<p>wow, i never realized how much of a piece of shit Ross Ulbricht is</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> </p><p>very aside but my &quot;ounce of prevention is worth more than pound of cure&quot; fix for this, using a garbage collector, has not proved as popular as i hoped</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@tedmielczarek" class="u-url mention">@<span>tedmielczarek</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> I think those might have actually influenced my decisionmaking!</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> yeah I made stickers about that (+ a website since I can&#39;t hand out stickers over the internet): <a href="https://keepcalmandcallclone.website/" target="_blank" rel="nofollow noopener" translate="no"><span class="invisible">https://</span><span class="">keepcalmandcallclone.website/</span><span class="invisible"></span></a></p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> this is a *really* hot loop and the last time I implemented a language interpreter the speed was truly abysmal so this time I&#39;m allowing myself to think about optimization a *little* early</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> like, Rust programmers have this shared flaw of obsessing over small costs which in many cases are simply not measurable but create a very real cost to implement</p><p>I benchmarked SmallVec in prjunnamed earlier and every attempt to add that optimization made the runtime worse despite less malloc&#39;ing</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> Thank you for the responses</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> I have absolutely no idea in this case. I worked on a TCP/IP stack and on an FPGA toolchain and neither has any unsafe code for performance (or other reasons) so I would personally simply not worry about it</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> Cell does not let you borrow things, it just copies them (hence why it needs to be Copy)</p><p>reading Rust generated LLVM IR or assembly is kinda horrific so I would suggest benchmarking it instead</p>