---- | ----------- | ------| -------- |
 816 | PO | x   | y     | z/XXX| XX/yx/XX/sk | XO    |TODO-Form |

svshape4:

 945 | 0:5|6:9 |10|11:15  |16:20  | 21:24  | 25 | 26:31 |  Form      |
 946 | -- |----|--| ---   | ----- | ------ | -- | ------| --------   |
 947 | PO |offs|yx| XXXX  | SVd   | YYYY   | sk | XO    | SVM2-Form  |

fitting as closely with svshape2 also saves decoder gates.

hmmm... really, SVd needs to move to bits 6:10 across svshape2
svshape3 and svshape4.

damn damn damn. i really don't like making changes like this, but
it is what it s.

-- 
You are receiving this mail because:
You are on the CC list for the bug.