Mamta Shukla

Compiler-rt is a set of runtime routines present in LLVM infrastructure to support and provide an implementation of various builtin functions for multiple architectures. It consists of – builtin libraries, sanitizer runtimes, profile runtimes, and block runtimes. This talk focuses on one of the compiler rt components- sanitizers. Compiler-rt sanitizers are used to perform runtime checks on compiled code. These checks can include detecting and preventing common programming errors, such as buffer overflows and memory leaks, as well as more sophisticated checks, such as undefined behavior sanitization, which can help detect and prevent subtle bugs that can be difficult to find using other methods. This talk will provide an introduction to compiler rt sanitizers in LLVM, and how to use them on various architectures starting with an example of memory sanitizer for x86 architecture. Second part of the talk will cover – how compiler-rt sanitizers can be used for 32-bit and 64-bit ARM architecture and integrate in development and testing using Software development kit (SDK) built with Yocto. Also useful for the development of applications for memory-constrained embedded devices using clang and LLVM.

At European Organization for Nuclear Research(CERN), for the alignment of large superconducting magnets and cryogenics, an interferometry based system is being devised to identify the position of their elements. This technique uses interferometry principle and uses sweeping laser to identify the distance of multiple points using Fourier Analysis. The data acquired from photo-detection module, received after a sweep of laser source, needs to undergo sophisticated post processing to obtain the final results. The system must monitor position of a large number of elements every second. Dealing with 1000s of target points in less than 1 second required time-optimized and precise calculation. Thus, GPU was employed to provide faster and precise results. This required to use signal processing algorithms like: Butterworth Filter, Hilbert Transform, Savitzky-Golay smoothing Filter in GPU. This talk will cover steps involved in adopting signal processing algorithm to GPU to achieve better performance and understand the effects of parallelism achieved. For the initial development, CuPy library for NVIDIA GPUs is used and later moved to implementation in C. CuPy provides wrapper for most of the CUDA toolkit in Python. We will also provide highlight about performance metrics with respect to increase in the data size and possible optimizations of its processing.