Create accountLogin

Recent searches

No recent searches

Search options

Only available when logged in.
livellosegreto.it is one of the many independent Mastodon servers you can use to participate in the fediverse.
Livello Segreto è il social etico che ha rispetto di te e del tuo tempo.

Administered by:

Server stats:

1.2K
active users

livellosegreto.it: AboutProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.3.7

#AutomaticDifferentiation

0 posts0 participants0 posts today
Andreas<p>Have you ever thought 💡 of using JAX as 🧮 <a href="https://mathstodon.xyz/tags/automaticdifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>automaticdifferentiation</span></a> engine in 💻 finite element simulations? Boost the performance 🏇 of computationally-expensive hyperelastic material models with <a href="https://mathstodon.xyz/tags/jit" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>jit</span></a> in 🔍 FElupe! 🚀 🚀</p><p><a href="https://github.com/adtzlr/felupe" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="">github.com/adtzlr/felupe</span><span class="invisible"></span></a></p><p><a href="https://mathstodon.xyz/tags/python" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>python</span></a> <a href="https://mathstodon.xyz/tags/jax" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>jax</span></a> <a href="https://mathstodon.xyz/tags/finiteelementmethod" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>finiteelementmethod</span></a> <a href="https://mathstodon.xyz/tags/scientificcomputing" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>scientificcomputing</span></a> <a href="https://mathstodon.xyz/tags/computationalmechanics" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>computationalmechanics</span></a> <a href="https://mathstodon.xyz/tags/fea" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>fea</span></a> <a href="https://mathstodon.xyz/tags/fem" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>fem</span></a> <a href="https://mathstodon.xyz/tags/hyperelasticity" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>hyperelasticity</span></a></p>
Virgile Andreani<p>I really enjoyed the talk by Manuel Drehwald at <a href="https://fosstodon.org/tags/RustSciComp23" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>RustSciComp23</span></a> who drew the lines of an exciting future for <a href="https://fosstodon.org/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AutomaticDifferentiation</span></a> in <a href="https://fosstodon.org/tags/Rust" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Rust</span></a> with <a href="https://fosstodon.org/tags/LLVM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>LLVM</span></a> <a href="https://fosstodon.org/tags/Enzyme" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Enzyme</span></a> , which should be directly integrated into the compiler at an horizon of a couple of months.</p><p>If I understood correctly, the idea is to differentiate code at the LLVM IR level, *after optimization* (and to do another pass of optimization after that). This can produce faster code than the AD engines that operate at the source code level.</p>
Marco 🌳 Zocca<p>not to brag or anything but my ad-delcont library has been an inspiration to this :) </p><p><a href="https://github.com/konn/ad-delcont-primop" rel="nofollow noopener noreferrer" target="_blank"><span class="invisible">https://</span><span class="ellipsis">github.com/konn/ad-delcont-pri</span><span class="invisible">mop</span></a> </p><p>this is a line of work that uses delimited continuations to implement reverse-mode <a href="https://sigmoid.social/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AutomaticDifferentiation</span></a> , rather than reifing the program into a graph. As such, it enables a nice purely functional API and this latest incarnation performs pretty well too</p><p><a href="https://sigmoid.social/tags/machinelearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>machinelearning</span></a> <a href="https://sigmoid.social/tags/functionalprogramming" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>functionalprogramming</span></a> <a href="https://sigmoid.social/tags/haskell" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>haskell</span></a></p>
SearchLive feeds
About