Uli Schäfer 1 FPGAs for high performance – high density applications Intro Requirements of future trigger systems Features of recent FPGA families 9U *

Slides:



Advertisements
Similar presentations
By Sunil G. Kulkarni, SO/F, Pelletron-Linac Facility, BARC-TIFR. 21/01/2011 ASET.
Advertisements

Reconfigurable Computing (EN2911X, Fall07) Lecture 04: Programmable Logic Technology (2/3) Prof. Sherief Reda Division of Engineering, Brown University.
Digital RF Stabilization System Based on MicroTCA Technology - Libera LLRF Robert Černe May 2010, RT10, Lisboa
Uli Schäfer JEM Status and Test Results Hardware status JEM0 Hardware status JEM1 RAL test results.
Spartan II Features  Plentiful logic and memory resources –15K to 200K system gates (up to 5,292 logic cells) –Up to 57 Kb block RAM storage  Flexible.
Reconfigurable Computing - Clocks John Morris Chung-Ang University The University of Auckland ‘Iolanthe’ at 13 knots on Cockburn Sound, Western Australia.
FPGA-Based System Design: Chapter 3 Copyright  2004 Prentice Hall PTR SRAM-based FPGA n SRAM-based LE –Registers in logic elements –LUT-based logic element.
Uli Schäfer JEM Status and plans Hardware status JEM0 Hardware status JEM1 Plans.
GOLD Generic Opto Link Demonstrator Uli Schäfer 1.
Phase-0 Topological Processor Uli Schäfer Johannes Gutenberg-Universität Mainz Uli Schäfer 1.
Programmable logic devices / tools Programmable logic devices are digital logic devices, providing combinatorial logic (gates, look-up tables) and flip-flops.
High-speed optical links and processors for the ATLAS Level-1 Calorimeter Trigger upgrade B.Bauss, V.Büscher, R.Degele, A.Ebling, W.Ji, C.Meyer, S.Moritz,
Phase-0 topological processor Uli Schäfer Johannes Gutenberg-Universität Mainz Uli Schäfer 1.
Level-1 Topology Processor for Phase 0/1 - Hardware Studies and Plans - Uli Schäfer Johannes Gutenberg-Universität Mainz Uli Schäfer 1.
Uli Schäfer 1 S-L1Calo upstream links architecture -- interfaces -- technology.
Uli Schäfer 1 (Not just) Backplane transmission options.
Programmable logic and FPGA
Uli Schäfer 1 CP/JEP backplane test module What’s the maximum data rate into the S-CMM for phase-1 upgrade ?
Hardware status GOLD Generic Opto Link Demonstrator Uli Schäfer 1.
Uli Schäfer JEM hardware / test JEM0 test programme Mainz standalone RAL sub-slice test JEM re-design Heidelberg slice test.
Configuration. Mirjana Stojanovic Process of loading bitstream of a design into the configuration memory. Bitstream is the transmission.
Uli Schäfer 1 (Not just) Backplane transmission options.
Introduction to FPGA and DSPs Joe College, Chris Doyle, Ann Marie Rynning.
General FPGA Architecture Field Programmable Gate Array.
EKT303/4 PRINCIPLES OF PRINCIPLES OF COMPUTER ARCHITECTURE (PoCA)
Highest Performance Programmable DSP Solution September 17, 2015.
ASIC 120: Digital Systems and Standard-Cell ASIC Design Tutorial 4: Digital Systems Concepts November 16, 2005.
Lecture #35 Page 1 ECE 4110– Sequential Logic Design Lecture #35 Agenda 1.Clocking Techniques Announcements Next: 1.HW #15 due. 2.Final review.
Electronics in High Energy Physics Introduction to Electronics in HEP Field Programmable Gate Arrays Part 1 based on the lecture of S.Haas.
Hardware status GOLD Generic Opto Link Demonstrator Assess the use of optical backplane connectivity for use on L1Calo Uli Schäfer 1.
J. Christiansen, CERN - EP/MIC
® SPARTAN Series High Volume System Solution. ® Spartan/XL Estimated design size (system gates) 30K 5K180K XC4000XL/A XC4000XV Virtex S05/XL.
FPGA-Based System Design: Chapter 3 Copyright  2004 Prentice Hall PTR FPGA Fabric n Elements of an FPGA fabric –Logic element –Placement –Wiring –I/O.
