Profile inference revisited

Profile-guided optimization (PGO) is an important component in modern compilers. By allowing the compiler to leverage the program’s dynamic behavior, it can often generate substantially faster binaries. Sampling-based profiling is the state-of-the-art technique for collecting execution profiles in data-center environments. However, the lowered profile accuracy caused by sampling fully optimized binary often hurts the benefits of PGO; thus, an important problem is to overcome the inaccuracy in a profile after it is collected. In this paper we tackle the problem, which is also known as profile inference and profile rectification . We investigate the classical approach for profile inference, based on computing minimum-cost maximum flows in a control-flow graph, and develop an extended model capturing the desired properties of real-world profiles. Next we provide a solid theoretical foundation of the corresponding optimization problem by studying its algorithmic aspects. We then describe a new efficient algorithm for the problem along with its implementation in an open-source compiler. An extensive evaluation of the algorithm and existing profile inference techniques on a variety of applications, including Facebook production workloads and SPEC CPU benchmarks, indicates that the new method outperforms its competitors by significantly improving the accuracy of profile data and the performance of generated binaries.

ENTERPRISE COST MANAGEMENT METHODS

The article is devoted to the research analysis of current patterns in the methodology of cost management at the enterprise. Modern methods of production cost management are analyzed. Methods that take into account the influence of the internal and external environment on production costs are determined. It is established that each method used autonomously does not solve the entire complex of management tasks. Therefore, the improvement of cost management methods should be based on the synthesis of a new method that includes the advantages of existing cost methods and eliminates their disadvantages

Applying M-sequences Decimation to Generate Interleaved Sequence

Abstract—M-sequences are widely used in for many purposes, from synchronization, whitening, communications and cryptography. We analyze decimation techniques and introduce two methods to generate decimation sequences which don’t have to calculate intermediate states. Then we apply these methods to interleaved sequence as a new method to pre-calculate for set of interleaved order which is more effective in implementation. Tóm tắt—M-dãy đang được sử dụng rất rộng rãi trong nhiều lĩnh vực, từ việc đồng bộ, làm trắng thông tin, viễn thông và kỹ thuật mật mã. Chúng tôi phân tích kỹ thuật phân rã m-dãy theo bước và giới thiệu hai phương pháp sinh dãy phân rã theo bước mà không cần tính các trạng thái trung gian. Áp dụng phương pháp này vào dãy lồng ghép, ta có một phương pháp mới để tính trước tập các thứ tự lồng ghép có tính hiệu quả trong cài đặt thực tế.

Gyro-average method for global gyrokinetic particle simulation in realistic tokamak geometry

Gyro-average is a crucial operation to capture the essential finite Larmor radius effect (FLR) in gyrokinetic simulation. In order to simulate strongly shaped plasmas, an innovative multi-point average method based on non-orthogonal coordinates has been developed to improve the accuracy of the original multi-point average method in gyrokinetic particle simulation. This new gyro-average method has been implemented in the gyrokinetic toroidal code (GTC). Benchmarks have been carried out to prove the accuracy of this new method. In the limit of concircular tokamak, ion temperature gradient (ITG) instability is accurately recovered for this new method and consistency is achieved. The new gyro-average method is also used to solve the gyrokinetic Poisson equation, and its correctness has been confirmed in the long wavelength limit for realistic shaped plasmas. The improved GTC code with the new gyro-average method has been used to investigate the ITG instability with EAST magnetic geometry. The simulation results show that the correction induced by this new method in the linear growth rate is more significant for short wavelength modes where the finite Larmor radius (FLR) effect becomes important. Due to its simplicity and accuracy, this new gyro-average method can find broader applications in simulating the shaped plasmas in realistic tokamaks.

Object-Based Geomorphological Mapping: Application on an Alpine Deep-Seated Gravitational Slope Deformation Contest (Germanasca Valley, Western Alps—Italy)

This research reports the use of a new method of geomorphological mapping in GIS environments, using a full-coverage, object-based method, following the guidelines of the new geomorphological legend proposed by ISPRA–AIGEO–CNG. This methodology is applied to a tributary valley of the Germanasca Valley, shaped into calcschist and greenschist, of the Piedmont Zone (Penninic Domain, Western Alps). The investigated sector is extensively affected by dep-seated gravitational slope deformation (DSGSD) that strongly influences the geological setting and the geomorphological features of the area. The mapping of these gravitational landforms in a traditional way creates some difficulties, essentially connected to the high density of information in the same site and the impossibility of specifying the relationships between different elements. The use of the full-coverage, object-based method instead is advantageous in mapping gravitational evidence. In detail, it allows for the representation of various landforms in the same sector, and their relationships, specifying the size of landforms, and with the possibility of multiscale representation in the GIS environment; and, it can progressively be update with the development of knowledge. This research confirms that the use of the full-coverage, object-based method allows for better mapping of the geomorphological features of DSGSD evidence compared to classical representation.