Proposal

No annual report

Introduction

Proposal ID 2015-028
Submitted January 15, 2015
Owner pfaendtner
Department UW Information Technology
Category Remote Computing
Funding Status Partially Funded
Metric Score 3.68

Contacts

Primary
  • Name
  • Title
  • Email
  • Phone
  • Mailbox
  • Jim Pfaendtner
  • Assistant Professor
  • jpfaendt@uw.edu
  • 206-619-9515
  • 1750
Budget
  • Name
  • Title
  • Email
  • Phone
  • Mailbox
  • Cheryl DeBoise
  • Director of Business and Finance
  • cdeboise@uw.edu
  • 206.616.0208
  • cdeboise@uw.edu
Dean
  • Name
  • Title
  • Email
  • Phone
  • Mailbox
  • Michael Bragg
  • Dean, College of Engineering
  • mbragg@uw.edu
  • 206 543-1829
  • Box 352180

Descriptions

Abstract

This proposal will spark a new era of student access to research computing resources at the University of Washington. There are currently almost no research computing resources of any meaningful scale available to undergraduates at UW. This leaves them at a distinct competitive disadvantage compared to their peers at other leading peer institutions. We propose investing $1,002,098 to supplement the existing ~$12M already invested in the Hyak supercomputer and associated cyberinfrastructure (CI) in order to provide all students everywhere at UW with guaranteed access to more than 3,000 CPU cores delivering more than 70 teraflops of computing as well as an additional ~8,000 cores available through Hyak’s backfill queue. The deployment will also provide access to more than 200TB of working and archival data storage. This resource will be operated for four years with all operational and infrastructure expenses and capital costs (> $600k in direct costs) borne by UW-IT or covered by previous investments.

For more than a decade computational simulation, as enabled by high performance computing (HPC), has been recognized as the “Third Leg” of science and engineering research, alongside theory and experiment. More recently, Big Data has been acknowledged as an emerging “Fourth Leg”. In each case, the role of HPC in advancing research is critical. For many classes of research, the most interesting problems are experimentally intractable. In some cases, such as astrophysics, it is impractical to perform controlled experiments because the scale is too large, and in others, including some areas of nuclear physics or chemical engineering because the scale is too small. In others, researchers are interested in making time-sensitive predictions about future events – weather modeling, is a classic example.

In many cases science and engineering students must be prepared to use the tools of HPC to be effective in their academic careers, both as students and in their employment after graduation. Leading research universities, such as Georgia Tech and the University of Utah have recognized this and provide their students with access to cutting edge HPC facilities.

This proposal seeks to do the same at UW, building on nearly five years of experience by UW-IT with deploying and operating a large scale, shared cyberinfrastructure (the collection of computer hardware, software, and human expertise necessary to advance research) in the form of the Hyak supercomputer, lolo storage services, and High Speed Research Network and Science DMZ.

Category Justification

Hyak is operated as a remote compute resource. Users log into a head node (there are several) and submit work to the cluster via a scheduler. The scheduler supports interactive sessions allowing users to access to graphical/interactive applications as well as traditional, batch-oriented workloads.

Hyak is designed and operated in a fashion similar to leadership-class supercomputers managed by the NSF and DoE. The scheduler is the same, and best practices for managing workflow are similar. As a result, experience gained using Hyak is directly applicable for users hoping to migrate to larger supercomputers at other sites. Hyak and the lolo storage systems are directly connected to the recently upgraded 100 Gigabits per second (Gbs) High Speed Research Network (HSRN) and Science DMZ which supports data transfers to other supercomputer centers at speeds up to ten gigabytes per second (GBs).

In addition to providing a platform for developing applications and workflows for eventual execution on leadership-class systems, Hyak supports two workloads not well addressed by any other compute resource available to UW researchers:
- Big Data Pipelines
- The Speed of Science

Big Data Pipelines allow UW researchers to develop automated workflows in which data from high-throughput instruments, such as gene sequencers and electron microscopes, are processed by hundreds or thousands of CPU cores in near real time. The Speed of Science (also known as the Speed of Thought) describes a mode of computation in which researchers ask questions which exceed the capabilities of their workstations or laboratory-scale systems, in an ad-hoc fashion, interacting with a responsive supercomputer on the fly.

Background

Motivation and a campus-wide need for shared cyberinfrastructure:

This proposal is motivated by the fact that our undergraduate and graduate students at the UW do not have access to any scalable (i.e., beyond a few dozen cores) cyberinfrastructure (CI) or HPC to pursue independent research. Furthermore, faculty who teach cutting edge electives that involve computing do not have access to an HPC resource for their students to use on assigned projects. This crosscutting need stretches across dozens of departments (see letters of support), touches at least six of the colleges and professional schools at UW and hundreds, potentially even thousands, of students.
Several units within the UW (namely UW-IT, the office of the Provost and the colleges of A&S, Engineering, Environment, iSchool, Medicine, and UW Bothell) have already made significant investments over the past 5 years in a centralized campus CI/HPC totaling nearly $16 million. These investments provide all the infrastructure necessary to support high performance computing, as well as nearly 900 compute nodes currently deployed in Hyak, the lolo Archive and Collaboration data storage services, as well as advanced networks capable of transferring data between UW and remote compute and data resources at 100 gigabits per second.

Hyak operates as a condominium cluster -- a supercomputer management model growing in popularity at major research universities and pioneered at UW. In this model, users fund the purchase of nodes which are deployed in a shared infrastructure. When node owners, the students in this case, submit work to the machine they are guaranteed immediate access to their own nodes. And, when they require access to more computer power than they have funded, they may submit their work to a separate backfill queue which will execute on idle nodes anywhere in the cluster, regardless of who owns them. In this way Hyak is an elastic supercomputer, allowing participants to average 100% (or, in some cases more than 100%) utilization over time.

To date, Hyak has averaged 76% overall utilization over four and a half years of operation, delivering more than 130 million CPU hours of computing (nearly 15,000 years of computing if executed on a single CPU). By design, the Hyak infrastructure was engineered to support the replacement of nodes after three or four years of operation. The marginal cost of this replacement is the cost of the nodes alone. This allows the average infrastructure investment of $3,000/node to be spread over two generations of nodes. This proposal will replace 168 of the original Hyak nodes with new equipment delivering more than ten times the performance of the nodes they replace. Over the next year we anticipate the replacement of an additional 300 first-generation nodes or more, all funded by individual faculty participants in the condominium.

In addition to access to Hayk CPU resources, the system provides a rich software environment, including dozens of applications important in a variety of research domains. Examples include the Intel compiler suite, along with all important math and performance libraries, several high-level languages, including Python, R, and Mathematica (Matlab is coming!), and a variety of commercial engineering software funded by the College of Engineering, including Fluent, Ansys, and Abaqus.

Hyak operations are entirely supported by sponsorship from the participating colleges, UW-IT, and the eScience Institute. It is maintained by a team of professional IT engineers, and deployed in the secure, state of the art UW Tower data center. Hyak has experienced no significant unscheduled downtime in its entire operational history. There are no on-going expenses to the STF associated with this proposal.

The current state of computing resources on campus is ready for a large-scale investment by the STF to improve the experience of students. The prior investments in the infrastructure means most of the STF money will go directly toward increasing capacity (e.g., nodes and disks). The resource is already managed by UW-IT and can be easily expanded to accommodate this STF request. And finally, as detailed below the acquisition of this resource by the STF will initiate a large scale effort to improve outreach and training for all students to improve their basic skills and capabilities to use HPC for research and teaching purposes. These future efforts are detailed below in the “Benefits” section.

Specific details about the proposed budget:

This proposal will fund the purchase of 168 Hyak nodes, each with 20 CPU cores (3,360 total cores), along with 115TB of very fast scratch storage dedicated to student use (685GB/node). All nodes will be operated for four years, consistent with standard Hyak practice. Nodes will be supported by Lenovo for three years. In the fourth year UW-IT will make its best effort to ensure their continued operation in the event of a hardware failure. All storage components will be supported by the IBM/RAID Inc. for five years. The server, disk storage units, and GPFS software are all required for the delivery of storage to the system.

In addition to the storage funded through this proposal, for each Hyak node deployed UW-IT provides the following storage resources at no additional cost:

100GB of capacity in the Hyak fast scratch storage system
50GB of capacity in the lolo Collaboration filesystem (for sharing data outside Hyak)
500GB of capacity in the lolo Archive filesystem (for archiving Hyak results)

The Hyak cluster uses a homogenous infrastructure based on IBM/Lenovo Blade Center H technology and IBM/Lenovo storage hardware and software. All purchases are made using a sole-source justification. All Hyak storage is based on technology licensed from LSI by IBM and other vendors. IBM no longer offers the storage solution used to-date in Hyak, but the equipment specified here is an OEM version of the same unit sourced by another vendor. Efficient operation of Hyak depends on our ability to maintain, as near as possible, homogeneous equipment.

Purchases are tax exempt because more than 50% of overall Hyak utilization complies with the relevant state of Washington regulations.

It is likely that prices of equipment will change between the time this proposal is submitted and the budget becomes available, if funded. It is also likely that technology and options offered by the Hyak vendors will evolve. Before any purchases are made, we will review the system design and budget to ensure the best possible configuration is used.

Benefits to Students and the University

This section details some of the many benefits that will result from this STF investment.

1) Increased CI/HPC access to UW students. The increased computing capabilities will nucleate a new registered student organization (RSO) (provisional name: UW High Performance Computing Club), which co-PIs Pfaendtner and Li will serve as faculty advisors for. Two graduate students (Kayla Vanous and Patrick Lestrange) have agreed to provide the initial leadership for the group, which will serve the following roles:
- Management of outreach and signing up new students.
- Management of code repository for useful codes (provided by various PIs who use HPC on campus).
- As needed, manage allocations and fair distribution of CPU cycles. See further details below in the subsection ‘Access Restrictions’.
- Provide a basic set of usage and access instructions on a website for faculty interested in using the resource for their own teaching activities.
- Document success of the utilization of the resource and encourage student users to share their research success at venues like the Mary Gates Undergraduate Research Symposium.
- Provide a single resource for cataloging training and classes related to HPC such as the eScience institute’s offering of Python “boot camps” and CHEM 465 “Computations in Chemistry”

2) Leverage existing investments. Because this proposal piggybacks on existing infrastructure and UW-IT staffing, more of the money spent by STF will go towards actual equipment use by UW students.

