Download Document

FABRIC
A pilot study of distributed correlation
Huib Jan van Langevelde
Ruud Oerlemans
Sergei Pogrebenko
and many other JIVErs…
Aim of the project
• Research the possibility of distributed correlation
• Using the Grid for getting the CPU cycles
• Can it be employed for the next generation VLBI correlation?
• Exercise the advantages of software correlation
• Using floating point accuracy and special filtering
• Explore (push) the boundaries of the Grid paradigm
• “Real time” applications, data transfer limitations
• To lead to a modest size demo
• With some possible real applications:
• Monitoring EVN network performance
• Continuous available eVLBI network with few telescopes
•Monitoring transient sources
•Astrometry, possibly of spectral line sources
• Special correlator modes: spacecraft navigation, pulsar gating
• Test bed for broadband eVLBI research
huib 23/6/06
Something to try on the roadmap for the next generation correlator,
even if you do not believe it is the solution…
NGC Groningen 29 June 2006
2/12
SCARIe FABRIC
•EC funded project EXPReS (03/2006)
• To turn eVLBI into an operational system
• Plus: Joint Research Activity: FABRIC
• Future Arrays of Broadband Radio-telescopes on Internet Computing
•One work-package on 4Gb/s data acquisition and transport
(Jodrell Bank, Metsahovi, Onsala, Bonn, ASTRON)
•One work-package on distributed correlation (JIVE, PNSC Poznan)
•Dutch NWO funded project SCARIe (10/2006)
• Software Correlator Architecture Research and Implementation for eVLBI
• Use Dutch Grid with configurable high connectivity
• Software correlation with data originating from JIVE
•Complementary projects with matching funding
• International and national expertise from other partners
• Poznan Supercomputer centre
• SARA and University of Amsterdam
• Total of 9 man year at JIVE, plus some matching from staff
huib 23/6/06
• plus similar amount at partners
NGC Groningen 29 June 2006
3/12
Previous experience on Software correlation
• Builds on previous
experience at JIVE
• regular and automated network
performance tests
• Using Japanese software
correlator from NICT
• Huygens extreme narrow band
correlation
• Home grown superFX with subHz resolution
huib 23/6/06
NGC Groningen 29 June 2006
4/12
Basic idea
•Use the Grid for correlation
•CPU cycles on compute nodes
•The Net could be crossbar switch?
•Correlation will be asynchronous
•Based on floating point arithmetic
•Portable code, standard environment
typical VLBI problems
huib 23/6/06
description
1 Gb/s full array
typical eVLBI continuum
typical spectral line
FABRIC demo
future VLBI
N
N
data-rate
N
telescopes subbands [Mb/s] spect/prod
16
16
1024
16
8
8
128
16
10
2
16
512
4
2
16
32
32
32
4096
256
Tflops
83.89
2.62
16.38
0.16
21474.84
Rough estimate based on XF correlation
NGC Groningen 29 June 2006
5/12
Work packages
• Grid resource allocation
• Grid workflow management
• Tool to allocate correlator resources and schedule correlation
• Data flow from telescopes to appropriate correlator resources
• Expertise from the Poznan group in Virtual Laboratories
• Will this application fit on Grid?
• As it is very data intensive
• And time-critical if not real-time
• Software correlation
• correlator algorithm design
• High precision correlation on standard computing
• Scalable to cluster computers
• Portable for grid computers and interfaced to standard
middleware
• Interactive visualization and output definition
huib 23/6/06
• Collect & merge data in EVN archive
• Standard format and proprietary rights
NGC Groningen 29 June 2006
6/12
Workflow Management
• Must interact with normal VLBI schedules
• Divide data, route to compute nodes, setup correlation
• Dynamic resource allocation, keep up with incoming data!
Effort from Poznan, based on their Virtual Lab.
