-

This site is deprecated and will be decommissioned shortly. For current information regarding HPC visit our new site: hpc.njit.edu

Difference between revisions of "ForPCCStarTable"

From NJIT-ARCS HPC Wiki
Jump to: navigation, search
(Importing text file)
 
(Importing text file)
 
(15 intermediate revisions by the same user not shown)
Line 2: Line 2:
 
<div class="noautonum">__TOC__</div>
 
<div class="noautonum">__TOC__</div>
  
−
== NSF Campus Cyberinfrastucture Program Area 4: Public-Access GPUs Proposal ==
+
== NSF CC*Program Area 4: Public-Access GPUs Proposal: Project Timeline ==
  
 
== 1. ==
 
== 1. ==
 +
<strong> What </strong>
 +
A summary table of the science drivers and their computing environments
  
−
=== What ===
+
<strong> Who </strong>
−
A summary table of the science drivers and their computingenvironments
+
−
 
+
−
=== Lead ===
+
 
DP
 
DP
  
−
=== Due by ===
+
<strong> Due by </strong>
 
Wed 10 Feb 2021
 
Wed 10 Feb 2021
  
−
=== Percent done ===
+
<strong> Percent done </strong>
 
30
 
30
  
−
=== NSF guidelines ===
+
<strong> NSF guidelines </strong>
 
These requirements should be specified in clear terms
 
These requirements should be specified in clear terms
 
reflecting a specific understanding of the required computing resources and
 
reflecting a specific understanding of the required computing resources and
 
environment, for example, CPU/GPU type, compute job profile parameter
 
environment, for example, CPU/GPU type, compute job profile parameter
 
ranges, core count ranges per job, times to completion or as part of a
 
ranges, core count ranges per job, times to completion or as part of a
−
composition or scientific workflow profile;
+
composition or scientific workflow profile
 +
 
 +
<strong> Comments </strong>
 +
<em>None</em>
 +
 
 +
== 2. ==
 +
<strong> What </strong>
 +
Platform architecture specifying cluster components, including
 +
compute node type and count, per-node memory, interconnect fabric, storage,
 +
and open source software/platform
 +
 
 +
<strong> Who </strong>
 +
GW, MH, KW
 +
 
 +
<strong> Due by </strong>
 +
Wed 10 Feb 2021
 +
 
 +
<strong> Percent done </strong>
 +
0
 +
 
 +
<strong> NSF guidelines </strong>
 +
<em>None</em>
 +
 
 +
<strong> Comments </strong>
 +
<em>None</em>
 +
 
 +
== 3. ==
 +
<strong> What </strong>
 +
An open source-based approach to cluster monitoring, measurement, management, and instrumentation
 +
 
 +
<strong> Who </strong>
 +
GW, (MH)
 +
 
 +
<strong> Due by </strong>
 +
Wed 10 Feb 2021
 +
 
 +
<strong> Percent done </strong>
 +
0
 +
 
 +
<strong> NSF guidelines </strong>
 +
<em>None</em>
 +
 
 +
<strong> Comments </strong>
 +
--
 +
 
 +
== 4. ==
 +
<strong> What </strong>
 +
A sustainability plan addressing the institution's commitment to
 +
providing an ongoing level of sustained access to computational resources
 +
 
 +
<strong> Who </strong>
 +
KS
 +
 
 +
<strong> Due by </strong>
 +
Fri 12 Feb 2021
 +
 
 +
<strong> Percent done </strong>
 +
0
 +
 
 +
<strong> NSF guidelines </strong>
 +
<em>None</em>
 +
 
 +
<strong> Comments </strong>
 +
--
 +
 
 +
== 5. ==
 +
<strong> What </strong>
 +
A High-Performance Network Connectivity and Specification
 +
 
 +
<strong> Who </strong>
 +
GW, MH
 +
 
 +
<strong> Due by </strong>
 +
Fri 12 Feb 2021
 +
 
 +
<strong> Percent done </strong>
 +
0
 +
 
 +
<strong> NSF guidelines </strong>
 +
Proposals must describe the network connectivity of the proposed computing resource, both
 +
intra-campus [for example, the campus network path(s) connecting the
 +
resource with the researchers and driving science applications on campus],
 +
and inter-campus (for example, showing the network path connecting with the
 +
regional exchange point or Internet2). Proposals should include a network
 +
diagram showing the connected topology of the proposed cluster resource.
 +
Proposals should include in their plans the deployment of a PerfSonar based
 +
network performance measurement capability to initially measure achievable
 +
end-to-end network performance for scientific data flows between the
 +
resource and relevant end points of researchers.
 +
 
 +
 
 +
<strong> Comments </strong>
 +
--
 +
 
 +
== 6. ==
 +
<strong> What </strong>
 +
Description of the cluster as a Shared Resource Intra-campus and
 +
Inter-campus
 +
 
