<div dir="auto"><div><br><div class="gmail_extra"><br><div class="gmail_quote">On Aug 11, 2017 5:14 PM, "Kenneth Graunke" <<a href="mailto:kenneth@whitecape.org">kenneth@whitecape.org</a>> wrote:<br type="attribution"><blockquote class="quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="quoted-text">On Thursday, August 10, 2017 11:50:30 PM PDT Tapani Pälli wrote:<br>
> I do wonder what the target machine is (I haven't seen one that would<br>
> not have ARCH_X86_HAVE_SSE4_1 true, both 32bit and 64bit) but falling<br>
> back to memcpy makes perfect sense without USE_SSE4_1;<br>
><br>
> Reviewed-by: Tapani Pälli <<a href="mailto:tapani.palli@intel.com">tapani.palli@intel.com</a>><br>
<br>
</div>I agree - you really want SSE 4.1 support for decent performance on Atoms.<br>
The MOVNTDQA texture upload path, for example, gave us some really big<br>
performance gains on some ChromeOS workloads.<br>
<br>
So, while we should patch this, it'd still be worth figuring out if it's<br>
possible to enable it in Android-x86.<br>
<br>
--Ken</blockquote></div></div></div><div dir="auto"><br></div><div dir="auto">Hi, </div><div dir="auto"><br></div><div dir="auto">I checked on nougat-x86 build 32bit and the patches are Ok.</div><div dir="auto"><br></div><div dir="auto">The users of android-x86 have the option to build with <a href="http://silvermont.mk">silvermont.mk</a> and later</div><div dir="auto">to have an optimized build, but we currently try to support SSE3 devices.</div><div dir="auto">GNU/Linux distributions for desktop, like their kernel, may have that goal too.</div><div dir="auto"><br></div><div dir="auto">Thanks</div><div dir="auto">M.</div><div dir="auto"><div class="gmail_extra"><br></div></div></div>