Stephen Hoffman<p><span class="h-card" translate="no"><a href="https://mastodon.social/@regehr" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>regehr</span></a></span> I don’t know how much got written about it, but Itanium fell over around its general differences from other compilers (meaning differing or unique optimizations and algorithms), being in-order (Poulson had some limited OOO), and the complete lack of visibility into memory access latencies available at compile time.</p><p>ILP is stupidly hard, too. A whole lot of the Itanium code I’ve looked at was just filled with NOPs. (There were multiple different types of NOP, too.)</p><p>Business-wise, Intel wanted a unique platform they controlled, and neither AMD nor particularly Microsoft liked that idea. HP, SGI, and limited support elsewhere wasn’t a big enough market, and the x86-32 performance intended for easing migration stank large. (Got lots of questions about that lack of performance, even though the particular OS wasn’t even using that.) </p><p>And Itanium was very late, and very slow when it arrived. Started in 1988 at HP, announced in 1997, and Merced in 2000. Merced was not at all speedy.</p><p>Itanium bet big on poor branch prediction performance, which was a problem back when itmwas designed (NetBurst being hot and bad at branch prediction) and that problem continued right up until Intel Core. Core was faster, less hot, and much better at branching.</p><p><a href="https://www.realworldtech.com/poulson/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="">realworldtech.com/poulson/</span><span class="invisible"></span></a></p><p><a href="https://devblogs.microsoft.com/oldnewthing/20150727-00/?p=90821" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">devblogs.microsoft.com/oldnewt</span><span class="invisible">hing/20150727-00/?p=90821</span></a></p><p>And some Itanium-contemporary books:</p><p>IA-64 Linux Kernel: Design and Implementation by David Mosberger and Stéphane Eranian, in the HP Professional Books series, Prentice Hall PTR, 2002, ISBN 0-13-061014-3</p><p>Itanium Architecture for Software Developers by Walter Triebel, Intel Press, 2000, ISBN 0-9702846-4-0</p><p>Programming Itanium-based Systems: Developing High Performance Applications for Intel's New Architecture by Walter Triebel, Joe Bissell, and Rick Booth, Intel Press, 2001, ISBN 0-9702846-2-4</p><p><a href="https://infosec.exchange/tags/retrocomputing" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>retrocomputing</span></a> <a href="https://infosec.exchange/tags/itanium" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>itanium</span></a> <a href="https://infosec.exchange/tags/itanic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>itanic</span></a></p>