Merge Split Branch

preview-update-submodules/ASP.html

@@ -246,7 +246,7 @@ <h4 class="mt-0">
   <figcaption class="figure-caption">Dr. Severini helping a student during ASP2022.<br /></figcaption>
 </figure>
 
-<p>Almost 200 students from 41 countries were selected to participate in the <a href="https://www.africanschoolofphysics.org/asp2022/">7th ASP 2022</a> at <a href="https://science.mandela.ac.za/ASP-2022">Nelson Mandela University in Gqeberha, South Africa</a>. With the school being shortened to two weeks, a parallel learning system was implemented, where participants could choose lectures to attend to improve their educational growth. <a href="https://www.nhn.ou.edu/~hs/">Dr. Horst Severini</a> is a Research Scientist and Adjunct Professor in <a href="http://www-hep.nhn.ou.edu/">High Energy Physics</a> and <a href="http://it.ou.edu/">Information Technology</a> from the University of Oklahoma (OU) and a co-leader of the high-performance computing workshop. He anticipated maybe 25 students attending on his track, “…we had about that many laptops,” he remarked, “and then we ended up with over 50 students!”</p>
+<p>Almost 200 students from 41 countries were selected to participate in the <a href="https://www.africanschoolofphysics.org/asp2022/">7th ASP 2022</a> at Nelson Mandela University in Gqeberha, South Africa. With the school being shortened to two weeks, a parallel learning system was implemented, where participants could choose lectures to attend to improve their educational growth. <a href="https://www.nhn.ou.edu/~hs/">Dr. Horst Severini</a> is a Research Scientist and Adjunct Professor in <a href="http://www-hep.nhn.ou.edu/">High Energy Physics</a> and <a href="http://it.ou.edu/">Information Technology</a> from the University of Oklahoma (OU) and a co-leader of the high-performance computing workshop. He anticipated maybe 25 students attending on his track, “…we had about that many laptops,” he remarked, “and then we ended up with over 50 students!”</p>
 
 <p>Severini was first introduced to distributed computing during his postdoc at OU. Then in the spring of 2012, Severini was introduced to Kétévi Assamagan, one of the founders of the ASP. Assamagan met with Severini and invited him and his colleagues to participate, leading to a  scramble to create a curriculum for this new lecture series. They were eager to show students how distributed computing could help with their work.</p>
 

preview-update-submodules/CLAS12.html

@@ -352,7 +352,7 @@ <h2 id="enhancing-science-and-improving-workflows">Enhancing science and improvi
   <figcaption class="figure-caption"> <br />Maurizio Ungaro, Jefferson Lab Staff Scientist.</figcaption>
 </figure>
 
-<p>When asked about workflows and job submissions, <a href="https://gemc.jlab.org/gemc/html/index.html">Maurizio Ungaro</a>, a Jefferson Lab Staff Scientist who helps 
+<p>When asked about workflows and job submissions, Maurizio Ungaro, a Jefferson Lab Staff Scientist who helps 
 coordinate CLAS12’s monte-carlo simulations, expresses: <strong>“This is actually where OSG services are really useful. Containers allow us to encapsulate the 
 software that we run, and HTCondor coordinates the submission of our jobs. Because of this, we’re able to solve two problems: one being CPU usage, and 
 the other being simulation organization.”</strong></p>

preview-update-submodules/Fulvio.html

@@ -248,7 +248,7 @@ <h4 class="mt-0">
 
 <p>If you could use a method of analysis that results in better insights into your research, you’d want to use that option. The catch? It can take months to analyze one set of data.</p>
 
-<p><a href="https://postlab.psych.wisc.edu/people/west/jacqueline-fulvio1/">Jacqueline M. Fulvio</a>, a research scientist for the <a href="https://postlab.psych.wisc.edu/">Postle Lab</a> at the University of Wisconsin-Madison, explained at <a href="https://agenda.hep.wisc.edu/event/1733/timetable/?view=standard">HTCondor Week 2022</a> how she overcame this problem using high throughput computing (HTC) in her analysis of neural oscillations’ role in visual working memory.</p>
+<p>Jacqueline M. Fulvio, a research scientist for the <a href="https://postlab.psych.wisc.edu/">Postle Lab</a> at the University of Wisconsin-Madison, explained at <a href="https://agenda.hep.wisc.edu/event/1733/timetable/?view=standard">HTCondor Week 2022</a> how she overcame this problem using high throughput computing (HTC) in her analysis of neural oscillations’ role in visual working memory.</p>
 