3) Expand use of an already successful resource.

Hyak, the lolo storage service, and the Science DMZ have been in operation since July 2010. Over that entire period, utilization has exceeded 76%, providing more than 130M CPU hours.

This is possible because of Hyak’s condominium-style execution model which guarantees users immediate access to the nodes they fund while also allowing everyone access to idle CPUs throughout the cluster. The effect is to provide an elastic supercomputing resource in which users sometimes use very little of the capacity they have funded, while at other times they use far more. In this way, over time, groups are able to achieve 100% utilization despite having inevitably bursty demand.

When groups deploy HPC clusters separate from Hyak it is rare for them to achieve greater than 50% utilization. Moreover, they are limited to maximum throughput at any moment by the size of their stand-alone system. In a shared system such as Hyak average utilization over time can be much greater, and maximum throughput at any moment can be immense, making up for those times when a group’s utilization is low.

Departmental Endorsements

The endorsements in this section are from UW administrators and faculty members in various departments. Please take note of the high level university support from the Vice Provost for Research, Vice President for UW-IT, and Dean of Infrastructure from the College of Engineering.

Letter from Kelli Trosvig Vice President for UW IT

Members of the 2015 STF Committee:

It is my pleasure to endorse the proposal High Performance Computing for All UW Students.

Since its inception in 2010 the UW-IT supported cyberinfrastructure, including the Hyak supercomputer, lolo storage services, high performance research networks, and UW Tower data center have enabled cutting edge research. Hundreds of scholarly publications have been produced using these resources, and more than a dozen world class faculty have been recruited or retained by UW.

The success of our efforts is perhaps best measured by the willingness of the faculty and their departments to invest in Hyak with more than $8M of their own funds, supplemented by more than $4M from UW-IT, the provost, and the eScience Institute.

To date, however, Hyak access has been largely limited to roughly fifty leading computational labs and a thousand or so faculty, their staff, post docs, and students. With more than five years of experience behind us, UW-IT is ready to expand our reach to embrace the mission of providing HPC access to all students for their independent research, collaborations with faculty, and in their classes.

Toward that end, UW-IT is committing to sponsor the Hyak infrastructure and operational costs, associated with this proposal. These total more than $600k in capital commitments for Hyak infrastructure to be dedicated to this effort as well as operational expenses.

Thank you for the opportunity to partner with UW-IT to provide all UW students with the computational resources they need to be competitive in science and engineering, no matter their college or field of study.

Sincerely,
Kelli Trosvig

Letter from Prof. Mary Lidstrom, Vice Provost for Research of UW
Members of the 2015 STF Committee:

I am writing to enthusiastically endorse the proposal High Performance Computing for All UW Students. The capabilities funded through this grant will enable a new era of student research and collaboration at the University of Washington, making us a leader among our peers.

Beginning in 2007 the Office of Research (OR) has taken a leading role in promoting and supporting centrally managed high performance and research computing at UW. Based on extensive faculty input between 2007 and 2009, I committed OR and the fledgling eScience Institute to the task of developing and deploying a capable campus cyberinfrastructure (CI) by 2010. Working closely with UW-IT, we achieved this goal in the form of the Hyak supercomputer, lolo storage services, and the associated research focused computer networks.

In the four and a half years since beginning operations, Hyak and the associated CI have served a growing number of researchers in an increasing variety of disciplines, ranging from nuclear physics to social network analysis. As a Hyak user myself, and as chair of the Hyak Governance Board, I have witnessed this growth first-hand. As satisfying as this has been, it has become increasingly clear that the greatest unserved need for access to research computing at UW is among our students.

The capabilities provided by this proposal will provide an effective response to this need. The proposed student organization will involve students from across the university in outreach and education in the use of these powerful resources. This transparent, inclusive, and accessible approach to the problem of providing capable research computing is the model used by the Hyak Governance Board and has proven exceptionally successful at the faculty level. We have every hope that success will be repeated among our students.

I look forward to the next four years of research computing at UW. In partnership with the existing Hyak user community and the STF I expect great things!

Sincerely,
Mary Lidstrom

Letter from Prof. Dawn Lehman, Associate Dean for Infrastructure for UW Engineering

Dear Chance,

I am writing to express the college’s enthusiastic support of this STF proposal to extend and expand the use of this important high-computing resource, Hyak. As you know, computing touches every part of the academic mission and is important in all disciplines of engineering. To date, the college has relied on individual or limited cluster computing resources to support every student’s learning through teaching and research. However, as teaching and research demands more powerful computing resources, it becomes imperative that Hyak is more broadly available to our student population. I want to thank you and your leadership team for putting this proposal together. The College and its 10 departments (most of which have faculty and students using Hyak) look forward to integrating this critical resource more broadly in undergraduate and graduate classroom learning as well as to develop solutions to cutting-edge research problems.

Dawn Lehman
Associate Dean of Infrastructure
Professor of Civil and Environmental Engineering

Letter from Greg Miller (Professor and Chair of Civil and Env. Engineering)

Chance and Jim:
Please consider this as my endorsement of the STF proposal you sent me entitled “High Performance Computing for All UW Students”. Providing students with convenient access to larger scale computing resources beyond the typical standalone laptop/desktop class of machine would open numerous doors for expanded educational activities at all levels. In my department (Civil and Environmental Engineering), in particular, many aspects of our profession are driven by large scale computational models, and providing students hands on experience working with such models would be beneficial to their learning and their subsequent careers.

The Hyak model has worked well at the UW for supporting multiple scales of research computing, and this should map well to broader student use as proposed. The UW research community has helped to create an environment with a good suite of software packages amenable for student applications, and this can provide a low barrier to entry for a new user community.

Greg
__________________________________
Gregory R. Miller
Professor and Chair
Department of Civil & Environmental Engineering
201 More Hall
University of Washington 352700
Seattle, WA 98195-2700

Phone: +1 206-543-0350
Fax: 206-543-1543
email: gmiller@uw.edu

__

Letter from Adam Leache (Assistant Professor, Department of Biology)

Advances in genetics are occurring at an accelerating rate, with the interface between biological research and computer information technology becoming broader and more critical to student learning every year. Students need hands-on experience with the use of computational genetics and bioinformatics to become the next generation of scientists in virtually all fields of biology. Access by undergraduate and graduate students in my lab, and those of many other faculty in Biology, to the equipment requested here will greatly enhance the educational opportunities and research experiences of those students. I teach an upper division undergraduate and graduate level course in the Department of Biology called Applied Phylogenetics that teaches students how to analyze new genetic data to answer complex biological questions. The computing infrastructure requested in this proposal would accelerate the pace of learning in my course by removing some of the constraints that we have in place concerning the size and complexity of the data that we can analyze.

Letter from Prof. James Carothers (Chemical Engineering and MolES Institute)

Also, would love to be able to use Hyak for our cross-listed Synthetic Biology classes (Intro, Lab, Advanced) for teaching synthetic biological system modeling. These courses enroll >50 students from ChemE, EE, BioE, CSE and college of arts and sciences.

A little about the Advanced course — cut and paste if it helps. More details in Eduction tab at synbio.uw.edu

The course emphasizes approaches for modeling and designing systems-level functions in microbes and mammalian cells and teaches computational simulation, coarse-grained mechanistic modeling, and global sensitivity analysis. Quarter-long projects employ simulation analysis to design novel synthetic biological circuits, pathways and systems.

Advanced Synthetic Biology is the third class in a three-course synthetic biology sequence (Introduction, Laboratory Methods, Advanced). These interdisciplinary courses, organized through the UW Center for Synthetic Biology, are cross-listed in four different departments and have substantial enrollments (>50 per year) from students in engineering and Arts & Sciences. On-going collaboration between the teaching faculty generates curricula that are continually improved through the integration of new material and research discoveries in this fast changing field.

Letter from Dr. Sefa Dag, Materials Science and Engineering

Dear STF Committee:
I am a Research Scientist and lecturer at Materials Science and Engineering department of University of Washington. My core scientific research is focused on new development of computational simulation techniques and methodologies for designing efficient and low-cost materials in the biophysical and solar energy domains. I successfully manage this task for more than 10 years in this field and during this period of time, I was supported by nation’s most prestigious computing facilities, NERSC and NCCS. These facilities provided large-scale, state-of-the-art computing, storage, and networking for my research achievements.
Currently, I am giving new lecture series, “Computational Nano and Materials Science”, as a joint course, in Chemical Engineering, Molecular Engineering and Materials Science and Engineering departments in University of Washington. I am lecturing to 20 UW young scientists and students to be able to teach new simulation techniques and developments. However, one of our main issues in this course is shortness of computational resources and lack of massively parallel systems. As you can esteem, development of nanotechnology brought a new challenges in computational field and with this development, new high-end computing systems and comprehensive scientific client services have been established. As a research scientist and lecturer, it is inevitable for me to collaborate with scientifically productive supercomputing user facility for both research and teaching missions.
We certainly need newly developed primary computational provider of computing and data resources in in University of Washington. HYAK will play an increasingly important role to create breakthroughs in advanced scientific research and train new young researchers and teams who need to run simulations from thousands to millions of petascale simulations.
I am writing to express very strong support for the proposal as the additional computing resources would be beneficial to accelerate both scientific discovery at the exascale and train new research scientists and users for future computational needs of large community.
Thank you for your time and consideration of this request.
Sincerely,
Sefa Dag
NIH Research Fellow
Research Scientist, Materials Science and Engineering
University of Washington
sefadag@uw.edu
__

Letter from Keith Laidig from Baker group / Biochemistry

Folks,

I am writing in support of Professor Pfaendtner's 2105 STF proposal entitled "High Performance Computing for All UW Students", seeking investment in the UW-IT's HYAK computer as way to provide broad access to high performance computing (HPC) and large scale computing (LSC) to the UW student body.

