<div dir="ltr">I plan to get rid of the cold attribute in a future patch.<div><br></div><div>Thanks,</div><div>Easwaran</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Dec 9, 2015 at 11:37 PM, Justin Bogner <span dir="ltr"><<a href="mailto:mail@justinbogner.com" target="_blank">mail@justinbogner.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Easwaran Raman <<a href="mailto:eraman@google.com">eraman@google.com</a>> writes:<br>
> Justin and Duncan, does this patch look good or do you have any concerns?<br>
<br>
It bothers me a bit that this doesn't actually remove the special cases<br>
in profiling code, since they still need to deal with the "cold"<br>
attribute. I kind of feel like there's a more general way to handle all<br>
of this, but I'm okay with this as an incremental step. Thanks!<br>
<div class="HOEnZb"><div class="h5"><br>
> On Mon, Dec 7, 2015 at 10:31 AM, Justin Bogner <<a href="mailto:mail@justinbogner.com">mail@justinbogner.com</a>> wrote:<br>
>> Easwaran Raman <<a href="mailto:eraman@google.com">eraman@google.com</a>> writes:<br>
>>> On Fri, Dec 4, 2015 at 4:13 PM, Justin Bogner <<a href="mailto:mail@justinbogner.com">mail@justinbogner.com</a>> wrote:<br>
>>>> Moving this out of the frontend to here makes sense, but I have a couple<br>
>>>> of questions.<br>
>>>><br>
>>>> - Does this only apply when we've supplied actual profile data, or can<br>
>>>> we hit it based off of the heuristic counts in no-profile-data mode?<br>
>>>> If the latter, this isn't quite NFC.<br>
>>><br>
>>> It applies only when there is profile data since entry count and maximum<br>
>>> function count come only from profile data.<br>
>>><br>
>>>> - Can the same be done for the "cold" attribute the frontend also<br>
>>>> applies?<br>
>>><br>
>>> The difference between the inlinehint and cold attributes is that<br>
>>> the latter is used not just by the inliner. For instance, branch<br>
>>> probability info. I can move it to the backend and replace all calls<br>
>>> to hasFnAttr(Attribute::Cold) with a wrapper that also returns true<br>
>>> if the count is below the threshold. Do you want that change to be<br>
>>> in this patch?<br>
>><br>
>> That sounds more invasive, so it probably makes sense to consider it<br>
>> separately.<br>
</div></div></blockquote></div><br></div>