 <p>The Postle Lab analyzed the patterns of brain waves recorded from participants as they performed working memory tasks using an HTC workflow. Visual working memory is the brain process that temporarily allows us to maintain and manipulate visual information to solve a task. First, participants were given a sample with two images to memorize for two seconds. Then the image disappeared, and, following a five-second delay, participants were given a cue that indicated which item in memory would later be tested. The experimenter then delivered a single pulse of transcranial magnetic stimulation (TMS) to the participants’ scalp on half of the trials. TMS alters brain function, so Fulvio and her collaborators looked for corresponding impacts on participants’ brain waves recorded in an electroencephalogram (EEG). Finally, the participants indicated whether the image shown on the screen matched the original sample item.</p>
 

preview-update-submodules/Hiemstra.html

@@ -249,7 +249,7 @@ <h4 class="mt-0">
 
 <p>First, the researcher must write their ML code using a deep learning framework, such as PyTorch or TensorFlow.</p>
 
-<p>Second, the researcher needs to pick a GPU type. “You can run ML jobs on a normal server without GPUs, but certain machine learning processes (e.g., neural networks) run much faster if you use one,” notes Christina Koch, one of Hiemstra’s supervisors for his work. When using the HTCondor Software Suite, the researcher <a href="https://support.opensciencegrid.org/support/solutions/articles/5000653025-using-gpus-on-the-ospool">can choose a specific GPU type by specifying a CUDA compute capability</a> in the HTCondor job submit file.</p>
+<p>Second, the researcher needs to pick a GPU type. “You can run ML jobs on a normal server without GPUs, but certain machine learning processes (e.g., neural networks) run much faster if you use one,” notes Christina Koch, one of Hiemstra’s supervisors for his work. When using the HTCondor Software Suite, the researcher <a href="https://portal.osg-htc.org/documentation/htc_workloads/specific_resource/gpu-jobs/">can choose a specific GPU type by specifying a CUDA compute capability</a> in the HTCondor job submit file.</p>
 
 <p>Third, the researcher has to pick a CUDA runtime library. This library handles communication between the GPU and the application space, allowing the ML code to run its computations.</p>
 

preview-update-submodules/Lightning-Talks.html

@@ -354,7 +354,7 @@ <h4 class="mt-0">
 
 <p>Bandopadhyay’s project focuses on <a href="https://www.ligo.caltech.edu/">LIGO</a>’s ability to detect gravitational wave signals coming from binary neutron star mergers involving sub-solar mass component stars, which she determines from a graph which shows the detectability of the signals as a function of the component masses comprising the binary system.</p>
 
-<p>The fitting factors for the signals would have initially taken her laptop a little less than a year to run. She learned how to use <a href="https://osg-htc.org/services/open_science_pool.html">OSPool</a> capacity from the School, where it takes her jobs only 2-3 days to run. Other lessons that Bandopadhyay hopes to apply are data organization and management as she scales up the number of jobs. Additionally, she intends to implement <a href="https://htcondor.readthedocs.io/en/latest/users-manual/container-universe-jobs.html">containers</a> to help collaborate with and build upon the work of researchers in related areas.</p>
+<p>The fitting factors for the signals would have initially taken her laptop a little less than a year to run. She learned how to use <a href="https://osg-htc.org/services/open_science_pool.html">OSPool</a> capacity from the School, where it takes her jobs only 2-3 days to run. Other lessons that Bandopadhyay hopes to apply are data organization and management as she scales up the number of jobs. Additionally, she intends to implement <a href="https://htcondor.readthedocs.io/en/latest/users-manual/choosing-an-htcondor-universe.html#container-universe">containers</a> to help collaborate with and build upon the work of researchers in related areas.</p>
 