I will leave it up to other to explain the value of HPC/LSC to UW students that hope to find themselves on the cutting edge of today's science and technology - (the value seems almost self-evident to me). Instead, I am writing this in support of HYAK as an efficient and cost-effective approach to this end. Building out HPC/LSC solutions is an expensive, painful and costly endeavor. I'm writing to tell you that the UW-IT's HYAK is possibly the best return on investment in the nation for a shared academic research computing model and is an outstanding value if you're looking to provide any sort of Big-<fill-in-the-blank> computing.

For the past 18 years, I've been in the business of designing, procuring, installing, configuring and managing HPC/LSC systems here at the UW. I've used all sorts of strategies to create computing systems for Professor David Baker's research group; building cluster from many desktops, stuffing dozens of white-boxes into closets and exhausting the heat into the ceiling spaces, building out small computer rooms within buildings with dedicated cooling and power, installing gear in professional datacenters off-campus, and stuffing the UW's datacenters full of roll-your-own solutions. In total, I've spent millions of Baker's research dollars to build HPC/LSC solutions and the infrastructure necessary to support them. For us, HYAK has proven a superb solution for much of our HPC computing needs - it is cheaper that using the "cloud" for us (it's the data upload/download/storage that kills us), it rivals the cost of us doing all of this ourselves (our IT groups has ~40 years combined of relevant expertise), and HYAK leverages the UW's world class networking infrastructure.

If you are considering an investment to provide access to HPC/LSC for the UW student body then I can assert that HYAK is the most cost-efficient solution you will find.

Yours,
Keith E. Laidig

__
Letter from Darwin Alonso,
I am writing in support of Professor Pfaendtner's 2105 STF proposal entitled "High Performance Computing for All UW Students", seeking investment in the UW-IT's HYAK computer as way to provide broad access to high performance computing (HPC) and large scale computing (LSC) to the UW student body.

Providing access to HPC for students in the sciences is of self evident value as both training for future work and as an invaluable aid to increasing their productivity at the University of Washington. Access to HYAK has been key to our students success in Dr David Baker's protein structure prediction and design lab, and I would like to see access provided to all students.

I would also like to point out the value to students in the arts. I've worked with my wife in the Ethnomusicology program in the School of Music over the last couple of years. They are in the process of archiving thousands of hours of exquisite and irreplaceable video and audio field recordings to the HYAK/Lolo archival storage. Providing this service for students would again be valuable as training and as a tool to aid in their research. Ethnomusicology is not currently using HYAK for video editing, but I see the day when that could be used to cut out the slow and CPU intensive steps in the production of sound and video recordings. This data rich processing is just the kind of work that is suited to be done in-house rather in the cloud.
Darwin Alonso
dovalonso@gmail.com

__
Dear STF committee,

My name is Ekaterina Badaeva. I am writing this letter to strongly support the HPC cluster application by professor Li.
I am currently a Materials and Processes engineer at Boeing Chemical Technology. I did my PhD in Professor Li’s group and have graduated in 2011 with the degree in computational/physical chemistry. While doing my PhD, I was involved in computational research and in teaching computational chemistry classes.
At professor Li’s group, I did research on predicting properties of magnetic semiconductor quantum structures, collaborating with professor Gamelin’s group. With the computational approach, we achieved several break-troughs in understanding materials properties and published several papers in highly ranked journals, including Nature. Modeling of these materials would not be possible without strong computational power that we had at that time. The nano-science research is evolving more in Dr. Li’s group, and the requirements to the computational modeling become more and more demanding, software becomes more advanced thus requesting enhanced computational power.
At Boeing I am working on a variety of projects related to the aerospace applications, many of those are computational modeling of materials. The modeling part that I do is directly related to the research that I did at professor Li’s group. Modeling properties of complex matter, such as polymers, semiconductors, fluids and metals requires very intense computational power. We are using several huge high performance computing clusters (thousands of cores) spread out through the United States. I may not express how valuable for me was the experience operating on the HPC cluster that I received when doing my research at UW! I also may not exaggerate how important the access to good computational power is for successful modeling predictions.

In addition, I would like to mention that access to enhanced computational resources is very important for teaching classes such as computational chemistry. I have been a TA for Dr. Li’s class, where we let students perform jobs on the UW clusters. Using HPC resources allowed students to complete studies of the more advanced and more realistic systems compared to what they would be doing on simple PC’s. As the result, students were more prepared to go into real industrial or academic research after they were done with their degrees. I personally received very positive feedback from the students who were able to work on the clusters.

At the Boeing Company, we are growing and hiring more computational chemists. One of the requirements to the related positions is experience operating on the HPC environments. I believe that enhanced computational resource is critical for success of the computational research and extremely important in teaching advanced classes, such as computational chemistry.

Thank you for your time and consideration of this request.

Sincerely,
Ekaterina.

Ekaterina A. Badaeva, PhD
BR&T | Chemical Technology | Polymer Science, Synthesis and Simulations
The Boeing Company
P.O. Box 3707 MC 19-LX
Seattle, WA 98124-2207
Ekaterina.a.badaeva@boeing.com
work phone: 425-237-5986
cell phone: 425-220-2060

Installation Timeline

Hyak is an established condominium-style cluster. Typical node deployments take less than six weeks from the time an order is placed with the vendor. This deployment should be no exception. Student accounts may be added at any time and typically take less than two days to provision.

Resources Provided by Department

Direct support for the STF investment Hyak is approximately $672k. This value is determined by the Hyak budget which shows infrastructure and operational costs to support a single node is $1,000/year. Infrastructure includes all system components aside from the compute node and supplemental storage. That is, this proposal will bear no costs associated with assembling the nodes into a cluster, their operation, or any of the software required to make effective use of the resource.

Overall supporting investments in Hyak exceed $20M. This includes $10M already invested in Hyak by its current users, funding, among other things, the ~900 nodes already in operation. In addition to direct investments in Hyak, contributions include advanced networking ($1.65M), data center capital costs (~$8M).

Access Restrictions (if any)

• Access to the STF-funded Hyak nodes and storage will be available to any student with a computational project that can use the resource. This will be managed by the UW High Performance Computing (HPC) Club.

• The HPC Club will manage allocations as needed. Priority high to low :
o Undergraduate and graduate students pursuing independent research projects
? No restriction, management as needed
o Undergraduate and graduate students supplementing research projects in labs they already work in: if they can justify the need for more expansive computing resources
o Instruction in HPC methods. Instructors wanting to access the STF Hyak allocation will submit their proposals to the HPC Club for evaluation.

Student Endorsements

There are 63 student endorsements pasted in the text below spanning dozens of departments and four UW colleges. Additionally, the Student Advisory Council of the College of Environment has submitted a letter of support.

Each of the 63 letters identifies unique and specific ways that the computing resource will be used to support student research on campus - it is very exciting!

Hi Jim,

Thank you for your thoughtful and detailed responses to our questions. After surveying the College of the Environment's Student Advisory Council (SAC), we would like to offer our support for the STF request for Hyak equipment.

We look forward to finding out more about how the CoE students can utilize these resources.

Best,
Jaci Saunders
CoE SAC co-chair
?
__
Dear STF Committee,
Endorsement for the New Computer Cluster Proposal
My name is Nuttavikhom Phanthuwongpakdee. I was an undergraduate student in the Department of Chemistry, which is part of the College of Arts and Sciences. My research involved the study of doping inorganic semiconductor quantum dots (QDs) with transition-metal ions. I used high-performance computing daily. Even though, I am not working on this topic, the knowledge, I had gained from this research, has helped me tremendously in my current research as a Ph.D. candidate at King’s College London and the National University of Singapore. I am writing to express very strong support for the proposal as the additional computation support would be beneficial in the following ways:
1. Faster processing time means that we can create more results;
2. More efficient computational support means better results with fewer errors;
3. Keeping up to date with the latest technology to provide the highest quality research.
Thank you for your time and consideration of this request.
Sincerely,
Nuttavikhom Phanthuwongpakdee
nuttavikhom.phan@kcl.ac.uk
nuttavikhom@nus.edu.sg
__

Dear STF committee,

My name is Kevin Epperly. I am an employee at the Burke Museum. I work with grad and undergrad students in the Klicka and Leache Labs which are part of the Department of Biology in the College of Arts and Sciences. These students' research include the phylogenetic and phylogeographic study of birds, amphibians, and lizards. We use high performance computing daily. I writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways:
Allowing students to de-mulitiplex larger genetic data-sets obtained through next generation sequencing techniques such as RADseq and Sequence Capture.
Performing computational heavy bioinformatic analysis of said data.
Developing and testing new models and analysis for these large data-sets.

Thank you for your time and consideration of this request.
Sincerely,
Kevin Epperly
kepperly@uw.edu
__
Dear STF Committee,
As a graduate student in the chemical engineering department at the University of Washington, the accessibility to computational resources has directly shaped and enhanced my career through increased papers, collaborations, and research. During my MS degree, I studied organic photovoltaics and investigated new materials. An important question for this field is “why some materials work better than others?” My group was puzzled over drastic differences in solar cell efficiency from two seemingly similar materials. STF-funded computational resources allowed me to calculate and model photovoltaic energy conversion with high level density functional theory. This yielded unexpected differences between the materials providing a direct answer to this challenging question, which was the centerpiece of a recent paper published in J. Am. Chem Soc (Beyond Fullerenes: Design of Nonfullerene Acceptors for Efficient Organic Photovoltaics). Results from STF resources were essential to this publication. Moreover, these computational resources have proliferated throughout my PhD studies: this work gave me the confidence to model absorption and emission of a singlet oxygen sensor, confirming experimental data. This work will be published soon. In addition, I am now able to direct my research with computational techniques. I am currently attempting to design new dopants for diamond, which are extremely challenging to make. By modeling defects, I can predict dopants characteristics, streamlining research choices and providing absorption and emission data to confirm new materials.
In short, STF-funded resources have drastically shaped and will continue to impact my research to date. The expansion of computing power on campus will enable current users to conduct more impactful research and afford more entry-level students the chance to learn this incredibly versatile and powerful tool. In addition, the community around UW will ensure that every student who wants to conduct computational resources can do so. I firmly every dollar spent on computing at the University of Washington will translate into increased excellence of research and broaden the toolset of every student.
Sincerely,
Matthew Crane
matthewjosephcrane@gmail.com
__
Dear STF committee,

My name is Michael McKeirnan. I am a 3rd year undergrad in the department of Atmospheric Sciences, which is part of the college of the environment. My research involves identifying cloud systems based on satellite data. I use high performance computing rarely, but would like to learn how to use such resources effectively. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in teaching the effective use of high performance computers, and for giving students access to resources they would almost never be able to find elsewhere - especially not for the same cost.
Thank you for your time and consideration of this request.

