add elana

Author: hychiang-git

_includes/misc.html

@@ -1,12 +1,13 @@
 
 <div class="misc">
     <h2>Academic service</h2>
-    Conference Reviewer: NeurIPS (2024, 2025), ICLR 2025, ICML 2025, AAAI 2026
+    Conference Reviewer: NeurIPS (2024, 2025), ICLR (2025, 2026), ICML 2025, AAAI 2026
     <br><br>
     <h2>Open-source contributions</h2>
     <ul>
         <li><a href="https://github.com/facebookresearch/HolisticTraceAnalysis">HolisticTraceAnalysis</a> -- A PyTorch profiling tool by Facebook Research</li>
         <li><a href="https://github.com/Dao-AILab/fast-hadamard-transform">fast-hadamard-transform</a> -- An efficient Hadamard transform implementation from Dao-AILab (Tri Dao)</li>
+        <li><a href="https://github.com/enyac-group/Elana">Elana</a> -- Elana: A Simple Energy & Latency Analyzer for LLMs</li>
     </ul>
     <h2>Pro bono office hours</h2>
     I've committed to set aside 1 hour every week to provide mentorship and/or guidance for whoever is in need. If you find my research, work, oversea and/or life experience useful, welcoming to fill in this <a href="https://docs.google.com/forms/d/e/1FAIpQLScxlNtC1qnStXdmbGNCYNJGVOVdZgHAKeNEtzw5cPd3CqS6UA/viewform?usp=sf_link">form</a>.

_news/20250321_open_to_work.md

@@ -2,7 +2,7 @@
 layout: post
 date: 2025-03-21
 inline: true
-top: true
+top: false
 ---
 :briefcase: **<span style="color:red">I will be on the job market in late 2025 and actively seeking full-time research scientist positions</span>** <br/>
 Topics: Efficient Machine Learning, including optimization techniques for deep learning models, diffusion models, and large language models.

_news/20251211_elana_released.md

@@ -0,0 +1,7 @@
+---
+layout: post
+date: 2025-12-11
+inline: true
+---
+:computer: **<span style="color:red">Tool Released</span>** <br/>
+We release an open-sourced profiling tool: *ELANA: A Simple Energy and Latency Analyzer for LLMs*, please checkout on our [github](https://github.com/enyac-group/Elana/) and the [tenchnical report](https://arxiv.org/pdf/2512.09946)

_projects/elana.md

@@ -0,0 +1,122 @@
+---
+layout: page
+title: ELANA
+full_title: "ELANA: A Simple Energy and Latency Analyzer for LLMs"
+authors: Hung-Yueh Chiang, Bokun Wang, Diana Marculescu
+description: "Energy and Latency Analyzern for LLMs"
+img: assets/img/publication_preview/elana.png
+importance: 2
+category: research
+---
+
+<style>  
+li {  
+    font-size: 1.1rem; /* Adjust as needed */  
+}  
+</style>  
+
+<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/4.7.0/css/font-awesome.min.css">
+
+<div style="text-align: center; padding-bottom: 1rem;">
+<!-- <abbr class="badge" style="background-color:#00369f; margin-left:0.1rem; margin-right:0.1rem; font-size:1.1rem;">Arxiv</abbr> -->
+<abbr class="badge" style="background-color:#BF5700; margin-left:0.1rem; margin-right:0.1rem; font-size:1.1rem; width:80px; display:inline-block; text-align:center;">Arxiv</abbr>
+</div>
+
+<div class="authors"> 
+    <a href="https://hychiang.info">Hung-Yueh Chiang</a>,
+    <a href="https://bokun-wang.github.io">Bokun Wang</a>, 
+    <a href="https://users.ece.utexas.edu/~dianam/">Diana Marculescu</a>
+</div>
+<div class="authors">
+    The University of Texas at Austin
+</div>
+<div style="text-align: center; margin-top:12px;">
+    <a href="https://arxiv.org/pdf/2512.09946"><i class="fa fa-file-pdf-o" style="font-size:24px;color"></i><b> Report </b></a> &nbsp;
+    <a href="https://github.com/enyac-group/Elana/"><i class="fa fa-github" style="font-size:24px;color"></i><b> Code </b></a> &nbsp;
+</div>
+
+
+<br>
+<div style="text-align: center;">
+<p style="font-family: Comic Neue; font-size: 1.4rem;">
+    &#129303; Support models on Hugging Face, including Transformers, SSMs, and hybrid models <br>
+    &#128295; Support energy, latency, model size, KV cache size profiling <br>
+    &#128269; Support detailed kernel latency profiling <br>
+</p>
+</div>
+<div class="row mt-3">
+    <div class="col-sm-10 mt-3 mt-md-0 offset-1">
+        {% include figure.html path="assets/img/projects/elana/elana.png" title="elana logo" class="img-fluid rounded z-depth-1" %}
+    </div>
+</div>
+<br>
+
+# Compare to Zeus profiling framework
+<div class="row">
+    <div class="col-sm-10 mt-3 mt-md-0 offset-1">
+        {% include gif.html path="assets/img/projects/elana/compare.png" title="compare" class="img-fluid rounded z-depth-1" %}
+    </div>
+</div>
+<br>
+
+# Hugging Face interface
+<div class="row mt-3">
+    <div class="col-sm-10 mt-3 mt-md-0 offset-1">
+        {% include figure.html path="assets/img/projects/elana/hf_interface.png" title="hf interface" class="img-fluid rounded z-depth-1" %}
+    </div>
+</div>
+<br>
+
+# Profile model size
+<div class="row mt-3">
+    <div class="col-sm-10 mt-3 mt-md-0 offset-1">
+        {% include figure.html path="assets/img/projects/elana/model_size.png" title="model size" class="img-fluid rounded z-depth-1" %}
+    </div>
+</div>
+<br>
+
+
+
+# Profiling results
+### Profile models on A6000
+<div class="row mt-4">
+    <div class="col-sm-10 mt-4 mt-md-0 offset-1">
+        {% include figure.html path="assets/img/projects/elana/A6000.png" title="A6000 results" class="img-fluid rounded z-depth-1" %}
+    </div>
+</div>
+<br>
+
+### Profile models on Jetson series
+<div class="row mt-4">
+    <div class="col-sm-10 mt-4 mt-md-0 offset-1">
+        {% include figure.html path="assets/img/projects/elana/jetson.png" title="jetson results" class="img-fluid rounded z-depth-1" %}
+    </div>
+</div>
+<br>
+
+
+# Profiling kernels
+We use torch profile to generate a json trace file and visualize it with [Perfetto](https://ui.perfetto.dev/). 
+<div class="row mt-4">
+    <div class="col-sm-10 mt-4 mt-md-0 offset-1">
+        {% include figure.html path="assets/img/projects/elana/perfetto_kernel.png" title="kernel results" class="img-fluid rounded z-depth-1" %}
+    </div>
+</div>
+<br>
+
+
+# Citation
+{% raw %}
+```latex
+@article{chiang2025elana,
+  title = {ELANA: A Simple Energy and Latency Analyzer for LLMs},
+  author = {Chiang, Hung-Yueh and Wang, Bokun and Marculescu, Diana},
+  journal = {arXiv preprint arXiv:2512.09946},
+  year = {2025},
+}
+```
+{% endraw %}
+
+<br>
+# Acknowledgements
+This work was supported in part by the ONR Minerva program, NSF CCF Grant No. 2107085, iMAGiNE - the Intelligent Machine Engineering Consortium at UT Austin, UT Cockrell School of Engineering Doctoral Fellowships.