 <figure class="figure float-end" style="margin-left: 1em">
   <img src="https://raw.githubusercontent.com/CHTC/Articles/main/images/Meng-headshot.jpeg" height="200" width="200" class="figure-img img-fluid rounded" alt="Meng Luo" />

preview-update-submodules/Record.html

@@ -240,13 +240,13 @@ <h4 class="mt-0">
 
 <p>Nearly 60 projects from different fields contributed to the number of jobs processed during this record-breaking week, including these with substantial usage:</p>
 <ul>
-  <li>BioMedInfo: University of Pittsburgh PI <a href="https://www.wrightlabscience.com/p/index.html">Erik Wright</a> of the Wright Lab, develops and applies software tools to perform large-scale biomedical informatics on microbial genome sequence data.</li>
+  <li>BioMedInfo: University of Pittsburgh PI Erik Wright of the Wright Lab, develops and applies software tools to perform large-scale biomedical informatics on microbial genome sequence data.</li>
   <li>Michigan_Riles: University of Michigan PI <a href="https://gallatin.physics.lsa.umich.edu/~keithr/">Keith Riles</a> leads the Michigan Gravitational Wave Group, researching continuous gravitational waves.</li>
   <li>chemml: PI <a href="https://www.cmu.edu/chemistry/people/faculty/isayev.html">Olexandr Isayev</a> from Carnegie-Mellon University, whose group develops machine learning (ML) models for molecular simulations.</li>
   <li>CompBinFormMod: Researcher PI  <a href="https://hutchisonlab.org/">Geoffrey Hutchison</a> from the University of Pittsburgh, looking at data-driven ML as surrogates for quantum chemical methods to improve existing processes and next-generation atomistic force fields.</li>
 </ul>
 
-<p>Any researcher tackling a problem that can run as many self-contained jobs can harness the capacity of the OSPool.  If you have any questions about the Open Science Pool or how to create an account, please visit the <a href="https://support.opensciencegrid.org/support/solutions/articles/5000634384-frequently-asked-questions">FAQ page</a> on the OSG Help Desk website. <a href="https://opensciencegrid.org/projects.html">Descriptions of active OSG projects can be found here</a>.</p>
+<p>Any researcher tackling a problem that can run as many self-contained jobs can harness the capacity of the OSPool.  If you have any questions about the Open Science Pool or how to create an account, please visit the <a href="https://portal.osg-htc.org/documentation/overview/references/frequently-asked-questions/">FAQ page</a> on the OSG Help Desk website. <a href="https://opensciencegrid.org/projects.html">Descriptions of active OSG projects can be found here</a>.</p>
 
 <p>…</p>
 

preview-update-submodules/Science-Gateway.html

@@ -240,7 +240,7 @@ <h4 class="mt-0">
 
 <p><em>The UNL students involved in the capstone project, on graduation day. Order from left to right: Evan, Josh, Dan, Daniel, and Conner.</em></p>
 
-<p>When a science gateway built by a group of undergraduate students is deployed this fall, it will open the door for researchers to leverage the capabilities of advanced software and the capacity of the <a href="https://opensciencegrid.org/about/open_science_pool/">Open Science Pool</a> (OSPool). Working under the guidance of researcher <a href="https://chem.unl.edu/joseph-yesselman">Joe Yesselman</a> and longtime OSG contributor <a href="https://derekweitzel.com/">Derek Weitzel</a>, the students united advanced simulation technology and a national, open source of high throughput computing capacity –– all within an intuitive, web-accessible science gateway.</p>
+<p>When a science gateway built by a group of undergraduate students is deployed this fall, it will open the door for researchers to leverage the capabilities of advanced software and the capacity of the <a href="https://opensciencegrid.org/about/open_science_pool/">Open Science Pool</a> (OSPool). Working under the guidance of researcher Joe Yesselman and longtime OSG contributor <a href="https://derekweitzel.com/">Derek Weitzel</a>, the students united advanced simulation technology and a national, open source of high throughput computing capacity –– all within an intuitive, web-accessible science gateway.</p>
 
 <p>Joe, a biochemist, has been fascinated by computers and mathematical languages for as long as he can remember. Reminiscing to when he first adopted computer programming and coding as a hobby back in high school, he reflects: “English was difficult for me to learn, but for some reason mathematical languages make a lot of sense to me.”</p>
 

preview-update-submodules/california-megafires.html

@@ -241,7 +241,7 @@ <h4 class="mt-0">
 team examined how severe wildfires impact the habitats of over 600 species across two megafire 
 years in California. Ayars, a former Post-Baccalaureate Research Fellow at the Rocky Mountain 
 Research Station located in Colorado, investigated this impact on wildlife with Dr. <a href="https://www.fs.usda.gov/research/about/people/gavin.jones">Gavin Jones</a> 
-and Dr. <a href="https://peery.russell.wisc.edu/anu-kramer/">Anu Kramer</a> of the University of Wisconsin—Madison. 
+and Dr. Anu Kramer of the University of Wisconsin—Madison. 
 Their research was enabled by capacity provided by the <a href="https://chtc.cs.wisc.edu/">Center for High Throughput Computing (CHTC)</a> 
 and <a href="https://www.pnas.org/doi/10.1073/pnas.2312909120#sec-3">published</a> in the Proceedings of the National Academy of Sciences.</p>
 