Sincerely,
Michael McKeirnan
mckeimic@uw.edu

__
Dear STF committee,
My name is David Baggett. I am a graduate student in the department of Medicinal Chemistry, which is part of the school of Pharmacy . My research involves computationally modelling molecular systems to determine how proteins will behave in solution. My aim is to use these simulations to develop therapeutics for diseases such as Alzheimer's disease. To do this, I use high performance computing several days a week. I writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: The molecular dynamics simulations, computational docking simulations to develop therapeutics, and analysis programs that help determine patterns in my simulations.

Thank you for your time and consideration of this request.
Sincerely,
David Baggett
baggettd@uw.edu
__d
Dear STF committee,

My name is Willy Voje. I am a 3rd year graduate student in the department of chemical engineering, which is part of the college of engineering. My research is concerned with computationally designing genetic components that go into biofuel, pharmaceutical, and consumable producing cells. This research requires a large number of simulations to be completed in order to validate the computational designs. Consequently, I use high performance computing at least weekly to design devices that I will later build and test experimentally. I am writing to express very strong support for the proposal as the additional computing support will enable me to ask bigger, more difficult questions that I currently cannot investigate with the available resources.

Thank you for your time and consideration of this request.

Sincerely,
Willy Voje
wmvoje@uw.edu
__

I am writing in support of the STF proposal for acquisition of Hyak nodes. I am a fifth year graduate student in Jim Pfaendtner's research group, and I have been using Hyak to conduct research during my studies here at UW. Hyak is by far the most valuable tool we have in conducting our research, and I know there are several ways in which access to Hyak could benefit the community of student researchers at large. Highly scalable simulations can be used in two capacities. First, one can speed up the solution of a problem by employing thousands of processors to complete it. In our case, we use molecular simulation, but there is no reason that this method could not be used for coarser finite difference methods or more detailed models such as quantum mechanical calculations. This allows us to probe systems of varying time and length-scales from femtoseconds to hours and from the subatomic to the macroscopic. Second, one can screen a large array of variables very quickly with access to more computing power. One specific way that I would make use of this power would be to screen new solvents. We have the ability to predict the themodynamic and physical properties of almost any organic solvent using molecular dynamics or Monte Carlo simulations. By computationally screening solvents, we can direct researchers to specific solvents that would fit their experimental constraints, or we can help predict the properties of new theoretical solvents that have not yet been synthesized in a lab. The problem with implimenting this technique is that we need to test a large library of possible solvents for any given problem that we would like to solve. We have the tools to automate the process of conducting the simulations and analysis on Hyak; we just need more computational power to make this a reality. This is only one way of many that a large high performance computational resource such as Hyak can be used to advance research on our campus. This is truly an opportunity to put useful and impactful equipment into the hands of the students here at UW, and I wholeheartedly support it.

Thank you,

Vance Jaeger
vjaeger@uw.edu

__

Dear STF committee,
My name is Gwenn Hennon and I am a 5th year graduate student in the School of Oceanography, College of the Environment. My research involves manipulating large data sets from both oceanographic cruises and gene expression from diatoms grown in the lab. I use high performance computing on a monthly basis to analyze these data sets and I advise undergraduates that also assist me with high performance computing. I am writing to express strong support for the proposal for additional computing resources available to students. With the new escience big data initiative and universities focus on supporting data science it seems like an ideal time to increase student access to high performance computing. The skills I have gained in this area are invaluable to my career as an oceanographer and I believe one of the most important areas for undergraduates to have real world experience with for their own career development.
Thank you for your time and consideration of this request.
Sincerely,
Gwenn Hennon
gwennm@uw.edu

__
Dear STF committee,

I am a fifth year graduate student in Jim Pfaendtner’s research lab. My work primarily focuses on simulations of chemical reactions involved in biomass processing. We use quantum mechanics/molecular mechanics to probe details of reactions that break down cellulosic chains to turn plants into ethanol and reactions that convert bio-oil from algae into biodiesel. Both of these processes require enzyme catalysts, which are large proteins found in bacteria and fungi.

Using the computational resources we have, we have only been able to model the transition states of these reactions at the quantum level with an implicit solvent model to account for water or other solvents that are typically used to break down plant matter. Implicit solvation models the solvent molecules as a continuum force field, instead of the explicit molecules that can change the reaction and geometry of the reacting molecules. In addition, we can only model molecules that are representative of the pieces of the enzyme that are reacting with the substrate molecules.

If we were able to use large scale computing, we would be able to model transition state theory computations for the entire active site of the enzymes that are involved in converting bio-oil and cellulosic material into usable fuel. Access to this type of detail would dramatically speed up the long-term goal of the project, which is to design catalysts that mimic the enzymes found in nature using the same reaction properties. Additionally, we could model these reactions in various solvents to further increase our understanding of the efficiency of enzymes working symbiotically with solvents to break down biomass. Overall, access to this type of computational power would better enable us to design catalysts that would make biomass processing cheap enough that cellulosic biofuel could compete with the fossil fuel market.
Kelly Fleming
klflemin@uw.edu
__
Dear STF committee,

My name is Kiran Kanekal. I am a second-year graduate student in the Chemical Engineering department, which is part of the College of Engineering at UW. My research involves relating conductive polymer dynamics to their bulk electric properties using molecular dynamics simulations.

Conductive polymers are also known as conjugated polymers due to the conjugated bonds that make up the backbone of these macromolecules, allowing charge transport to occur. Another consequence of their conjugated backbones is that the polymers have an extremely rigid structure, and therefore cannot be easily processed (either by dispersing in solvent or melting). To overcome this obstacle, the conjugated polymer backbones are modified with substitution moieties, such as alkyl chains. These polymers form the basis for much research being done on novel organic electronic materials for solar cell and battery applications. Extensive work has already been done in the synthesis and structural characterization of new conductive polymers with a broad range of backbones and substitution moieties. Thus, the relationship between the morphology of these polymers and their relevant bulk properties has been enumerated significantly. However, the link between dynamic fluctuations of conjugated polymers and these same bulk properties has yet to be understood. One experimental method for investigating this relationship is quasi-elastic neutron scattering (QENS), which is used to determine the characteristic time scales of polymer dynamics at different length scales. Our group has already collected QENS data for a model class of conjugated polymers. By itself, QENS data is difficult to interpret without making several assumptions as to which specific segments are undergoing a particular dynamic fluctuation. By using this data to validate molecular dynamics simulations of the same system, however, we can determine which parts of the polymer chain have the greatest contribution to their dynamics. This will establish a set of guidelines that will help synthetic chemists tailor their conjugated polymer designs to specific applications (e.g. solar cell vs. field effect transistor). In order to complete this work, I will be using high performance computing on a weekly basis. I am writing to express my very strong support for the proposal, as the additional computing support would be beneficial in the following ways:

1. It would allow for larger simulations that could be run for longer durations, better representing the material being simulated.

2. It will allow for more simulations to be run in parallel, providing greater simultaneous access to more initial molecular configurations.

Thanks,

Kiran Kanekal

__
Dear STF committee,
My name is Maggie McKeon. I am a grad student in the department of Civil & Environmental Engineering, which is part of the college of engineering. My research involves modeling the influence of climate change on hydrodynamics and sediment transport in urban estuaries. I use high performance computing daily. I writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: allowing increased access to high performance computing for students. Numerical modeling frameworks are becoming more and more accessible, but often students do not have access to the kind of high performance computing resources that can fully take advantage of their full capabilities. Furthermore, as more and more numerical models are being used to answer questions related to climate change, high performance computing become necessary. This would really provide a great opportunity for students to do meaningful scientific work.

Thank you for your time and consideration of this request.
Sincerely,
Maggie McKeon
mckeon@uw.edu

__
Dear STF committee,
My name is Kyle Caldwell, and I am a 4th year graduate student in the department of chemical engineering, which is part of the college of engineering. My research involves tailoring the fiber- matrix interphase of fiber reinforced composite materials, as well as predicting polymer-polymer adhesion for use in other types of engineering composite materials. I use high performance computing weekly for my research. I writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: polymer-polymer and polymer-fiber surface MD simulations investigating the spatial distribution of properties near and far from the interface would be accessible, the computing time required for the many DFT simulations required for my research would be greatly reduced. Additional computing power would also enable simulations of micelle formation in non-aqueous media, which is another branch of my groups research.
Thank you for your time and consideration of this request. Sincerely,
Kyle Caldwell kcal@uw.edu
__
Dear STF committee,

My name is Una Nattermann. I am a graduate student in the Department of Biochemistry, which is part of the College of Medicine. My research involves the computational design of self-assembling protein nanomaterials. I use high performance computing monthly. I am writing to express support for the proposal as the additional computing support would be beneficial in the following ways: (1) decrease waiting time to launch jobs, (2) increase # of jobs that can be submitted at one time, (3) decrease time spent waiting for jobs to finish in order to proceed to the next step of my project, and (4) increase usage of high performance computing, freeing up some of the other options from overload as well.

Thank you for your time and consideration of this request.

