[libre-riscv-dev] daily kan-ban update 29may2020

[libre-riscv-dev] daily kan-ban update 29may2020

[Luke Kenneth Casson Leighton]

yesterday:

* continued review/assistance Tobias DataMerger
* continued guess-work on regfile formal proof
* started a compunit variant of test_pipe_caller for ALU.  discovered
that MultiCompUnit is not pipeline-compliant.

today:

* continue investigating non-compliance with pipeline API.  Cesar:
this *does not* affect what you are doing.

l.


---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68

_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

Re: [libre-riscv-dev] daily kan-ban update 29may2020

[Yehowshua]

Today:

Going to take a stab at one of the applications NLNet sent us for future grants.

Yehowshua
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

Re: [libre-riscv-dev] daily kan-ban update 29may2020

[Luke Kenneth Casson Leighton]

On Fri, May 29, 2020 at 2:51 PM Yehowshua <yimmanuel3@gatech.edu> wrote:
>
> Today:
>
> Going to take a stab at one of the applications NLNet sent us for future grants.

great.  can i suggest using the wiki -
https://libre-soc.org/nlnet_proposals/ - so that there is full
transparency, and also so that everyone can participate in reviewing
and writing the application?

l.

_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

Re: [libre-riscv-dev] daily kan-ban update 29may2020

[Yehowshua]

> 
> great.  can i suggest using the wiki -
> https://libre-soc.org/nlnet_proposals/ - so that there is full
> transparency, and also so that everyone can participate in reviewing
> and writing the application?
> 

Absolutely.

> l.
> 
> _______________________________________________
> libre-riscv-dev mailing list
> libre-riscv-dev@lists.libre-riscv.org
> http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev


_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

Re: [libre-riscv-dev] daily kan-ban update 29may2020

[Cesar Strauss]

Going along nicely, if slowly, with the parallel unit test of
src/soc/experiment/compalu_multi.py. Thankfully, the weekend is near, so
I expect to finally finish this task, and move along the queue.

Got a nice unplanned example of a race condition when, surprisingly,
running the test twice in a row passed, then failed, without changing
anything. Hint: when passing information from one co-routine to another,
do it around a synchronization point / barrier.

Also, learned how to use Settle() to respond asynchronously to outputs
of the DUT. It's useful for simulating a zero-delay response to a
request, or a purely combinatorial ALU.

Overall, I am quite enjoying the exercise.

Regards,

Cesar

_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

Re: [libre-riscv-dev] daily kan-ban update 29may2020

[Luke Kenneth Casson Leighton]

On Friday, May 29, 2020, Cesar Strauss <cestrauss@gmail.com> wrote:

> Going along nicely, if slowly, with the parallel unit test of
> src/soc/experiment/compalu_multi.py. Thankfully, the weekend is near, so
> I expect to finally finish this task, and move along the queue.
>
> Got a nice unplanned example of a race condition when, surprisingly,
> running the test twice in a row passed, then failed, without changing
> anything. Hint: when passing information from one co-routine to another,
> do it around a synchronization point / barrier.


yyep.

now you know why i put, in the TODO comments, to first wait for busy being
LOW in a loop, then for busy HI in a loop, and *then* proceed.




> Also, learned how to use Settle() to respond asynchronously to outputs
> of the DUT. It's useful for simulating a zero-delay response to a
> request, or a purely combinatorial ALU.


oh.  nice.  i will experiment with that.  i can think of several places i
should have used it.


> Overall, I am quite enjoying the exercise.


:)


>
> Regards,
>
> Cesar
>
> _______________________________________________
> libre-riscv-dev mailing list
> libre-riscv-dev@lists.libre-riscv.org
> http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
>


-- 
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

Re: [libre-riscv-dev] daily kan-ban update 29may2020

[Jacob Lifshay]

today: working on div and mul pipelines

probably won't get a lot done due to helping my friend build his computer
after finally getting the right motherboard. spent several hours only to
discover that the new mobo is defective.

Jacob
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

Re: [libre-riscv-dev] daily kan-ban update 29may2020

[Luke Kenneth Casson Leighton]

---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68

On Sat, May 30, 2020 at 12:20 AM Jacob Lifshay <programmerjake@gmail.com> wrote:
>
> today: working on div and mul pipelines

cool.

> probably won't get a lot done due to helping my friend build his computer
> after finally getting the right motherboard. spent several hours only to
> discover that the new mobo is defective.

aiyaaa :)  i hate it when that happens.  i've built... over... 20
systems in the past (including a 386, back in.... 1992-3 i think,
where i saved money by putting it into a small cardboard box).  i've
always been lucky that they always worked - except for those 8 systems
that had memory faults that needed a BIOS update.

good luck :)

l.

_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

Re: [libre-riscv-dev] daily kan-ban update 29may2020

[Luke Kenneth Casson Leighton]

On Fri, May 29, 2020 at 5:52 PM Cesar Strauss <cestrauss@gmail.com> wrote:

> Also, learned how to use Settle() to respond asynchronously to outputs
> of the DUT. It's useful for simulating a zero-delay response to a
> request, or a purely combinatorial ALU.

i can confirm: it does as intended.

https://git.libre-soc.org/?p=soc.git;a=commitdiff;h=cf573b0dfd3cb3f69dc28a162109dd92b0da72fd

so thank you for bringing that to everyone's attention, Cesar.  i was
previously writing tests that waited an entire cycle - as if they were
synchronous - and wondering what the hell was going on.

l.

_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev