<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    Hey James,<br>
    <br>
    Feel free to ping me off list if you are wanting to bounce ideas off
    of someone!<br>
    <br>
    (For context - I wrote and maintain a high precision GPU math
    library at Codeplay as one of my many hats)<br>
    <br>
    Cheers,<br>
    -Neil.<br>
    <br>
    <div class="moz-cite-prefix">On 05/04/16 08:00, James Molloy via
      llvm-dev wrote:<br>
    </div>
    <blockquote
cite="mid:CALCTSA0cidDLgVwWb=mP-9LC25Zh_PW2As4hmLsq+OAYy5sFtQ@mail.gmail.com"
      type="cite">
      <div dir="ltr">Hi,
        <div><br>
        </div>
        <div>OK, writing an APFloat implementation of log and exp is a
          far more tractable problem. Thanks for the feedback all!</div>
        <div><br>
        </div>
        <div>James</div>
      </div>
      <br>
      <div class="gmail_quote">
        <div dir="ltr">On Tue, 5 Apr 2016 at 04:06 Chandler Carruth via
          llvm-dev <<a moz-do-not-send="true"
            href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a>>
          wrote:<br>
        </div>
        <blockquote class="gmail_quote" style="margin:0 0 0
          .8ex;border-left:1px #ccc solid;padding-left:1ex">
          <div dir="ltr">
            <div class="gmail_quote">
              <div dir="ltr">On Mon, Apr 4, 2016 at 10:41 AM Joerg
                Sonnenberger via llvm-dev <<a moz-do-not-send="true"
                  href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>>
                wrote:<br>
              </div>
              <blockquote class="gmail_quote" style="margin:0 0 0
                .8ex;border-left:1px #ccc solid;padding-left:1ex">On
                Mon, Apr 04, 2016 at 09:49:24AM -0700, Reid Kleckner via
                llvm-dev wrote:<br>
                > An optional MPFR dependency would also be pretty
                painful. I expect it will<br>
                > frequently be missing and will not be exercised by
                most buildbots.<br>
                <br>
                IMO if constant folding of transcendental functions
                makes a significant<br>
                difference for your program, you likely are doing
                something strange<br>
                already. I don't think it matters much for a lot of use
                cases, so having<br>
                an optional dependency for this seems to be fine.</blockquote>
              <div><br>
              </div>
            </div>
          </div>
          <div dir="ltr">
            <div class="gmail_quote">
              <div>I think that if transcendental functions are
                sufficiently unimportant, we shouldn't support them at
                all. We should have consistent behavior rather than
                ending up with nearly impossible to triage issues
                because one person's Clang used MPFR and the other
                didn't.</div>
              <div><br>
              </div>
              <div>While I think the optimizer should be able to
                constant fold transcendentals as a matter of core
                competency, I don't have very strong opinions about it
                for the reasons you cite -- I don't have use cases.</div>
            </div>
          </div>
          <div dir="ltr">
            <div class="gmail_quote">
              <div><br>
              </div>
              <div> </div>
              <blockquote class="gmail_quote" style="margin:0 0 0
                .8ex;border-left:1px #ccc solid;padding-left:1ex"> Note
                that the<br>
                non-transcendental functions are quite a different deal,
                especially<br>
                reasonable well behaving functions like log and exp.<br>
              </blockquote>
              <div><br>
              </div>
            </div>
          </div>
          <div dir="ltr">
            <div class="gmail_quote">
              <div>Agreed, and I feel very strongly that we should
                support both log and exp, and it has no business being
                optional.</div>
              <div><br>
              </div>
              <div>But clearly the right way to do this is to start from
                the bottom and work our way up. We should add the
                easiest ones first and incrementally add support. I
                particularly like the suggestions from Hal about how to
                make this a tractable engineering task.</div>
            </div>
          </div>
          _______________________________________________<br>
          LLVM Developers mailing list<br>
          <a moz-do-not-send="true"
            href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a><br>
          <a moz-do-not-send="true"
            href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev"
            rel="noreferrer" target="_blank">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><br>
        </blockquote>
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
      <pre wrap="">_______________________________________________
LLVM Developers mailing list
<a class="moz-txt-link-abbreviated" href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a>
<a class="moz-txt-link-freetext" href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a>
</pre>
    </blockquote>
    <br>
  </body>
</html>