Sincerely,

Una Nattermann

--
Una Nattermann
Biological Physics, Structure and Design Graduate Student
Baker Lab
Department of Biochemistry
University of Washington, Seattle
natteruw@uw.edu
__
Dear STF committee,

My name is Shushan He. I am a second year graduate student in the department of Chemistry, which is part of the college of Arts and Sciences. My research involves large scale simulation of biophysical system such as model lipid bylayers. I use high performance computing almost everyday for my research. I am writing to express very strong interest and support for the proposal as the additional computing support would be beneficial in the following ways:

1) More efficient computer clusters drastically shortens the length of each 'simulation --> analysis --> preliminary results --> more simulation' research cycle, allowing us to finish each step of the project in a more timely and productive manner
2) With the help of more cutting edge, high performing computing power, much time can be saved from simply waiting for jobs to finish or dealing with issues cased by limited resources, and this allows graduate students like me to devote more of our time and efforts to other important aspects of graduate school, such as teaching.
3) A lot of potential graduate students in the theoretical/computational field highly value the importance of quality computing resources when choosing between institutions - having that kind of resources will definitely put UW in a more advantageous position when competing for better incoming graduate students in such fields.

Thank you for your time and consideration of this request. I really hope the committee consider this proposal a huge benefit to a lot of the students on the UW campus.

Sincerely,

Shushan He

hes8@uw.edu
__
Dear STF committee,

My name is Brandon Frenz. I am a second year graduate student in the department of Biochemistry. My research involves the determination of protein structures from sparse experimental data. I use high performance computing daily. I am writing to express support for the proposal as the additional computing resources would be helpful as my research often requires huge amount of computational time as part of the methods development.

Thank you for your time and consideration of this request.

Sincerely,

Brandon Frenz

brandon.frenz@gmail.com
__
Dear STF committee,
My name is Bo Peng. I am a 4th year graduate student in the department of chemistry, which is part of the college of art and science. My research involves supercomputer assisted inorganic material and small peptide design and electronic structure development. I use high performance computing almost everyday. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: It will definitely boost the research progress and publication speed.
Thank you for your time and consideration of this request.
Sincerely,
Bo Peng
bpeng@uw.edu
__
Dear STF committee,

My name is Adam Richie-Halford. I am a second year graduate student in the Department of Physics, which is part of the College of Arts and Sciences. My research involves the study of real-time dynamics of quantized vortices in a unitary fermi gas. Or to put it more simply, I study how swirly things interact with each other in very exotic states of matter. I use high performance computing resources at least weekly and during peak periods of research, I use HYAK daily. I am writing to express my very strong support for the proposal because the additional computing support would be beneficial in the following ways:
• Additional computing resources would allow me to solve problems more quickly. It takes time to get results and our group's limited resource allocation sometimes makes compute time main limitation to completing our research more quickly.
• Even better, additional computing resources would allow me to solve bigger, more challenging problems. The new HYAK cores would serve both as a place to test new and innovative ideas and as a tool to attack the bigger problems on our research agenda.
• The additional computing resources and the student high-performance computing (HPC) organization would help build the technical skills that I need to succeed after I graduate from the UW. They would help to make the UW a center of excellence in HPC and would foster more collaboration between the UW's computer science, applied math, and natural science communities.
I thank you for your time and consideration of this request. If you have any questions, please don't hesitate to contact me.

Sincerely,
Adam Richie-Halford
richford@uw.edu
__
Dear STF committee,

My name is Feizhi Ding. I am a fifth-year graduate student in the department of chemistry which is part of the college of arts & sciences. My research involves computational design and simulation of organic polymers for photovoltaic applications. I use high performance computing daily. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: (1) since the number of users of our computer clusters is increasing significantly during the past few years, the current computing resources become more and more inadequate. There is a significant amount of time wasted waiting in the queue of computations. The additional computing support would help improve the situation by reducing the waiting time for computations. (2) Since the current computer clusters are aging quickly, the performance of the computations becomes worse in the sense that both the data transfer and the calculations take longer and longer time. The new additional computing support would hopefully have the improved computing performance that meets our research demands.

Thank you for your time and consideration of this request.

Sincerely,

Feizhi Ding

Email: fding@uw.edu
__
Dear STF committee,

My name is David Sparkman-Yager. I am a first year graduate student in the Molecular Engineering Ph.D. program housed within the graduate school. My research involves the computational prediction of RNA structure, as a means to identify novel aptazymes (small-molecule-induced autocatalytic RNA switches). I use high performance computing nearly every day. I am writing to express my extremely strong support for the proposal, as the additional computing support would allow me to increase both the number of sequences screened for desired function, as well as the thoroughness of that screening. Additionally, I would be free to pursue other self-directed projects without fear of congesting laboratory nodes.

Thank you for your time and consideration of this request.
Sincerely,
David S-Y

David Sparkman-Yager
Molecular Engineering Ph.D. Student
Carothers Lab
University of Washington
dsparkmanyager@gmail.com
__
Dear STF committee,

My name is Jonas Wilhelm. I am a first year grad student in the department of Physics,
which is part of the college of Arts & Sciences. My research involves Lattice QCD
calculations. I use high performance computing daily. I writing to express very strong
support for the proposal as the additional computing support would be
beneficial in the following ways:

- calculations would be faster
- calculations which need higher computational speed could be done

Thank you for your time and consideration of this request.

Sincerely,

Jonas Wilhelm

jwilhelm@uw.edu
__

Dear STF committee,

My name is Mishaal Aleem, and I am an undergraduate in the Aeronautics & Astronautics department, which is a part of the College of Engineering.

My research involves visualization of direct numerical simulations of droplet-laden turbulence. This research is a unique opportunity for me, as an undergraduate, to be involved in industry-relevant, high-fidelity simulations of complex fluids. As part of this research, I use high performance computing weekly. I am writing to express very strong support for the proposal as the additional computing support would significantly increase the computing potential for not only this project, but open up further research opportunities. As I am very interested in continuing on to the graduate-level, this proposal would be of significant benefit.

Thank you for your time and consideration of this request.

Sincerely,
Mishaal Aleem

aleemm@uw.edu
__
Dear STF committee,

My name is Joshua Goings. I am a third year graduate student in the department of Chemistry, which is part of
the college of Arts and Sciences. My research involves computing the properties of molecules under electromagnetic radiation. The accurate predictions as the result of my work go towards developing new materials that advance national energy and information processing interests. I use high performance computing daily. I writing to express very strong support for the proposal as the additional computing support would be beneficial
in the following ways: rapid computational analysis of new materials created in collaboration with experimental groups at the UW, development of accurate computational methodologies to predict new properties of cutting-edge materials, as well as detailed insight into fundamental processes in the molecular sciences.

Thank you for your time and consideration of this request.

Sincerely,

Joshua Goings

jjgoings@uw.edu

__

Dear STF committee,

My name is Chang Dou. I am a PhD student in the department of SEFS, which is part of the college of Environment. My research involves biofuel production from lignocellulosic biomass. I use high performance computing weekly. I writing to express very strong support for the proposal as the additional computing support would be beneficial in the high performance computing in running the chemical process optimization model and life cycle assessment model.
Thank you for your time and consideration of this request.
Sincerely,

Chang Dou
changdou@uw.edu
__
Dear Prof. Pfaendtner,

Please find below my endorsement for Hyak. Thank you for collecting these endorsements.

Dear STF committee,

My name is Michael Dodd. I am a fourth year graduate student in the department of Aeronautics and Astronautics, which is part of the College of Engineering. My research involves performing direct numerical simulation of multiphase turbulent flows. I use high performance computing weekly. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: reduce the time to solution, allow more students in the Computational Fluid Mechanics lab to have computing jobs running simultaneously, and facilitate the development of petascale numerical methods.

Thank you for your time and consideration of this request.

Sincerely,
Michael Dodd
doddms@uw.edu
__
Dear STF committee,

My name is Matthew Childers, and I am second year graduate student in the Department of Bioengineering, which is a joint program in the School of Medicine and College of Engineering. I work in Valerie Daggett's research group performing realistic simulations of protein dynamics. Specifically, I am currently using computational methods to study the intrinsic conformational sampling (i.e. distribution in Ramachandran or dihedral angle space) of the naturally occurring amino acids. Due to the high computational demand of these simulations, I use high performance computing on a daily basis. I writing to express very strong support for the proposal as the additional computing support would be beneficial in many ways. Namely, I am currently enrolled in a molecular engineering course that introduces several theoretical techniques beyond classical dynamics. It is my belief that the skills I gain in this course will allow me to employ a variety of simulation techniques during my research to gain insight into the dynamic behavior of proteins across multiple spatial and temporal scales.

Thank you for your time and consideration of this request.
Sincerely,

Matthew C Childers
Graduate Student - Daggett Research Group
Department of Bioengineering - University of Washington

Email: mcc7fb@uw.edu
Phone: 206.543.9305
__
Dear STF committee,

My name is Jianqing Qi. I am a 5th PhD student in the department of Electrical Engineering, which is part of the college of engineering. My research involves charge transport through biological molecules. I use high performance computing on a daily basis. I am writing to express very strong support for the proposal as the additional computing support would be beneficial for students’ research.

Thank you for your time and consideration of this request.

Sincerely,

Jianqing Qi
jqqi@uw.edu
__
Dear STF committee,

My name is Walter Perkins. I am a 2nd-year graduate student in the Department of Atmospheric Science, which is part of the College of Environment. My research currently involves the design of simple a computational model to help us utilize large sets of paleoclimate observations in reconstructing past climates. While the models we use are simple, the datasets we use are of a size where our productivity is greatly increased by the availability of distributed computing environments.

I am writing to express my support for the proposal to provide undergraduates with access to high-performance computing resources. I had the luxury of being exposed to this type of computing environment in my undergraduate education, and this skill proved invaluable for an easy transition into my graduate research. Providing students with exposure to these systems and offering training will provide useful skills immediately applicable to real-world science. It is becoming increasingly important to have experience in a variety of computational tools by the end of a Bachelor's degree, and I believe this proposal and subsequent project will add an important tool to that set.

Thank you for your time and consideration of this request.

Sincerely,
Walter Perkins
wperkins@uw.edu
__
Dear STF committee,

My name is Miriam Williamson. I am a 4th year undergraduate in the Department of Biochemistry, which is part of the School of Medicine. My research involves using the software suite Rosetta to make structure predictions of small, cyclic proteins. I use high performance computing daily. I am writing to express support for the proposal as the additional computing support would be beneficial in the following ways: I will allow me to explore larger folding trajectories, more nodes would allow me to run more trajectories in parallel, taking less time.
Thank you for your time and consideration of this request.

