<html><body><p>I think adding short float is a good idea, as long as it doesn't interfere with _Float16 where that's available and as long as it's either optional or can be implemented as binary32 where there's no hardware support for anything smaller (as section 4.1 allows).<br><br>ARM has 2 __fp16 types. One is binary16. One is not, having a slightly bigger exponent range instead of Infinities and NaNs (like the CELL SPUs).<br><br>Nvidia half is the same as binary16. I think half2 is a vector of 2 half s, getting 2x speed.<br><br>OpenEXR HALF is almost the same as binary16, except only supporting the default rounding mode.<br><br>In section 3.1 "<font size="5" face="LMMono12-Regular">_Float16</font><font size="4" face="NimbusSanL-Regu"> . . . The specification is general in that it can handle a wide range of lengths. However, it only permits an exact size.</font>"<br><br>That's true, but _Float16<b>x</b> allows any other representation with (1) IEEE semantics including Infinities and NaNs, (2) at least the exponent range of _Float16, and (3) at least the precision of _Float16. That might handle some of the other "half" precision variants listed in section 2.3. It might handle TI's 20 bit. It would not handle OpenGL's 10 or 11 bit (or 14, which I was unable to find in the OpenGL spec) and maybe not TI's 20 bit.<br><br>In section 2.1, Adobe also has 24 bit floating-point, not just 16.<br><br>In section 5.4, should promotion be to float except where float would be promoted to double? On many CPUs, "float f; short float sf; . . . f+sf" would be faster when evaluated as a float than as a double. Some CPUs/GPUs process half or half2 precision up to twice as fast as single precision, so is promotion necessary on them?<br><br>- Ian McIntosh IBM Canada Lab Compiler Back End Support and Development<br><br><br><img width="16" height="16" src="cid:1__=8FBBF5E5DFF4A7DE8f9e8a93df938690918c8FB@" border="0" alt="Inactive hide details for "Rajan Bhakta" ---2016-03-14 12:44:52 PM---FYI, this new document is being brought up for the next WG"><font color="#424282">"Rajan Bhakta" ---2016-03-14 12:44:52 PM---FYI, this new document is being brought up for the next WG14 meeting by another group.</font><br><br><font size="2" color="#5F5F5F">From: </font><font size="2">"Rajan Bhakta" <rbhakta@us.ibm.com></font><br><font size="2" color="#5F5F5F">To: </font><font size="2">cfp-interest@oakapple.net</font><br><font size="2" color="#5F5F5F">Date: </font><font size="2">2016-03-14 12:44 PM</font><br><font size="2" color="#5F5F5F">Subject: </font><font size="2">[Cfp-interest] Fw: ISO/IEC JTC 1/SC 22/WG 14 New documents available on ISOTC eCommittee</font><br><font size="2" color="#5F5F5F">Sent by: </font><font size="2">cfp-interest-bounces@oakapple.net</font><br><hr width="100%" size="2" align="left" noshade style="color:#8091A5; "><br><br><br>FYI, this new document is being brought up for the next WG14 meeting by another group.<font size="4"><br></font><br>One point I want to make is in 3.1 the _Float16x (which we do *not* specify in Part 3 of the TS) may be appropriate in some cases, but the "readability" argument and non-IEEE argument is still there.<font size="4"><br><br><br></font><br>Regards,<br><br>Rajan Bhakta<br>z/OS XL C/C++ Compiler Technical Architect<br>ISO C Standards Representative for Canada<br>C Compiler Development<br>Contact: rbhakta@us.ibm.com, Rajan Bhakta/Houston/IBM<font size="2" color="#800080"><br>----- Forwarded by Rajan Bhakta/Houston/IBM on 03/14/2016 10:08 AM -----</font><font size="4"><br></font><font size="2" color="#5F5F5F"><br>From: </font><font size="2">Daniel Plakosh <livelinkisotc@iso.org></font><font size="2" color="#5F5F5F"><br>To: </font><font size="2">Rajan Bhakta/Houston/IBM@IBMUS</font><font size="2" color="#5F5F5F"><br>Date: </font><font size="2">03/14/2016 09:18 AM</font><font