FPGA-Based System Design: Chapter 3 Copyright  2004 Prentice Hall PTR Topics n FPGA fabric architecture concepts.
Programmable Logic Devices
Introduction to FPGA Created & Presented By Ali Masoudi For Advanced Digital Communication Lab (ADC-Lab) At Isfahan University Of technology (IUT) Department.
Basic Sequential Components CT101 – Computing Systems Organization.
BR 1/991 Issues in FPGA Technologies Complexity of Logic Element –How many inputs/outputs for the logic element? –Does the basic logic element contain.
S. Rave, U. Schäfer For L1Calo Mainz
ATLAS Trigger / current L1Calo Uli Schäfer 1 Jet/Energy module calo µ CTP L1.
JFEX Uli Schäfer 1. Constraints & Numerology Assumption: one crate, several modules. Each module covers full phi, limited eta range Data sharing with.
The ATLAS Global Trigger Processor U. Schäfer Phase-2 Upgrade Uli Schäfer 1.
JFEX Uli Schäfer 1 Mainz. L1Calo Phase-1 System Uli Schäfer 2 CPM JEM CMX Hub L1Topo ROD JMM PPR From Digital Processing System CPM JEM CMX Hub L1Topo.
ESS | FPGA for Dummies | | Maurizio Donna FPGA for Dummies Basic FPGA architecture.
Agata Week – LNL 14 November 2007 LLP ATCA Carrier Status M. Bellato on behalf of the LLP Carrier Working Group.
Introduction to Field Programmable Gate Arrays Lecture 1/3 CERN Accelerator School on Digital Signal Processing Sigtuna, Sweden, 31 May – 9 June 2007 Javier.
XLV INTERNATIONAL WINTER MEETING ON NUCLEAR PHYSICS Tiago Pérez II Physikalisches Institut For the PANDA collaboration FPGA Compute node for the PANDA.
JFEX Uli Schäfer 1 Mainz. L1Calo Phase-1 System Uli Schäfer 2 CPM JEM CMX Hub L1Topo ROD JMM PPR From Digital Processing System CPM JEM CMX Hub L1Topo.
High Speed Digital Systems Lab Spring/Winter 2010 Project definition Instructor: Rolf Hilgendorf Students: Elad Mor, Ilya Zavolsky Integration of an A/D.
DDRIII BASED GENERAL PURPOSE FIFO ON VIRTEX-6 FPGA ML605 BOARD PART B PRESENTATION STUDENTS: OLEG KORENEV EUGENE REZNIK SUPERVISOR: ROLF HILGENDORF 1 Semester:
ESS | FPGA for Dummies | | Maurizio Donna FPGA for Dummies Modern FPGA architecture.
EE121 John Wakerly Lecture #15
16 February 2011Ian Brawn1 The High Speed Demonstrator and Slice Demonstrator Programme The Proposed High-Speed Demonstrator –Overview –Design Methodology.
Trigger Hardware Development Modular Trigger Processing Architecture Matt Stettler, Magnus Hansen CERN Costas Foudas, Greg Iles, John Jones Imperial College.
Trigger Workshop: Greg Iles Feb Optical Global Trigger Interface Card Dual CMC card with Virtex 5 LX110T 16 bidirectional.
Clock Networks and PLLs in Altera’s Stratix III Devices VLSI Systems I Fall 2007 Hamid Abbaalizadeh.
Delivered by.. Love Jain p08ec907. Design Styles  Full-custom  Cell-based  Gate array  Programmable logic Field programmable gate array (FPGA)
FPGA-Based System Design: Chapter 3 Copyright  2004 Prentice Hall PTR Topics n FPGA fabric architecture concepts.
Introduction to Field Programmable Gate Arrays (FPGAs) EDL Spring 2016 Johns Hopkins University Electrical and Computer Engineering March 2, 2016.
GBT protocol implementation on Xilinx FPGAs Csaba SOOS PH-ESE-BE-OT.
Multi-Gigabit transmission BLT  GOLD Andreas Ebling, Isabel Koltermann, Jonas Kunze Andi Ebling 1.
Status and Plans for Xilinx Development
집적회로설계 1 Spring 2007 Prof. Sang Sik AHN Signal Processing LAB.
Issues in FPGA Technologies
Programmable Logic Devices
Topics SRAM-based FPGA fabrics: Xilinx. Altera..
Electronics for Physicists
The Xilinx Virtex Series FPGA
The Xilinx Virtex Series FPGA
Electronics for Physicists
Presentation transcript:

