Fwd: Application for CMT hardware

[Brian Madden]

Lets cross our fingers!

-B

---------- Forwarded message ----------
From: Denis Sheahan <Denis....@sun.com>
Date: Thu, Apr 9, 2009 at 1:12 PM
Subject: Re: Application for CMT hardware
To: Brian Madden <untw...@gmail.com>

Hi Brian

I have made the application for the T2 hardware.  

Sun SPARC Enterprise T5120 Server, 8 Core 1.2GHz UltraSPARC T2 processor, 

32GB FBDIMM memory  (16 * 2GB), 2 * 146GB 10K RPM SAS disks, 

1 DVD+/-RW,  4 - 10/100/1000 Ethernet ports, 1 serial port, 4 USB ports, 

1 dedicated PCI-E low profile slot, 2 PCI-E low profile or XAUI (10Gb ethernet) slots, 

2 (N+1) 650W power supplies, Solaris 10 and Java Enterprise System software pre-installed, 

RoHS-6 compliant.  (Standard Configuration)

I let you know how it goes

regards

Denis

On Apr 9, 2009, at 7:18 AM, Brian Madden wrote:

Denis,

Here are your responses.  If you need more information let me know.

Thanks again,
Brian
 

On Wed, Apr 1, 2009 at 10:46 AM, Denis Sheahan <Denis....@sun.com> wrote:

Hi Brian

We can make an application to get you a donation of T2 hardware as part of the Open Sparc initiative.

It is best if a Professor sponsor the application

I need the following information:

University Name 

 University of Pittsburgh

Sun Sales Rep if known - I can find this out if necessary

Unknown

Professor name

Daniel Mosse

Professor email

mo...@cs.pitt.edu

Department

Computer Science

Tel. Number

412-624-8923

Mailing Address:

PO Box is not acceptable

Department of Computer Science,

University of Pittsburgh

Sennott Square Bldg.

210 S. Bouquet St.

Pittsburgh, PA 15260

Shipping Destination:

if different from mailing address

We also need to make a case for the donation, need the Project details and how it

can help OpenSPARC.   I can edit this if you send me the following details

What are the goals, objectives of this project and how will OpenSPARC be used?

 

There are many goals of the XOmB project, however the primary goals are to design a functional and useful exokernel both for educational and practical purposes.  We believe that the exokernel design will allow for greater flexibility and control over hardware with minimal control by the OS.  One of the major motivaters in the decision to implement an exokernel is due to the lack of forced abstraction.  We are pushing as much as possible to userland, and believe this will allow for safe, rapid, prototype and implement new subsystems.  To implement the kernel we would like to utilize Sun hardware for a number of purposes.  The first, the hardware threading capabilities, will afford minimal context switching, something we were trying to minimize in software by pushing as much as possible to userland anyway.  Now even in the case where a context switch will be necessary, by having multiple hardware threads we may be able to eliminate the need all together.  Second, the virtual domains provided by the hardware will be used to implement safe userland drivers.  One of the problems we've encountered while working on XOmB is the difficulty of moving some drivers such as the network drivers to userland.  By utilizing virtual domains we hope to achieve safe userland drivers.
 

Does this project require any Sun engineering, technical or IP support?

No

Is the organization willing to make research results available and/or share curriculum to OpenSPARC community?

 

Yes, as firm believers in open source we would love to contribute back to the OpenSPARC community.  With a strong interest in education, our project has intended to be a stepping stone for anyone interested in operating system design and implementation.  We are hoping to provide newcomers a platform to both learn from and work in, as well as comprehensive documents describing and explaining our implementation choices and tradeoffs.  We hope to serve future generations of systems programmers with the ground work that we are laying right now.
 

When(month, year) will the organization begin teaching OpenSPARC in their curriculum and when sharing their curriculum on opensparc.net?

 

As a completely member driven organization we are often slow moving due to other obligations, however we've continued to work for the past two years and will continue to work on XOmB for years to come.  While we are not the fastest moving project, we have made significant progress and have already contributed barebones X86-64 operating system code to the open source community.  As our project is open source, OpenSPARC community members will be more than welcome to browse our ever growing source tree and documentation.  When we hit major milestones we would be more than happy to contribute those in some way to the OpenSPARC community as presentations or tutorial/demonstration packages.
 

- WIll the research be part of teaching at the university

 

As of right now the XOmB project is run and perpetuated entirely by the student body.  We do have some faculty support and would love to get more involved with teaching, however time constraints and other obligations have limited this potential.  We would however love to have XOmB be a bigger part of the teaching at the University and will continue to work with systems professors at the University to try to work XOmB in as either homework or lecture material.

Let me know if you need  more information

regards

Denis

--
Brian Madden
www.pittgeeks.org

--
Brian Madden
www.pittgeeks.org

--
Brian Madden
www.pittgeeks.org