[libre-riscv-dev] IEEE754 FPU turning into ALU with Reservation Stations

Aleksandar Kostovic alexandar.kostovic at gmail.com
Thu Apr 4 06:39:47 BST 2019


okay, so that was funny. turns out I didn't exactly understand what to do,
despite looking and the pipeline_example.py file. Will wait to see how you
do get_a stage so I can do put_z similarly.

On Thu, Apr 4, 2019 at 7:18 AM Aleksandar Kostovic <
alexandar.kostovic at gmail.com> wrote:

> ok aleksander, you want to help morphing the fpmul.py code a little so
>> that it starts to look like PipelineStageObjectExample in
>> pipeline_example.py?
>
> yes, saw the code in the file. I think i know what to do. I'll start with
> put_z and you do get_op :)
>
> On Thu, Apr 4, 2019 at 7:09 AM Luke Kenneth Casson Leighton <lkcl at lkcl.net>
> wrote:
>
>> ok aleksander, you want to help morphing the fpmul.py code a little so
>> that it starts to look like PipelineStageObjectExample in
>> pipeline_example.py?
>>
>> where all the functions are being used, e.g. get_op and de_normalised
>> etc., all of those now need to *return* their arguments, instead of
>> performing an m.d.sync *inside* the function.
>>
>> the next phase will be: the m.d.syncs can then be replaced (all of
>> them) with assignments in each "with State" block, later, where
>> __setattr__ will over-ride (catch) the assignment and carry out the
>> m.d.sync *automatically*.
>>
>> however, all of the m.d.syncs need to be extracted to *outside* of the
>> functions, first.
>>
>> if i start at the top (get_op), you start at the bottom (put_z)?  i'll
>> work out how to transform get_op ok?
>>
>> l.
>>
>> _______________________________________________
>> libre-riscv-dev mailing list
>> libre-riscv-dev at lists.libre-riscv.org
>> http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
>>
>


More information about the libre-riscv-dev mailing list