Sincerely,
Miriam Williamson
miriamw@uw.edu

__
Dear STF committee,

My name is Heidi Rockney. I am currently finishing up my undergraduate senior research project in the department of Biology in the school of Arts & Sciences at the University of Washington. My research involves analyzing two different genes of frogs from Ghana in a conservation genetics project. It is necessary for me to use high performance computing at least three times a week. I am in extremely strong support for this proposal. Access to additional computers would greatly benefit me, as there are several students needing access to school computers, and more computers would mean less wait time and greatly improve my ability to finish my undergraduate research in a timely manner.

Sincerely,

Heidi Rockney
hjr2@uw.edu

__
Dear STF Commitee,

My name is Kovas Palunas and I am a third year student studying Chemical Engineering. As a member of Professor Jim Pfaendtner's group, I use the hyak high preformance computing cluster three to four times a week to do my work, which involves modeling of short peptide chains. The HPC cluster is instumental in allowing me to do my research, often unfortunately meaning that it is the bottleneck that restricts my progress. With more available cores, I could much more easily both test my model systems, and grow them to larger sizes. The growth more cores would allow would make my models more accurate (closer to reality) and would also allow me to run tests involving much larger protein chains. To give a sense of scale, I now run tests with a 14 amino acid long chain that can take several days to complete. Many functional proteins found in nature are ten times this size (~150 acids), and have very complicated behaviors that are hard to explain using my current setup. Unfortunately, it is hard to do research on a system when it takes weeks to just do a single run with it!

I strongly urge you to accept this proposal; it would greatly improve the quality of my work and allow me to explore more complicated, interesting, and relavent systems inaccessable to me now. Thank you for your time and consideration,

Kovas Palunas
__
Dear STF committee,

My name is Dave Slager. I am a 2nd year PhD student in the Biology
department, which is part of the College of Arts and Sciences. My
research involves using state-of-the art, massively parallel DNA
sequencing technology to make inferences about evolutionary
relationships, hybrid zones, and population genetics, primarily in
birds. This research requires high-performance computing on an
approximately monthly basis and often for extended runs. In fact,
access to high-performance computing is a primary limiting factor in
whether or not much of my proposed research is feasible, because the
computing time required for many of my research questions increases
faster than exponentially as sample size increases. I am writing to
express very strong support for this proposal as the additional
computing support would allow me to use cutting-edge bioinformatics
methods such as SNAPP in my research and compete on a level playing
field with graduate students at UW peer institutions already
possessing such computational resources.

Thank you for your time and consideration of this request.

Sincerely,

David L. Slager
slager@uw.edu
__
Dear STF committee,
My name is CJ Battey, and I am a second-year graduate student in the department of Biology, part of the college of Arts and Sciences. My research is focused on using genomic data to track gene flow in recently diverged populations in which the signals of introgression and ancestry are mixed. I use high performance computing daily to assemble consensus sequences from raw data and conduct computationally intensive Bayesian inference under a variety of population-genetic and phylogenetic models. Additional computing infrastructure would allow me to run my analyses in parallel and greatly reduce the time necessary to complete projects, as well as providing the flexibility to more completely explore the effect of various model parameters on final results. I strongly support the purchase of additional computing power available to graduate students across the university.
Thank you for your time and consideration of this request.
Sincerely,
CJ Battey
cjbattey@uw.edu
__
Dear STF committee,

My name is Yunqi Zhao. I am a 4th year graduate student in the department of electrical engineering, which is part of the college of engineering.

My research involves (1) developing high performance algorithm used in atomistic simulation of nanoscale devices. (2) modeling electron transport property in novel nano-structures by large-scale computation.

I use high performance computing everyday.
I writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: (1) greatly enhancing the efficiency of the quantum mechanical simulation we currently performed. (2) enlarging the size scale of the system we investigated, which is essential to provide insights in industrial applications.

Thank you for your time and consideration of this request.

Sincerely,
Yunqi Zhao
zhaoyq@uw.edu

__
Dear STF committee,

My name is Stephanie Berger, and I am a graduate student in the Bioengineering Department. My research in David Baker’s lab involves computational design of novel proteins for cancer treatment. Many of my colleagues in the Baker Lab use high performance computing on a daily basis, and I plan to use it as well in an upcoming project. I am writing to express very strong support for the proposal as the additional computing support would allow us to complete our computational simulations with shorter queue times and expedite our research.

Thank you for your time and consideration of this request.
Sincerely,
Stephanie Berger
berger389@gmail.com
__
Dear STF committee,

My name is David Lingerfelt. I am a third year graduate student in the department of chemistry, which is part of the college of arts and sciences. My research, most generally, involves real-time quantum mechanical simulations of chemical systems under non-equilibrium conditions. For example, in a recently published study, I simulated the time evolution of electron density in organic photovoltatic materials immediately following photon absorption to better understand the ultra-fast electronic response of these materials and aid in the computational design of high-efficiency solar energy harvesting materials. For this and similar purposes, I use high performance computing nearly every day. I am writing to express very strong support for the proposal, as the additional computing support would be beneficial in the following ways:

I will be able to validate the programs I'm developing in the field of molecular dynamics on larger, more experimentally interesting test-cases rather than the more simplistic model systems that I currently have the capacity to study. The specific variant of molecular dynamics I am implementing approximates a fully quantum-mechanical molecular dynamics simulation (i.e. one in which the quantum-mechnaical nature of both electrons and nuclei are taken into consideration) with numerous independent classical nuclear trajectories with the minimal amount of randomness built into the time evolution to properly account for the uncertainty that is infamously called for in the quantum formulation of physics. This method, while far less computationally demanding than the direct numerical integration of the time-dependent Schrodinger equation for the many-body electron-nuclear wavefunction, is still quite expensive for most interesting chemical systems. With access to the robust computational resources requested by the proposal, I can stay on track for timely graduation without having to compromise on the size/complexity of systems I can apply my developed methods to. In short, the approval of this proposal will promote the application of cutting-edge molecular dynamics tools developed at the UW to probe the physical phenomenon of relevance to outstanding problems in chemistry as diverse as energy harvest/storage and quantum computing.

Additional HPC resources will also enrich the experience and productivity of undergraduate researchers in the theoretical chemistry division. Rarely are even the most advanced undergraduate chemistry students well-versed enough in the fields of applied math and computer science to take on a research project that requires extensive scientific programming. However, undergraduates researchers can still play a fruitful role in the theoretical chemistry research process through the preparation of input files and post-processing of output files necessary to use existing software in a scientific study. During this process, students gain invaluable programming experience through writing basic scripts to perform these tasks in an automated fashion, and also learn the basics of navigating in a Unix-type, command line environment. Over the course of ~10 week summer research programs like those my research group takes part in annually, a students can only realistically hope to complete a project of relatively narrow scope. If only the most minimal of computational resources are allocated to the student, though, the scope of their study is especially reduced. The quality of resources available to these exceptional young researchers in this situation can represent the difference between a relatively derivative theoretical study that is purely for the student's academic enrichment and one that is novel and potentially publishable in a peer reviewed journal. Thank you for your time and consideration of this request.

Sincerely, David Lingerfelt
DBLinger@uw.edu

Dear STF committee,

My name is Addie Kingsland. I am a 3rd year graduate student in the Department of Chemistry. In my research I study DNA, especially with regard to base pair mismatches (which has numerous applications in cancer research) and with materials modifications to DNA (which has applications in self-assembling nanostructures). I use high performance computing almost daily, and on occasion for months straight. I am writing to express very strong support for the proposal as the additional computing support would be incredibly beneficial. More computing resources would mean less wait time for my computing jobs, as well as the ability to run more calculations at a time. This would expedite my gathering of data and lead to less "dead time" while I am working.

Thank you for your time and consideration of this request.

Sincerely,
Addie Kingsland
adelak@uw.edu
__
Dear STF committee,

My name is Sergey Ovchinnikov. I am a 4th year graduate student in the department of Molecular and Cellular Biology. My research involves predicting protein structures using Rosetta structure modeling suite and co-evolution data. High performance computing is essential to my everyday research. I am writing to express support for the proposal as the additional computing support would be beneficial in increasing the amount of sampling that I could do and would allow addressing more biological system.

Thank you for your time and consideration of this request.

Sincerely,

Sergey Ovchinnikov
krypton@uw.edu
__
To whom it may concern,
My name is Karl Oleson and I am a 2nd year graduate student in chemical engineering. The research I am involved in is focused on creating biofuel from Pacific Northwest logging residue. Specifically, I am researching how small defensive molecules inside of trees called “extractives” affect the biological processing steps of fermentation based biofuel.
Creating biofuel from logging slash is possible, but research and optimization of expensive process steps needs to be done before we run our cars on trees. In these fermentation based biofuel processes, one of the most expensive steps involves biomass-degrading enzymes that break down the wood fibers into sugar units for fermentation. Certain extractives (such as tannins) are known to bind to and deactivate many enzymes and could be substantially harming this break down process. My experimental work involves observing how these molecules and enzymes reduce and inhibit biomass break down. However, handling these chemicals and solution in lab can only give us a small amount of information about what is actually happening.
Molecular dynamics simulations using high performance computing on HYAK would be able to identify how extractives inhibit these important enzymes. Using the GROMACS simulation software (which needs high performance computing capabilities), better biologically engineered enzymes could be designed to reduce these negative effects. As enzymes are one of the most expensive steps of these biofuel strategies, it could lead to reduced operating cost and accelerate the commercialization of forestry based biofuels.
Simulating these enzyme-extractive interactions is beyond the capabilities currently available to me, but would be possible with the extra cores on HYAK if this proposal is approved. Therefore, I strongly support this STF proposal and ask for its approval. Thank you for your time.

Sincerely,
Karl Oleson
karlolesonche@gmail.com
__

Dear STF committee,

My name is Kerry Garrett. I am a 5th year graduate student in the department of Chemistry, which is part of the College of Arts and Sciences. My research involves calculating the electronic and optical properties of second-order nonlinear optical organic materials. I use high performance computing every day. I am writing to express very strong support for the proposal as the additional computing support would mean less time for my jobs to wait on the queue. This is important for my projects which require many calculations to be completed in a timely manner.

