<div dir="ltr">I may have missed something, but since sv uses virtual simd when it needs to, what's the point of having mvl at all? VL should take care of everything.</div><br><div class="gmail_quote"><div dir="ltr">On Fri, Sep 28, 2018 at 4:36 PM Luke Kenneth Casson Leighton <<a href="mailto:lkcl@lkcl.net">lkcl@lkcl.net</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">---<br>
crowd-funded eco-conscious hardware: <a href="https://www.crowdsupply.com/eoma68" rel="noreferrer" target="_blank">https://www.crowdsupply.com/eoma68</a><br>
<br>
<br>
On Fri, Sep 28, 2018 at 9:14 PM, Jacob Lifshay <<a href="mailto:programmerjake@gmail.com" target="_blank">programmerjake@gmail.com</a>> wrote:<br>
<br>
> Setting csrs doesn't necessarily create a pipeline stall, at least it<br>
> doesn't in my rv32 core (probably not a good example).<br>
><br>
> <a href="https://github.com/programmerjake/rv32" rel="noreferrer" target="_blank">https://github.com/programmerjake/rv32</a><br>
<br>
yep krste corrected bruce on isa-dev. what i may do is have two<br>
CSRs, one to set MVL, the other to set VL. one of the issues with SV<br>
is that MVL (max vector length) is flexible where in RVV it's a<br>
hardware set/fixed limit.<br>
<br>
l.<br>
</blockquote></div>