Talks

The latest enhancement to the GMRT’s transient science capability is the SPOTLIGHT project, designed to detect Fast Radio Bursts (FRBs) in real-time. This is achieved by forming 2000 post-correlation (PC) beams across the sky while piggybacking on regular GMRT observations over a frequency range from 300 MHz to 1460 MHz. In addition to real-time FRB detection, SPOTLIGHT includes a quasi-real-time pulsar search using 10% of the full field-of-view. The acquired data will undergo weekly processing using a Fourier Domain Acceleration Search pipeline, aiming to discover new pulsars from SPOTLIGHT observations. In this presentation, I will highlight the development and optimisation of the multi-GPU pulsar search pipeline designed for SPOTLIGHT. Additionally, I will discuss the key scientific goals of the pulsar search component, including the expected pulsar yield and the broader astrophysical implications of this project. Enabling commensal pulsar searches on GMRT, serving as both a science initiative and a technology demonstrator.

Fast Radio Bursts (FRBs) represent one of modern astronomy's most compelling mysteries. Despite cataloging approximately 800 FRBs, including 65 repeaters, their exact emission mechanisms remain unknown. This knowledge gap persists largely due to limitations in localization accuracy and timing. Current FRB detection systems face a critical tradeoff: wide-field instruments discover numerous FRBs but provide poor localization accuracy, while interferometric arrays offer precise positions but process data with substantial delays—typically a week between detection and localization. This delay severely limits multi-wavelength follow-up observations, particularly for transient repeaters that may remain active only briefly after discovery. SPOTLIGHT (Survey for sPoradic radiO bursTs via a commensaL, multIbeam, GPU-powered HPC at the GMRT) addresses this challenge through an innovative real-time imaging localization pipeline. Operating commensally with GMRT, the system searches for FRBs across 2000 phased array beams using combined HPC and AI techniques. Upon detection, SPOTLIGHT triggers high-time resolution (1.3ms) visibility recording across a 16-node cluster. These visibilities undergo immediate processing: de-dispersion at the candidate dispersion measure, conversion to CASA measurement sets, and imaging for precise localization. The entire process completes within approximately ~2 minutes post detection—dramatically faster than existing systems. The implementation leverages 8 servers with 384 CPUs dedicated to real-time imaging, capable of handling approximately 100 triggers daily. This capability enables immediate multi-wavelength follow-up observations, critical for capturing rapid afterglows and understanding repeater behavior. SPOTLIGHT represents the first operational survey providing real-time FRB localization—a transformative advancement for radio astronomy that maximizes scientific opportunities from each detection.

The SPOTLIGHT system is a real-time system developed for the upgraded GMRT, designed to capture and localize fast radio transients with high time and frequency resolution. It comprises a GPU-based search engine that performs real-time dedispersion and single-pulse
detection across multiple beams, followed by a triggering system that initiates voltage capture and high-resolution imaging for candidate localization. A critical component of this pipeline is the “candidate optimiser,” which is responsible for sifting through thousands of candidates generated every observing session. The optimiser employs a combination of clustering,
coincidence and anticoincidence filtering and machine learning-based classification to identify true astrophysical signals while discarding RFI and other false positives. This enables efficient use of computational resources and ensures timely triggering for follow-up observations. Here we describe the design, implementation, and performance of the candidate optimiser, and
highlights its key role in enhancing the reliability of FRB searches with SPOTLIGHT.

Fast Radio Bursts (FRBs) are a source of a very short-duration transient signal of an order of milliseconds, the highly energetic cosmological event, origin of which is not fully understood yet and that has become the center of attention in astronomy recently. SPOTLIGHT is a real-time multi-beam commensal system that will run in parallel with the regular GMRT observations and will search FRBs and pulsars in the field of view (FoV). For real-time search of FRBs, a state-of-the-art parallel multi-node computer cluster is utilized to form visibilities, post-correlation beam-formed high-time resolution time data, and machine learning algorithms to classify the event of interest from the unwanted noise (e.g., radio frequency interference).

The survey speed for the FRBs can be increased by meeting two requirements viz. 1. Increasing FoV : One way to acheive this is by forming large number of beams. 2) Precise localisation: While many FRBs are detected in the time domain, a very small fraction of FRBs have been localized, and hence, new techniques for localization are in high demand across the world. SPOTLIGHT is targeted to search FRBs in the observation field by splitting it into about 2000 post-correlated beams steered across FoV. Digital technology allows the formation of thousands of such synthetic beams tiled in a given region in the sky. Array configuration, i.e. antennas included, frequency of observation and the source directions (declination of source, hour angle of observation) with respect to the array, determines the post correlated beam size and its orientation. With the earth's rotation, the configuration of the array changes systematically over observation time and for different sources in the sky (e.g., calibrators, target field) for a given observing session. So, dynamic and efficient computing of post correlated beams and tiling thousands of such beams becomes crucial in source localisation.

Multi-beam simulation package is developed which models the post correlated beam for a given observational setup and optimally arranges the beams in a given region of sky. It supports three modes : 1. Sky survey mode : Covering maximum FoV (with moderate length of baseline ) 2. Targeted mode - Covers given FoV (for specific search e.g. globular cluster observation, size few arc-min). 3. Multi-source mode : Observes different sources in given FoV.