Thank you for your time and consideration of this request.

Sincerely,

Kerry Garrett
keg22@uw.edu
__
Dear STF committee,
My name is Giada Arney. I am a fifth year grad student in astronomy and astrobiology, which is part of the college of Arts and Sciences. My research involves simulating planetary atmospheres to better understand early Earth environments billions of years ago. I also simulate Earthlike exoplanets with hydrocarbon hazes in their atmospheres. I use high performance computing weekly to run tens to hundreds of radiative transfer simulations at once. In addition, I mentor an undergraduate student, Guadalupe Tovar, who uses Hyak on bi-weekly basis. I writing to express very strong support for the proposal. The additional computing support would be beneficial because the shared research group Hyak nodes I have access to are frequently filled to capacity by more senior members of my group.
Thank you for your time and consideration of this request.
Sincerely,
Giada Arney
giada@uw.edu

__

Dear STF committee,

My name is Jamie Oaks. I am a postdoctoral research fellow in the
Department of Biology, which is part of the College of Arts and
Sciences . My research involves developing new computational methods
for inferring evolutionary history. I use high performance computing
multiple times per week. I writing to express very strong support for
the proposal as the additional computing support would be beneficial
in the following ways: It will allow me to much more efficiently
develop, test, and deploy computational tools that will be useful to a
large community of evolutionary biologists. It will also allow me to
apply these methods to empirical data to test hypotheses about
important processes involved in biological diversification

Thank you for your time and consideration of this request.

Sincerely,

Jamie Oaks
__
Dear STF committee,

My name is Mehedi Maswood. I am a second year graduate student in the department of Civil and Environmental Engineering, which is part of the college of Engineering. My research involves computationally intensive hydrologic modeling for large-scale continental basins. I use high performance computing DAILY. I am writing to express very strong support for the proposal as the additional computing support would be beneficial for my graduate studies.

Thank you for your time and consideration of this request.

Sincerely,

Mehedi Maswood

Email: mmaswood@uw.edu
__
Dear STF committee,

My name is Lewis Johnson. I am a a recent PhD alum (2012) and current Visiting Scholar in the department of Chemistry, which is part of the College of Arts and Sciences. My graduate research at UW involved extensive, typically daily, use of STF funded high-performance computing resources (the Stuart cluster), which contributed to six published peer-reviewed papers. These resources were invaluable for both compute-intensive simulations of organic electro-optic materials, learning about HPC and Linux systems, and teaching summer research students about electronic structure calculations. I apply the skills that I developed as a grad student using STF resources on a daily basis in my current research as well, and also applied them in mentoring research students on computational projects at large-scale HPC sites during my prior teaching position at Pomona College. I strongly support the proposed student expansion of the Hyak cluster for providing state-of-the-art HPC resources to the next generation of students.

Thank you for your time and consideration of this request.

Sincerely,

Lewis E. Johnson
lewisj@uw.edu
__
Dear Student Technology Fee Committee,

My name is Edward Schwieterman and I am a graduate student in the astronomy department, which is part of the College of Arts and Sciences. I would like to express my strong support for the additional high performance computing resources that would be afforded by this proposal. My research involves simulating the spectral appearance of terrestrial planets orbiting other stars, as future space-based telescopes might see them. I make use of high performance computing on a daily basis for my research, because the amount of computation I need to complete my projects is beyond the capacity of my workstation and laptop computer. Additional nodes provided for student use will be beneficial because it will allow me to begin calculations immediately when the super cluster would otherwise have been subscribed to its limit. This will allow me to get results faster and proceed with my research projects more quickly. Thank you for considering this request.

Sincerely,

Edward Schwieterman
PhD Student, Astronomy & Astrobiology
eschwiet@uw.edu
__
Dear STF committee,
My name is Md. Safat Sikder. I am a second year graduate student in the department of civil and environmental engineering, which is part of the college of engineering. My research involves computationally intensive modeling of atmospheric processes using numerical modeling. I use high performance computing daily. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: it would be beneficial for my graduate study.

Thank you for your time and consideration of this request.

Sincerely,
Md. Safat Sikder
mssikder@u.washington.edu
__
Dear STF committee,
My name is Jared Grummer and I am a fourth year PhD candidate in the Department of Biology. I study population genetics and systematics of lizards and frogs using next-generation sequencing (NGS) technologies. NGS-based research necessarily involves a high level of computation because the large volume of sequence data generated from NGS platforms. I therefore use high performance computing on a weekly basis to both construct my datasets and analyze them. In fact, I have a manuscript accepted to the journal Molecular Ecology that acknowledges the Hyak Computing Cluster that was instrumental in analyzing data for my research.

I am writing this endorsement to express how important having these computational resources is to my research, and also how strongly I support this proposal. The additional computing support would be beneficial to me by allowing fast and accurate dataset construction, enabling the completion of analyses that are impossible to perform on desktop computers, and drastically decreasing analysis time that will enable me to output research at a faster rate.

Thank you very much for your time and consideration of our request.

Sincerely,
Jared Grummer

grummer@uw.edu
__
Dear STF committee,

My name is Patrick Lestrange and I am a third year graduate student in the Department of Chemistry. My research focuses on modeling ultrafast X-ray absorption and multi-dimensional spectroscopy. These techniques have a wide array of applications in solar energy and inorganic catalysis and theoretical modeling is often used to help interpret complex spectra. I use high performance computing on a daily or weekly basis. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in many ways. With the computing resources I currently have access to, my calculations will often spend a lot of time in the queue before they begin. This proposal will allow me to obtain results for my experimental collaborators more quickly and will also allow me to work on larger chemical systems that are very difficult or impossible for me to study currently.

Thank you for your time and consideration of this request.

Sincerely,

Patrick Lestrange

plestran@uw.edu
__
Dear STF committee,

My name is Brian Koepnick. I am a graduate student in my third year in the Department of Biochemistry at the UW. My research involves computational modeling of large molecules. I use high performance computing every week. I am writing to express support for the proposal as the additional computing support would allow me to compute additional jobs every week and accelerate my research.

Thank you for your time and consideration of this request.

Sincerely,
Brian Koepnick

koepnick@uw.edu
__
Dear STF committee,

My name is Kayla Sapp. I am a fourth year graduate student in the department of Chemistry, which is part of the college of Arts and sciences. My research involves molecular dynamics simulations of continuum level lipid bilayers coupled to membrane associated proteins. I investigate how the coupling between the membrane and protein induces an attractive interaction among proteins and how the membrane influences protein diffusion. I use high performance computing daily. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: an increase in capacity will be beneficial with the growing amount of research in use of high computing power and an upgrade would be helpful as well.

Thank you for your time and consideration of this request.

Sincerely,

Kayla Sapp
ksapp1@uw.edu
__
Dear STF committee,
My name is Molly Grear. I am a first year PhD student in the Civil and Environmental Engineering Department with Professor Mike Motley. My research involves modeling fluid-structure interactions in order to understand the potential impacts of marine and hydrokinetic energy devices on marine mammals. I hope to model injury to the animals, as well as the complex fluid dynamics involved. I use high performance computing monthly presently and plan to use it more as my research ramps up. I writing to express very strong support for the proposal as the additional computing support would be beneficial to ensuring students can easily have access and short queues to work with HYAK. Additionally, students can learn the skill of using high powered computers effectively, as this will unquestionably become increasingly valuable for fluid dynamics research.
Thank you for your time and consideration of this request.
Sincerely,
Molly Grear
mgrear@uw.edu
__
Dear STF committee,
My name is Matthew Bonnema. I am a graduate student in the department of Civil Engineering, which is part of the college of Engineering. My research involves computationally intensive modeling of satellite-data driven surface water modeling. I use high performance computing daily. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: It would be beneficial to my graduate studies.
Thank you for your time and consideration of this request.
Sincerely,
Matthew Bonnema
matthew.bonnema@gmail.com
__
Dear STF committee,

I'm Jason, a third-year PhD student in the departments of Bioengineering and Chemical Engineering. I research how we can engineer bacteria to make high-value chemicals cheaply and renewably. To do this, I create and simulate thousands of mathematical models of bacterial metabolisms to predict yields of the chemical product. Hyak is critical for this work, as each set of simulations takes thousands of core-hours of computation time. On Hyak, a simulation set currently takes around one day. Having more capacity on Hyak means faster simulation times and a quicker troubleshooting cycle, as well as the potential for more computational projects in our lab to operate simultaneously.

Thanks for your time and consideration!

Jason Stevens
jasoste@uw.edu
__

Dear STF committee,
My name is Sima Bouzid. I am a graduate student in the Department of Biology, which is part of the College of Arts and Sciences. My research involves estimating the evolutionary relationships of reptiles and amphibians at the species and population levels using genomic data. I plan to integrate genomic, environmental, and physiological data in a multidimensional, computationally intensive analysis that will extrapolate the factors influencing local adaptation due to climate change. In order to maximize the accuracy of my models, I must include a very fine level of detail, which is impossible without the use of high performance computing. I use high performance computing weekly. I am writing to express very strong support for the proposal as the additional computing support would facilitate my research and that of my colleagues by allowing for increasingly more complex analyses and decreasing the amount of time it takes for these high-powered analyses to run.
Thank you for your time and consideration of this request.
Sincerely,
Sima Bouzid
bouzidnm@uw.edu

Dear STF committee,

My name is Cooper French. I am a 1st year Graduate student in the department of Biology, which is part of the college of Science. My research involves avian phylogenetics and biogeography,exploring the relationships of extant bird species and populations using next-gen sequencing and analysis. I use high performance computing on a weekly basis and multi-core computing is an essential component of all contemporary statistical tools available for phylogenetics. I write to express very strong support for the proposal as the additional computing support would reduce downtime in experimental procedures (currently computation of a single ten-species tree can take as long as three weeks), and allow us to expand the size and coverage of our analyses in keeping with the current standards in molecular phylogenetics.
Thank you for your time and consideration of this request.
Sincerely,
Cooper French
cooper.french@gmail.com

Dear STF committee,

My name is Matthew McElroy and I am a 5th year PhD student in the department of Biology at University of Washington. My research involves studying thermal adaptation in a radiation of lizards from Puerto Rico. I combine field experiments with genetic and genomic approaches to understand what processes contribute to population divergence, and what regions of the genome are under selection in different thermal habitats. The genomic data that I work with requires having extensive computational resources that I use on a weekly basis, with some analysis taking weeks to complete! I am writing to express my strong support of this proposal because (1) these computational resources can help expedite the time that analyses take to complete, and (2) as more and more students and post-docs are using these resources the computers are often times "full", creating a bottle neck for graduate students to analyze their data.
Thank you for considering this proposal.
Sincerely,
Matthew McElroy
__
Dear STF committee,

My name is Xiaodong Chen. I am a forth year graduate student in the department of Civil and Environmental Engineering, which is part of the college of Engineering. My research involves computationally intensive modeling of atmospheric processes using numerical modeling. I use high performance computing daily. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: It will greatly benefit my graduate studies.
Thank you for your time and consideration of this request.