size="2" color="#5F5F5F"><br>Subject: </font><font size="2">ISO/IEC JTC 1/SC 22/WG 14 New documents available on ISOTC eCommittee</font><font size="4"><br></font><hr width="100%" size="2" align="left" noshade><font size="4"><br><br></font><font face="Tahoma"><br>Dear member,<br><br>Please note that the following new documents have been posted<br>on the ISO/IEC JTC 1/SC 22/WG 14 eCommittee:<br><br></font><table border="0" cellspacing="0" cellpadding="0"><tr valign="top"><td width="49" bgcolor="#CAF0FE" valign="middle"><div align="center"><font face="Tahoma">N Number</font></div></td><td width="156" bgcolor="#CAF0FE" valign="middle"><font face="Tahoma">Title (Description)</font></td><td width="48" bgcolor="#CAF0FE" valign="middle"><font face="Tahoma">Exp. Action</font></td><td width="39" bgcolor="#CAF0FE" valign="middle"><font face="Tahoma">Due Date</font></td><td width="69" bgcolor="#CAF0FE" valign="middle"><font face="Tahoma">Version Date</font></td></tr>
<tr valign="top"><td width="49" bgcolor="#FFFFFF" valign="middle"><div align="center"><a href="http://isotc.iso.org/livelink/livelink?func=ll&objId=17597838&objAction=Open"><u><font color="#2F2F2F" face="Tahoma">2016</font></u></a></div></td><td width="156" bgcolor="#FFFFFF" valign="middle"><font color="#2F2F2F" face="Tahoma">Adding Fundamental Type for Short Float</font></td><td width="48" bgcolor="#FFFFFF" valign="middle"><font color="#2F2F2F" face="Tahoma">INFO</font></td><td width="39" bgcolor="#FFFFFF" valign="middle"><font color="#2F2F2F" face="Tahoma">None</font></td><td width="69" bgcolor="#FFFFFF" valign="middle"><font color="#2F2F2F" face="Tahoma">2016-03-14</font></td></tr>
<tr valign="top"><td width="361" bgcolor="#E0E0E0" colspan="5" valign="middle"><img width="1" height="1" src="cid:2__=8FBBF5E5DFF4A7DE8f9e8a93df938690918c8FB@" border="0" alt=""></td></tr>
<tr valign="top"><td width="49" bgcolor="#EEEEEE" valign="middle"><div align="center"><a href="http://isotc.iso.org/livelink/livelink?func=ll&objId=17589139&objAction=Open"><u><font color="#2F2F2F" face="Tahoma">2011</font></u></a></div></td><td width="156" bgcolor="#EEEEEE" valign="middle"><font color="#2F2F2F" face="Tahoma">ISO/IEC TS 17961 with TC1 Incorporated</font></td><td width="48" bgcolor="#EEEEEE" valign="middle"><font color="#2F2F2F" face="Tahoma">INFO</font></td><td width="39" bgcolor="#EEEEEE" valign="middle"><font color="#2F2F2F" face="Tahoma">None</font></td><td width="69" bgcolor="#EEEEEE" valign="middle"><font color="#2F2F2F" face="Tahoma">2016-03-14</font></td></tr>
<tr valign="top"><td width="361" bgcolor="#E0E0E0" colspan="5" valign="middle"><img width="1" height="1" src="cid:2__=8FBBF5E5DFF4A7DE8f9e8a93df938690918c8FB@" border="0" alt=""></td></tr></table><font face="Tahoma"><br><br>Download all documents as ZIP : </font><a href="http://isotc.iso.org/livelink/livelink?func=sspndocuments.zipdownload&nid=404333"><u><font color="#0000FF" face="Tahoma">ZIP-File</font></u></a><font face="Tahoma"><br><br>Access to ISO/IEC JTC 1/SC 22/WG 14 : </font><a href="http://isotc.iso.org/livelink/livelink?func=ll&objId=8919800&objAction=browse"><u><font color="#0000FF" face="Tahoma">Committee Homepage</font></u></a><font face="Tahoma"> <br><br><br>Please inform us if you have any difficulty accessing the above documents. For help with technical <br>problems please contact helpdesk@iso.org .<br><br>Best regards<br><br>dplakosh@cert.org<br></font><font size="4"><br>[attachment "ISO-IECJTC1-SC22-WG14_N2016_Adding_Fundamental_Type_for_Shor.pdf" deleted by Ian McIntosh/Toronto/IBM] </font><tt>_______________________________________________<br>Cfp-interest mailing list<br>Cfp-interest@oakapple.net<br></tt><tt><a href="http://mailman.oakapple.net/mailman/listinfo/cfp-interest">http://mailman.oakapple.net/mailman/listinfo/cfp-interest</a></tt><tt><br></tt><br><br><BR>
</body></html>