In Sky Survey mode beams are tiled such that it covers the maximum FoV within primary beamwidth. Selects the optimum overlap ratio of the beam to give maximum coverage for the given observing session. However in Targeted mode, the number of beams and the overlap ratio is calculated based on the given FoV. Multi-Source mode is extended version of targeted mode, where multiple sources within given FoV can be observed.

We carried out functionality checks for modeling technique for post correlation beam shape and verification of tilling arrangement with variety of antenna configuration, source directions, and observing frequencies. We tested simulation of post correlation beam pattern by observing a field with 800 steered beams with source away from the phase center and compared the folded SNR map (fig. 1) for these beams with simulated beam patterns (fig. 2). The results showed that the source observed was located well within the accuracy of 1HPBW.

The simulation package simulates the post correlation beam shape and optimally tiles the beams such that it effectively covers a larger fraction of the field of view while keeping the sensitivity of detection fairly high so as not to miss any faint FRB event occurring across the field of view.

SPOTLIGHT is a time-domain survey instrument to perform a real-time commensal search for FRBs (Fast Radio Bursts) and Pulsars with a PetaFlop system hosting 60 CDAC’s indigenously developed Rudra servers equipped with 90 A100 GPUs connected to a 2 PB storage over a 100 Gbps Infiniband network. The system executes real-time HPC and AI applications to ensure simultaneous time-domain detection and arc-second imaging localisation of the detected bursts across the GMRT observing band of 300 to 1460 MHz. The 60 Rudra servers are arranged in three sub-clusters implementing the correlator, transient search and imaging analysis. The first cluster, consisting of 16 servers, performs correlation and beamformation forming 2000 beams at 1.3 millisecond time-resolution and sends the 2000 beams to the 24 servers of the second cluster for multi-beam transient detection pipeline to search for FRBs using Machine Learning for classification. The 2000 beams are optimally located well within the half-power beamwidth of the primary beam of a GMRT antenna. Imaging localisation on the detected FRBs and candidate confirmation is done on the 20 servers of the third cluster.

The first stage of the system is equipped to accept Nyquist sampled digitised data of baseband signals of both polarisations of all antennas of GMRT at the rate of 25 GB/s over 32 ten gigabit fibre links to perform correlation and beamformation. The data ingestion to the SPOTLIGHT system is done via UDP packet broadcasting from the GMRT samplers, ensuring piggy-back operation with the GMRT Wide-band Backend. The received digitised data on each server over two ten gigabit fibre links is buffered, time-sliced and shared between the servers on a 100 Gbps network such that each server gets a slice of contiguous time series digitised data of all antennas, on which it performs operations like FFT, Multiplication and Accumulation (MAC) and beamformation using dual A100 GPUs. MAC is designed to give visibilities at 1.3 ms time resolution. The post-correlation beams are formed using the phased addition of the 1.3 ms time resolution visibilities. The post-correlation beams reduce the effect of radio frequency interferences and systematics in the data by eliminating the auto-correlation products (Figure 1). Cleaning the dynamic spectrum and more importantly, the dispersion measure-time image significantly increases the efficiency of the AI classifier. The beam-steering implemented at the post-correlation stage significantly minimizes computational costs. The 2000 beams thus formed are distributed over the servers of the second stage such that each server receives data of ~84 beams. The beam data received is buffered in a shared memory which is accessed by the multi-beam transient detection pipeline for pulsar and FRB searches.

The 1.3 ms time resolution visibilities formed in the first stage are integrated up to 1 second or higher and sent to a dedicated host node for archiving. This visibility data at this reduced time-resolution is used to detect the non-working antennas to remove them from the array and also to perform real-time phase and amplitude calibration of the data. In the first stage, in parallel to performing correlation and beamforming, the digitised Nyquist data is written to a shared memory. Upon detection of a transient event in the second stage, this baseband shared memory is accessed to write the Nyquist sampled voltages to the 2 PB storage over 100 Gbps network. The availability of the full Nyquist samples from the full array covering the dispersion sweep of the FRBs opens up the possibility of probing the underlying emission physics aided with full polarisation information and the propagation effects at an unprecedented sensitivity. In addition to this baseband buffering, in the second stage, the 1.3 ms time resolution visibilities are buffered in another shared memory which is accessed to write these visibilities to the 2 PB storage upon detection of a burst. This facilitates real-time imaging localisation of the FRB at arc-second resolution, enabling the host galaxy association to use the burst as a cosmological tool. The sizes of shared memory for the beam data (~ 2.4 TB) and 1.3 millisecond resolution visibilities (~ 5 TB) in the second stage and the baseband shared memory (~ 4.8 TB at 4-bits per sample) in the first stage are large enough to hold the data on Random Access Memory to accommodate detection of bursts even at high dispersion measure.

The SPOTLIGHT system is being realized in a phased manner. The entire hardware installation was completed by October 2024. Following this, the software pipelines are being tested with an increasing number of beams progressively. The real-time system currently being tested at the GMRT can generate 800 beams, along with all promised data products, using 32 A100 GPUs, achieving 40% of the final specification.

We consider the problem of simulating dynamics of classical nonlinear dissipative systems with N>>1 degrees of freedom. To make the problem tractable for quantum computers, we add a weak Gaussian noise to the equation of motion and the initial state. Our main result is an end-to-end quantum algorithm for simulating the noisy dynamics of nonlinear systems satisfying certain sparsity and divergence-free conditions. For any constant nonzero noise rate, the quantum runtime scales polynomially with log(N), evolution time, inverse error tolerance, and the relative strength of nonlinearity and dissipation.