@@ -274,7 +274,7 @@ <h4 class="mt-0">
 
 <p>“I would recommend anyone interested in using HTC [high throughput computing] or just curious about whether or 
 not it would be helpful for their work to reach out to CHTC,” she said. “It’s a great resource and they’re great 
-at teaching you.” Ayars gave a special shout-out to <a href="https://www.cs.wisc.edu/staff/lombardi-rachel/">Rachel Lombardi</a> and <a href="https://www.cs.wisc.edu/staff/koch-christina/">Christina Koch,</a> two RCFs at CHTC. Research computing facilitators
+at teaching you.” Ayars gave a special shout-out to Rachel Lombardi and <a href="https://www.cs.wisc.edu/staff/koch-christina/">Christina Koch,</a> two RCFs at CHTC. Research computing facilitators
 help new or potential users of CHTC services understand what computing resources are available on and off the UW-Madison 
 campus and provide support for adapting workloads to harness HTC.</p>
 

preview-update-submodules/fellowships/index.html

@@ -249,14 +249,14 @@ <h2 class="mt-1">Applications are open!</h2>
 
       <h2 id="featured-fellow">Featured Fellow</h2>
 
-      <div id="kristina-zhao" class="bg-white team-card justify-content-center mb-3 flex-grow-1">
+      <div id="pratham-patel" class="bg-white team-card justify-content-center mb-3 flex-grow-1">
     <div class="uw-card-content shadow">
         <div class="row">
             <div class="col-3">
                 <div class="overflow-hidden d-flex justify-content-center">
 
 
-                    <img src="/web-preview/preview-update-submodules/staff-list/images/kristina_zhao.jpg" alt="Headshot for Kristina Zhao" />
+                    <img src="/web-preview/preview-update-submodules/staff-list/images/pratham_patel.jpg" alt="Headshot for Pratham Patel" />
 
                 </div>
             </div>
@@ -265,32 +265,27 @@ <h2 id="featured-fellow">Featured Fellow</h2>
                     <div>
                         <h2 class="uw-mini-bar mb-0 mt-3">
 
-                            Kristina Zhao
+                            Pratham Patel
 
                         </h2>
-                        <div><h6 class="mt-2">Mentor(s):</h6> Emma Turetsky and Ian Ross</div>
+                        <div><h6 class="mt-2">Mentor(s):</h6> Brian Lin</div>
                     </div>
                 </div>
             </div>
             <div class="col-12">
-                <h4 class="mt-3">Integrating PyTorch and Pelican</h4>
+                <h4 class="mt-3">Enhancing container image build system</h4>
                 <div class="pt-2">
-                    <p>PyTorch is one of the most popular machine learning frameworks. 
-An important aspect of using it is the data engineering: how 
-is input data fed into the model during training? Going from 
-“tutorial scale” problems to cutting-edge research requires 
-drastically different techniques around data handling.</p>
-
-<p>For this project, we aim to better integrate Pelican 
-into the PyTorch community, providing both technical 
-mechanisms (implementing the fsspec interface for Pelican)
-and documentation by providing tutorials and recipes for 
-scaling PyTorch-based training using a combination of HTCondor 
-and Pelican.</p>
+                    <p>Container images are a widely used technology to package and distribute 
+software and services for use in systems such as Docker or Kubernetes.
+The PATh project builds hundreds of these images on a weekly basis but 
+the build system needs improvement to support more images and additional
+use cases. This project will focus on taking the existing system and 
+adding configurable, per-image build options. After a successful summer,
+the student fellow will gain skills in Docker containers, GitHub actions, and Bash.</p>
 
                 </div>
                 <div>
-                    <a class="btn btn-primary" href="./reports/2024/kristina-zhao">Read Final Report</a>
+                    <a class="btn btn-primary" href="./reports/2024/pratham-patel">Read Final Report</a>
                 </div>
             </div>
         </div>