Sincerely,
Xiaodong Chen
xiaodc@uw.edu

__
Dear STF committee,

My name is Yang Hsia. I am a 3rd year graduate student in the department of Biochemistry, which is part of the School of Medicine. My research involves designing protein based nanomaterials using computational approaches. I use high performance computing on a monthly basis. I am writing to express support for the proposal as the additional computing support would be beneficial since often times we are required to run thousands of jobs, but often times processing power is limited by the amount of hardware available; thus making data collection take much longer than it should. The availability would also allow us to generate better data as we would be able to sample a much larger set of parameters given the same amount of time.

Thank you for your time and consideration of this request.

Sincerely,

Yang Hsia
yhsia@uw.edu

--
Yang Hsia
Baker Lab - PhD Graduate Student
University of Washington Biochemistry/BPSD
__
Dear STF committee,

My name is Daniel Hnatovic. I recently graduated with my bachelor's degree in the department of Aeronautics & Astronautics, which is part of the college of Engineering. I have been conducting research part time at UW for the past year and will continue research for the next few months. My research involves computational fluid dynamics (CFD). Specifically, I am studying the path instability of a gas bubble rising in a quiescent liguid using a multiphase CFD code. These simulations require very large computing resources, and I am writing to express very strong support for the proposal as the additional computing support would be beneficial in reducing the time required for my simulations, thus allowing me to get more work done, faster.

Thank you for your time and consideration of this request.

Sincerely,
Daniel Hnatovic
__
Dear STF committee,

My name is Jiawen Xie. I am a visiting 3rd-year PhD student in the department of Aeronautics and Astronautics, which is part of the college of Engineering. My research involves numerical simulations of elastic response and failure evolution of aerospace structures made of composite materials. I use high performance computing daily. Usually each of my job need 5 hours running with 4 nodes. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: 1) significantly reducing my waiting time of a submitted job; 2) increasing the available nodes to run the simulations.

Thank you for your time and consideration of this request.

Sincerely,

Jiawen Xie
jwxie@umich.edu

--
Dear STF committee,
My name is Zibo Chen. I am a 2nd year graduate student in the department of biochemistry. My research involves computational protein design. I use high performance computing daily. I writing to express very strong support for the proposal as the additional computing support would be beneficial.
Thank you for your time and consideration of this request.

Sincerely,
Zibo Chen
zibochen@uw.edu
__
Dear STF committee,

My name is Harley Pyles. I am a 2nd year graduate student in the Molecular and Cell Biology program at UW. My research involves computational modeling of protein-inorganic interfaces. I use high performance computing on a daly basis and it is critical to my thesis project. I am writing to express support for the proposal as the additional computing support would allow me to preform larger scale analyses and accomplish more during my time at UW.

Thank you for your time and consideration of this request.

Sincerely,

Harley Pyles
pylesh@uw.edu

__

Dear STF committee,

My name is Cassie Bryan. I am a fourth year graduate student in the department of Biochemistry, which is part of the college of Medicine. My research involves computationally designing proteins for use as cancer and autoimmune therapeutics. I use high performance computing daily. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: faster design means faster research means novel therapeutics getting to patients in need sooner.

Thank you for your time and consideration of this request.

Sincerely,

Cassie Bryan
cassio87@uw.edu

__
Dear STF committee,
My name is Aaron Chevalier and I am a fifth year Bioengineering graduate student in David Baker's lab. As you probably know we use a vast amount of computational resources in this lab and I personally use HYAK quite frequently. My particular research involves designing proteins to be anti-flu or anti-ebola therapeutics. I am in very strong support of additionally computing resources at HYAK. Specifically more nodes as I submit thousands of jobs per week, and results are often delayed for days as we have insufficient computers to handle the load.

Thanks,
Aaron Chevalier

--
Aaron Chevalier
Baker Lab
University of Washington Bioengineering
achev@uw.edu
__

Jiawen Xie
Aerospace Engineering
University of Michigan, Ann Arbor
__
My name is Irfan Syahdan.
I am a graduate student (master) in the department of Aeronautics and Astronautics, which is part of the college of enginnering. My research involves a simulation of a droplet break up in accelerating flow. Most of droplet breakup research has been conducted in constant flow, while only a few in accelerating flow. Due to the multiphase nature and the high density ratio between air and water, the simulation requires a very small time step (less than a micro second) and large total mesh. I use high performance computing almost daily. I am writing to express very strong support for the proposal as the additional computing support would be beneficial because more available processor will decrease the time to finish the full simulation, thus enabling to learn more from more variation of more cases.
Thank you for your time and consideration of this request.
Sincerely,

Irfan Syahdan
irfanms@uw.edu
__
Dear STF committee,
My name is Zehua Wei. I am a first year MS student in the department of chemical engineering, which is part of the college of engineering. My research involves molecular simulation of the interaction between peptide and 2D BN surface. I use high performance computing over 3 times per week, so I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following ways: ?enlarge the research field of computing chemistry which was strongly limited by the computing performance ?shorten the waiting time of the researchers, so getting more results in a certain research hour.
Thank you for your time and consideration of this request.
Sincerely,
Zehua Wei
Wedward@uw.edu
__
Dear STF committee
My name is Shi Jin. I am a second year PhD student in Department of Physics, which is part of the School of Art and Science. My research involves nuclear many-body theory, cold atom and other related system. I use high performance computing daily. I am writing to express very strong support for the proposal as the additional computing support would be beneficial in the following way:
1. shorten the computing time in our project due to use of more CPUs.
2. make it possible to simulate much more complicated systems.
Thank you for listening to our voice
Cheers
Shi Jin
PhD Student
Department of Physics
University of Washington
Seattle WA
__
Dear STF committee,

My name is Stephen Po-Chedley. I am a fifth year graduate student in the Department of Atmospheric Sciences, which is part of the College of the Environment. My research involves climate modeling and remote sensing. I use high performance computing on a weekly basis. I writing to express very strong support for the proposal as the additional computing support would be beneficial.
Being able to quickly run climate models on campus or to break up large computational tasks over a greater number of processes would greatly increase the speed of my research and the research of other students studying climate change. While this sort of computational resource is available commercially (e.g. Amazon) or via federally supported infrastructure, the cost and grant requirements can limit or slow down use. I believe this sort of investment will lead to incredible scientific advancement across several departments. Since I am not a computer scientist, I believe having this technology available would also generate opportunities to learn about high performance computing.
I was excited to see this proposal, because I have recently been working with my research group to increase our collective computer power. I came to the conclusion that more centralized, shared computing resources like those proposed made the most sense.
Thank you for your time and consideration of this request.
Sincerely,
Stephen Po-Chedley

Graduate Research Assistant
Atmospheric Sciences | University of Washington
http://www.atmos.uw.edu/~pochedls
pochedls@uw.edu

__

Dear STF committee,
My name is Benjamin Basanta. I am a second-year graduate student in the department of Biochemistry, which is part of the School of Medicine. My research involves computational modeling of protein structures for their design. I use high performance computing weekly. I am writing to express very strong support for the proposal, as the additional computing support would be beneficial in the following ways: It would allow me to increase the volume of designs generated and the performance of the algorithms used to evaluate them.
Thank you for your time and consideration of this request.
Sincerely,
Benjamin Basanta
basantab@uw.edu

Items

Group Funded Item Unit price Quantity Subtotal
None

Lenovo Blade Center HS23 node

$5,009.00 42 $210,378.00
Description

Each node contains multiple processors (Xeon E5-2670) with 20 compute cores total per node and 64 GB of ram per node with a 600GB internal hard drive and a high speed network card. The nodes are compatible with the existing Hyak infrastructure

Justification

The nodes are required for the HPC computing resource

IBM Disk Server

$160,586.00 1 $160,586.00
Description

The shared disk server contains:

IBM Server System x3650 M4 [outfit with 200TB of total storage)
GPFS-S3.5-12 server
Xanadu 460 Base Sys (8) 12G SAS Dual Ctrl (disk controller and disks)

Justification

This large shared disk (200 TB) is required to support HPC computing across the new nodes.

Total requested: $1,002,098.00

Total funded: $370,964.00

Comments

I was able to sit in on this presentation today at the STF Committee after I presented on another proposal. I wanted to offer my thoughts on what my experience has been with Hyak.

I just wanted to emphasize that Hyak is not intuitive and it is not that easy for a random person to pick up. There is no graphical interface. When you are given access to Hyak, you have to obtain an authentication token, establish an SSH session with Hyak and then write custom scripts to submit your job to the job scheduler. Even as a technology professional, I was not able to be much help to a student who recently wanted to get started with Hyak. I am not well-versed in Linux and High Performance computing and the Hyak team refused to assist us either beyond the Wiki.

I am glad that the Hyak team plans to offer more training resources for those wanting to utilize Hyak. I think that training will be essential. Using a HPC Cluster is a skill that must be acquired. Clearly many already have this skill but not everyone does.

Nick Paradis,
Senior Computer Specialist / IT Administrator, Mechanical Engineering

nparadis@uw.edu — February 11, 2015 @ 6:09 p.m.

No comments