 +
<strong> Who </strong>
 +
DP, GW
 +
 
 +
<strong> Due by </strong>
 +
Fri 12 Feb 2021
 +
 
 +
<strong> Percent done </strong>
 +
0
 +
 
 +
<strong> NSF guidelines </strong>
 +
The Cluster as a Shared Resource Intra-campus and Inter-campus: Proposals
 +
should describe (1) their approach to sharing the proposed computing
 +
resource across the science drivers and researchers at their institution;
 +
(2) how the resource will be accessed by external research groups; and (3)
 +
how the resource is coordinated with external resources allowing the
 +
institution's researchers to seamlessly access computing resources at other
 +
campuses, regional and national computing resources, and/or production
 +
cloud resources, if appropriate.
 +
 
 +
Proposals are required to commit to a minimum of 20% shared time on the
 +
cluster and describe their approach to making the cluster available as a
 +
shared resource external to the campus, with access and authorization
 +
according to local administrative policy. Conversely, the proposal should
 +
describe the approach to providing *on-demand* access to additional
 +
external computing resources for its targeted on-campus users and projects.
 +
One possible approach to implementing such a federated distributed
 +
computing solution is joining a multi-campus or national federated system
 +
such as the Open Science Grid. Whatever opportunistic, federated, scalable,
 +
distributed computing platform is chosen, the proposal is expected to
 +
justify the choice by including a discussion on the shared platform's track
 +
record in the community, its current scientific computing production
 +
capability, and its scaling properties. Proposals are encouraged to include
 +
a letter of collaboration from the selected platform and describe how they
 +
will track and report on meeting the 20% extramural usage goal each year.
 +
 
  
−
=== Lead comments ===
+
<strong> Comments </strong>
 +
--

Latest revision as of 19:48, 1 February 2021

NSF CC*Program Area 4: Public-Access GPUs Proposal: Project Timeline

1.

What A summary table of the science drivers and their computing environments

Who DP

Due by Wed 10 Feb 2021

Percent done 30

NSF guidelines These requirements should be specified in clear terms reflecting a specific understanding of the required computing resources and environment, for example, CPU/GPU type, compute job profile parameter ranges, core count ranges per job, times to completion or as part of a composition or scientific workflow profile

Comments None

2.

What Platform architecture specifying cluster components, including compute node type and count, per-node memory, interconnect fabric, storage, and open source software/platform

Who GW, MH, KW

Due by Wed 10 Feb 2021

Percent done 0

NSF guidelines None

Comments None

3.

What An open source-based approach to cluster monitoring, measurement, management, and instrumentation

Who GW, (MH)

Due by Wed 10 Feb 2021

Percent done 0

NSF guidelines None

Comments --

4.

What A sustainability plan addressing the institution's commitment to providing an ongoing level of sustained access to computational resources

Who KS

Due by Fri 12 Feb 2021

Percent done 0

NSF guidelines None

Comments --

5.

What A High-Performance Network Connectivity and Specification

Who GW, MH

Due by Fri 12 Feb 2021

Percent done 0

NSF guidelines Proposals must describe the network connectivity of the proposed computing resource, both intra-campus [for example, the campus network path(s) connecting the resource with the researchers and driving science applications on campus], and inter-campus (for example, showing the network path connecting with the regional exchange point or Internet2). Proposals should include a network diagram showing the connected topology of the proposed cluster resource. Proposals should include in their plans the deployment of a PerfSonar based network performance measurement capability to initially measure achievable end-to-end network performance for scientific data flows between the resource and relevant end points of researchers.


Comments --

6.

What Description of the cluster as a Shared Resource Intra-campus and Inter-campus

Who DP, GW

Due by Fri 12 Feb 2021

Percent done 0

NSF guidelines The Cluster as a Shared Resource Intra-campus and Inter-campus: Proposals should describe (1) their approach to sharing the proposed computing resource across the science drivers and researchers at their institution; (2) how the resource will be accessed by external research groups; and (3) how the resource is coordinated with external resources allowing the institution's researchers to seamlessly access computing resources at other campuses, regional and national computing resources, and/or production cloud resources, if appropriate.

Proposals are required to commit to a minimum of 20% shared time on the cluster and describe their approach to making the cluster available as a shared resource external to the campus, with access and authorization according to local administrative policy. Conversely, the proposal should describe the approach to providing *on-demand* access to additional external computing resources for its targeted on-campus users and projects. One possible approach to implementing such a federated distributed computing solution is joining a multi-campus or national federated system such as the Open Science Grid. Whatever opportunistic, federated, scalable, distributed computing platform is chosen, the proposal is expected to justify the choice by including a discussion on the shared platform's track record in the community, its current scientific computing production capability, and its scaling properties. Proposals are encouraged to include a letter of collaboration from the selected platform and describe how they will track and report on meeting the 20% extramural usage goal each year.


Comments --

Retrieved from "https://wiki.hpc.arcs.njit.edu/index.php?title=ForPCCStarTable&oldid=345"