huib 23/6/06
NGC Groningen 29 June 2006
7/12
Topology
•Slice in time
• Every node gets an interval
• A “new correlator” for every time
slice
• Employ clusters computers at
nodes
• Minimizes total data transport
• Bottleneck at compute node
• Probably good connectivity at
Grid nodes anyway
• Scales perfectly
• Easily estimated how many
nodes are needed
• Works with heterogeneous
nodes
• But leaves sorting to compute
nodes
• Memory access may limit
effectiveness
huib 23/6/06
NGC Groningen 29 June 2006
•Slice in baseline
• Assign a (or a range of)
products to a certain node
• E.g. two data streams meet in
some place
• Transport Bottleneck at
sources (telescopes)
• Maybe curable with multicast
transport mechanism which
forks at network nodes
• Some advantage when local
nodes at telescopes
• Does not scale very simply
• Simple schemes for ½N2 nodes
• Need to re-sort output
• But reduces the compute
problem
• Using the network as the
cross-bar switch
8/12
Broadband software correlation
Station 1
Station 2
Station N
EVN Mk4 equivalents
Raw data BW=16 MHz,
Mk4 format on Mk5 disk
From Mk5 to linux disk
Raw data 16 MHz,
Mk4 format on linux disk
DIM,TRM,
CRM
Channel extraction
Extracted data
SU
Pre-calculated,Delay tables
DCM,DMM,
FR
Delay corrections
Delay corrected data
Correlator
Chip
Correlation. SFXC
huib 23/6/06
Data Product
NGC Groningen 29 June 2006
9/12
Better SNR than Mk4 hardware
huib 23/6/06
NGC Groningen 29 June 2006
10/12
Software correlation
•Working on benchmarking
• Single core processors so far
• Different CPU’s available
SFX correlator: measuring CPU on single core
Auto and Cross correlations
4000
3500
• Already quite efficient
2500
CPU time (s)
• More work on memory
performance
3000
jop32
2000
pcint
cedar
1500
1000
•Must deploy on cluster
computers
•And then on Grid
500
0
0
4
8
12
16
20
24
28
32
36
40
44
number of stations
SFX correlator:CPU contributions
4000
•Organize the output to be
used for astronomy
3500
3000
CPU time (s)
2500
cedar
FFT only
2000
I/O only
FFT Auto
1500
1000
huib 23/6/06
500
0
0
4
8
12
16
20
24
28
32
36
40
44
number of stations
NGC Groningen 29 June 2006
11/12
Huygens, software correlation
•Experience with software
correlation from Huygens
•Carrier signal from Titan
lander
•Recorded on Mk5 disk system
• Saved Doppler data experiment
•Requires extreme narrow
band correlation
•And solar system model
•May reveal 3D trajectory at
1km accuracy
huib 23/6/06
NGC Groningen 29 June 2006
13/12
Goal of the project
• Develop: methods for high data rate e-VLBI using
distributed correlation
• High data rate eVLBI data acquisition and transport
• Develop a scalable prototype for broadband data acquisition
•Prototype acquisition system
• Establish a transportation protocol for broadband e-VLBI
•Build into prototype, establish interface normal system
• Interface e-VLBI public networks with LOFAR and e-MERLIN
dedicated networks
•Correlate wide band Onsala data on eMERLIN
•Demonstrate LOFAR connectivity
• Distributed correlation
• Setup data distribution over Grid
•Workflow management tool
• Develop a software correlator
huib 23/6/06
•Run a modest distributed eVLBI experiment
NGC Groningen 29 June 2006
14/12
2 major components
Part 1: Scalable connectivity
• 1.1. Data Acquisition
• 1.1.1. Data acquisition architecture (MRO)
• Scalable data acquisition system, off-the-shelf components
new version of PC-EVN?
• 1.1.2. Data acquisition prototype (MRO)
• Prototype for 4Gb/s?
• 1.1.3. Data acquisition control (MPI)
• Control data acquisition, interface for protocol, distributed
computing
• 1.2. Broadband Datapath
• 1.2.1. Broadband protocols (JBO)
• IP protocols, lambda switching, multicasting
• 1.2.2. Broadband data processor interface (JBO)
• Data from public network to eMERLIN correlator
huib 23/6/06
• 1.2.3. Integrate and test (OSO)
• 10 Gb/s test environment for OSO-eMERLIN (and LOFAR?)
• 1.2.4. Public to dedicated interface (ASTRON)
NGC Groningen 29 June 2006
15/12
Components (part 2)
Part 2: Distributed correlation
• 2.1. Grid resource allocation
• 2.1.1. Grid VLBI collaboration (PNSC)
• Establish relevant tools for eVLBI
• 2.1.2. Grid workflow management (PNSC)
• Tool to allocate correlator resources and schedule correlation
• 2.1.3. Grid routing (PNSC)
• Data flow from telescopes to appropriate correlator resources
• 2.2. Software correlation
• 2.2.1. correlator algorithm design
• High precision correlation on standard computing
• 2.2.2. Correlator computational core
• 2.2.3. Scaled up version for clusters
• 2.2.4. Distributed version, middleware
• Deploy on Grid computing
• 2.2.5. Interactive visualization
• 2.2.6. Output definition
huib 23/6/06
• Output data from individual correlators
• 2.2.7. Output merge
• Collect data in EVN archive
NGC Groningen 29 June 2006
16/12
On distributed computing
• What are the Grid resources
• Calibrate the require amount of computing
• Dynamical allocation possible?
• Interaction with observing schedule
• Topology of network
• Slice data in frequency, time or differently?
• Interface for routing data
• Multicast implementation on acquisition module
huib 23/6/06
NGC Groningen 29 June 2006
17/12
Distributed correlation
• Correlator model centrally generated?
• Or calculate at every node
• Plan for merging data back together
• How to get uvw coordinates in data
• Monitor progress centrally
huib 23/6/06
NGC Groningen 29 June 2006
18/12
Current eVLBI practice
observing schedule
in VEX format
BBC &
samplers
user correlator
parameters
field system
controls antenna
and acquisition
correlator control
including model
calculation
Mk4
formatter
Mk5
recorder
huib 23/6/06
NGC Groningen 29 June 2006
earth orientation
parameters
output
data
Mk4 data
in Mk5prop form
over TCPIP
Mk5
playback
19/12
FABRIC components
observing schedule
in VEX format
DBBC
VSI
field system
controls antenna
and acquisition
PC-EVN
#2
huib 23/6/06
NGC Groningen 29 June 2006
VSIe??
on??
GRID
resources data
user correlator
parameters
earth orientation
parameters
resource allocation
and routing
correlator control
including model
calculation
FABRIC
=
The GRID
output
data
20/12