Sphenix-tpc-l mailing list

[Thomas K Hemmick <Thomas.Hemmick AT stonybrook.edu>]

Hi guys

Here is a snippet of what I wrote for Bill's successful application for chips.  Please don't repeat these exact words, but it is the style of thing that they are apparently looking for as it did work.  In particular, coming up with additional examples of industrial applications for shipping around large volumes of data at high speed is likely a good idea.

============

As a field, physics has the well-earned reputation for efficient and effective data transfer from a myriad of sensors (our detectors) toward long term storage.  The effectiveness of an experiment is most often directly related to the quality and quantity of both throughput of their data acquisition system and the selectiveness of the data chosen for recording.  In physics this process if known as "triggering" in which a myriad of sensors must report their data to a decision-making node that will with certainty pass on the most interesting data but as-effectively-as-possible reduce the uninteresting data.

The most modern of physics detectors deploy ASICs to digitize and format the raw (often vanishingly small) signals into numbers that can be passed on through standard commercial media (e.g. fiber optic).  However, intelligent decision-making is essentially impossible at the front end since each ASIC has a limited view of the full picture (merely the few signals that it services).  Effective decision-making requires that these data be aggregated together so that a complete or nearly complete picture is available for the execution of proper decisions.

Although this paradigm sounds rather ideal, it can be overwhelming to funnel all available data to a single point.  A simplistic attempt at that job would result in excellent decisions made at a low rate.  Instead, one must at each stage of data aggregation apply compression and filtering to throttle down the data to the point that best and most valuable data is delivered to the decision node allowing the decisions to be both on excellent information and also made rapidly.  This fundamental capability is what distinguishes good from great physics experiments and is rapidly becoming the same issue in the commercial world.

We plan to develop a so-called "Data Aggregation Module" using the XILINX chips requested in this solicitation.  This module would support 48 sensors (via fiber optic IO) and would process, format, and throttle the data before delivery to a 16-lane PCI express bus.  Because the IO at both ends is quite generic and non-specific to our physics experiment (sPHENIX which is successor to PHENIX that held the world record in DAQ throughput for more than a decade), it can have a variety of commercial applications as well.  These work for any system where a widespread array of sensors must have their data aggregated into a decision node.  Examples include robotics, autonomous vehicles, surveillance footage, or simply any application wherein real world data needs to be throttled and presented intelligently to a decision node.

=============

Tom

On Wed, Apr 24, 2019 at 7:52 AM Sasha Milov <alexander.milov AT weizmann.ac.il> wrote:

hi all

can you advice what to fill into the xilinx request fields?

thanks

a.m.

On Wed, Mar 13, 2019 at 11:06 PM matt posik <mposik1983 AT gmail.com> wrote:

Thanks!

I just submitted the donation request. I will keep you posted.

Thanks,

Matt

On Wed, Mar 13, 2019, 4:09 PM Huang, Jin <jhuang AT bnl.gov> wrote:

Dear Matt

 

Sure. Will be happy to. Please find the suggestion below.

 

Cheers

 

Jin

 

______________________________

 

 

Following Tom's suggestion in the TPC general meeting today, I am writing to follow up with the details of the Xilinx University Donation program. The idea to apply to this program originates from a comment received during the CD-1 review. Tom worked through this process and received five FELIX FPGAs which represents a significant cost-saving in the TPC DAM R&D and production. We will appreciate further helps from other university collaborators too.

 

It does take some effort to apply to the Xilinx University Donation program. At the first step, the donation can be requested at https://www.xilinx.com/support/university/donation-program.html   

* Our research project (high throughput data acquisition R&D or some fancier names)

* Software Development Tools: (not required)

* In the Hardware Development Boards section

** Hardware Development Boards: Other

** List any other products you are requesting: Kintex-7 Ultrascale FPGA with the part number of XCKU115-2FLVF1924E.

** How many boards are you requesting: (about five seems possible)

After passing the first round of evaluation,  Xilinx will contact the PI regarding information on the project for their further assessment.

 

Please feel free to let Tom and I know if there is any question or suggestions.

 

______________________________

 

Jin HUANG

 

Associate Physicist

Brookhaven National Laboratory

Physics Department, Bldg 510 C

Upton, NY 11973-5000

 

Office: 631-344-5898

Cell:   757-604-9946

______________________________

 

From: Sphenix-tpc-l <sphenix-tpc-l-bounces AT lists.bnl.gov> On Behalf Of matt posik
Sent: Wednesday, March 13, 2019 4:04 PM
To: Klaus Dehmelt <klaus.dehmelt AT stonybrook.edu>
Cc: sphenix_tpc_general AT skipper.physics.sunysb.edu
Subject: Re: [Sphenix-tpc-l] Bi-weekly sPHENIX TPC General Meeting on Wednesday, March 13, 3 PM - 4 PM EDT

 

Hi All,

 

I took a look through my emails and can not find the one that was sent regarding the chips that were mentioned during today's meeting. 

Is there any chance that we can recirculate that email?

 

Thanks,
Matt

 

On Tue, Mar 12, 2019 at 9:06 PM Klaus Dehmelt <klaus.dehmelt AT stonybrook.edu> wrote:

Dear all,

 

we will continue with our  bi-weekly meeting tomorrow at 3 PM EDT. You can find a tentative agenda at https://indico.bnl.gov/event/5590/

Please let me know if you have anything to present.

Please note that in the US we are on DST now.

 

 

BlueJeans coordinates are

 

To join the Meeting:
https://bluejeans.com/881710315

To join via Browser:
https://bluejeans.com/881710315/browser

To join with Lync:
https://bluejeans.com/881710315/lync

To join via Cisco Jabber Video:
https://bluejeans.com/881710315/jabber


To join via Room System:
Video Conferencing System: bjn.vc<http://bjn.vc/> -or- 199.48.152.152
Meeting ID: 881710315

To join via Phone:
1) Dial:
+1.408.740.7256<tel:%2B1.408.740.7256>
+1.888.240.2560<tel:%2B1.888.240.2560>(US Toll Free)
+1.408.317.9253<tel:%2B1.408.317.9253>(Alternate number)
(see all numbers - http://bluejeans.com/numbers)
2) Enter Conference ID: 881710315

 

Cheers,

 

Klaus

********************************************

Klaus Dehmelt

Stony Brook University

Dept. of Physics and Astronomy

Stony Brook, NY 11794-3800, USA

Office C108

Phone +1 631 632 8115

E-mail klaus.dehmelt AT stonybrook.edu

My-SBU-webpage

*********************************************

 

 

--

 Matthew Posik

 Assistant Research Professor 

 Temple University 

 Department of Physics

 SERC 

 1925 N. 12th St.

 Philadelphia, PA 19122

 USA

 

 TU Office: SERC Room 451

Phone: 215-204-8690

 

Physics Office: SERC Room 406/4th Floor

Phone: 215-204-7421

--

Department of Particle Physics and Astrophysics
Weizmann Institute of Science
234 Herzl str. Rehovot 76100, Israel
tel/fax: +972 8 934 2848