Uli Schäfer 1 FPGAs for high performance – high density applications Intro Requirements of future trigger systems Features of recent FPGA families 9U * 40cm ATCA µTCA/AMC

Uli Schäfer 2 Intro : FPGA basics Large array of logic cells ~100k combinatorial : map any 4-variable equation into 4-input lookup table (LUT) sequential : flip-flop (FF) Interconnect ‘wires’ : segmented routing switch boxes connecting wires and logic cells dedicated global clock trees into all cells I/O pads route internal signals to pins define signal standard Clock management : condition the incoming clocks and generate multiples and fractions phase lock loop (PLL) delay lock loop (DLL) Cores RAM blocks for data storage Many other cores introduced in recent years, see below… Functionality of FPGA is defined upon power up by reading in a configuration data stream from non-volatile memory

Uli Schäfer 3 Higher granularity along with the need to keep fraction of duplicated channels at reasonable level requires higher density designs (higher channel count per FPGA and per module) Typical form factors and therefore card edges tend to get smaller: current L1calo ‘standard’ is 9U*400mm Telecom standards :ATCA: 8U * 280mm µTCA (AMC): 73.5 * 180.6mm Narrower data paths, but 10/12.5 Gbps per link Single ended data transmission stretched to limits at data rates and signal standards employed on current L1calo modules  go differential FPGA features in demand: on-chip high-speed serial links ( incoming trigger tower data ) differential high-speed data buses ( FIO ) logic resources (fabric) arithmetic units in case more demanding algorithms required suitable pinout and I/O properties for high density / high speed designs (signal integrity) Requirements of future L1calo processors

Uli Schäfer 4 Recent FPGA features/improvements Increase in clock speed Increase in logic resources (fabric) Increase in block memory Further hard cores: Processors Gbps serializer/deserializer units for parallel source synchronous data transmission (clock forwarding) Multi-Gbps links with embedded clock DSP / arithmetic circuitry I/O Differential high-speed standards (LVDS,PECL,…) Low voltage single ended Internal termination differential : 100Ω single ended : ‘programmable’ impedance On-chip bypass capacitors and signal integrity-optimised pinout

Uli Schäfer 5 Resources by manufacturer (*) All FPGA families have some means of phase adjustment (L,X) or multi-phase sampling (A) on their input lines, as well as SerDes. Not all features available on all I/O lines Virtex-4 have 6.5 Gbps serial links Altera Stratix III Lattice SC Xilinx Virtex-5 Global clocks MHz Clock management PLL1286 Clock management DLL012 Serial links (IIGX)3.83.7Gb/s Parallel differential I/O w. full speed / SerDes / phase (*) pairs Gb/s

Uli Schäfer 6 Lattice SC input delay control 144 tap delay unit, 40ps/tap 9-tap sampling within a window allows for calculation of optimum sampling point and automatic delay adjustment Available on every other differential pair only

Uli Schäfer 7 Xilinx Virtex-5 source synchronous interface (Gbps, double data rate) SerDes and programmable delay unit available in all I/O pads No hard core phase aligner, use soft core (fabric) to track data Eliminate cycle-to-cycle jitter at source with a PLL Due to the DLL the data are clocked into the deserialiser with a clock edge generated just a few ticks before the data bit  Low frequency jitter doesn’t matter

Uli Schäfer 8 Xilinx serial links (MGT) 3.7 Gbps serial link, low power 100mW/ch up to 24 channels per device  Data rate and channel count match SNAP12 optical link Transmitter:programmable signal level pre-emphasis Receiver: equalization Latency (RX+TX) : minimum of 12.5 ticks of byte clock byte clock could be as high as 320 MHz for a 40 MHz based system 40ps reference clock jitter requirement Re-design LHC clock distribution Use jitter attenuators (silabs.com) Go asynchronous Use local Xtal Require re-synchronisation to LHC clock (latency !) Allow for standard data rates / standard components

Uli Schäfer 9 Xilinx Virtex-5 resources (maximum) Resource Virtex-5(in XCV1000E) 6-input LUTs:200k(25k*4-input) Flipflops:200k(25k) Distributed RAM : 3.4 Mb(400kb) Block RAM :11.6Mb(400kb) “DSP” 25*18 bit multiplier/accumulator:640 PCI Express endpoint 1 Ethernet MAC (with internal or external PHY) 4

Uli Schäfer 10 Summary / Outlook Logic density gone up considerably. A single FPGA is equivalent to almost a full L1calo processor module Current FPGA families allow for high data rates on both ‘parallel’ and high-speed serial links Aggregate bandwidth is higher on ‘parallel’ links Xilinx Virtex-5 has same high-speed I/O resources on all user pins and is therefore particularly useful for typical trigger circuitry : many-in  few-out On-chip links with embedded clock do have surprisingly low latency but might need additional synchroniser stages due to jitter requirements Xilinx development boards ML506/ML555 available  let’s start work. Explore synchronous / asynchronous schemes