اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدGlow: Graph Lowering Compiler Techniques for Neural Networks
67
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
This paper presents the design of Glow, a machine learning compiler for heterogeneous hardware. It is a pragmatic approach to compilation that enables the generation of highly optimized code for multiple targets. Glow lowers the traditional neural network dataflow graph into a two-phase strongly-typed intermediate representation. The high-level intermediate representation allows the optimizer to perform domain-specific optimizations. The lower-level instruction-based address-only intermediate representation allows the compiler to perform memory-related optimizations, such as instruction scheduling, static memory allocation and copy elimination. At the lowest level, the optimizer performs machine-specific code generation to take advantage of specialized hardware features. Glow features a lowering phase which enables the compiler to support a high number of input operators as well as a large number of hardware targets by eliminating the need to implement all operators on all targets. The lowering phase is designed to reduce the input space and allow new hardware backends to focus on a small number of linear algebra primitives.
قيم البحث
اقرأ أيضاً
As gradual typing becomes increasingly popular in languages like Python and TypeScript, there is a growing need to infer type annotations automatically. While type annotations help with tasks like code completion and static error catching, these anno
tations cannot be fully determined by compilers and are tedious to annotate by hand. This paper proposes a probabilistic type inference scheme for TypeScript based on a graph neural network. Our approach first uses lightweight source code analysis to generate a program abstraction called a type dependency graph, which links type variables with logical constraints as well as name and usage information. Given this program abstraction, we then use a graph neural network to propagate information between related type variables and eventually make type predictions. Our neural architecture can predict both standard types, like number or string, as well as user-defined types that have not been encountered during training. Our experimental results show that our approach outperforms prior work in this space by $14%$ (absolute) on library types, while having the ability to make type predictions that are out of scope for existing techniques.
We present Calyx, a new intermediate language (IL) for compiling high-level programs into hardware designs. Calyx combines a hardware-like structural language with a software-like control flow representation with loops and conditionals. This split re
presentation enables a new class of hardware-focused optimizations that require both structural and control flow information which are crucial for high-level programming models for hardware design. The Calyx compiler lowers control flow constructs using finite-state machines and generates synthesizable hardware descriptions. We have implemented Calyx in an optimizing compiler that translates high-level programs to hardware. We demonstrate Calyx using two DSL-to-RTL compilers, a systolic array generator and one for a recent imperative accelerator language, and compare them to equivalent designs generated using high-level synthesis (HLS). The systolic arrays are $4.6times$ faster and $1.1times$ larger on average than HLS implementations, and the HLS-like imperative language compiler is within a few factors of a highly optimized commercial HLS toolchain. We also describe three optimizations implemented in the Calyx compiler.
Just-in-time (JIT) compilers are used by many modern programming systems in order to improve performance. Bugs in JIT compilers provide exploitable security vulnerabilities and debugging them is difficult as they are large, complex, and dynamic. Curr
ent debugging and visualization tools deal with static code and are not suitable in this domain. We describe a new approach for simplifying the large and complex intermediate representation, generated by a JIT compiler and visualize it with a metro map metaphor to aid developers in debugging.
REC (REGULAR EXPRESSION COMPILER) is a programming language of simple structure developed originally for the PDP-8 computer of the Digital Equipment, Corporation, but readily adaptable to any other general purpose computer. It has been used extensive
ly in teaching Algebra and Numerical Analysis in the Escuela Superior de Fisica y Matematicas of the Instituto Politecnico Nacional. Moreover, the fact that the same control language, REC, is equally applicable and equally efficient over the whole range of computer facilities available to the students gives a very welcome coherence to the entire teaching program, including the course of Mathematical Logic which is devoted to the theoretical aspects of such matters. REC; derives its appeal from the fact that computers can be regarded reasonably well as Turing Machines. The REC notation is simply a manner of writing regular expression, somewhat more amenable to programming the Turing Machine which they control. If one does not wish to think so strictly in terms of Turing Machines, REC expressions still provide a means of defining the flow of control in a program which is quite convenient for many applications.
Inspired by recent successes with parallel optimization techniques for solving Boolean satisfiability, we investigate a set of strategies and heuristics that aim to leverage parallel computing to improve the scalability of neural network verification
. We introduce an algorithm based on partitioning the verification problem in an iterative manner and explore two partitioning strategies, that work by partitioning the input space or by case splitting on the phases of the neuron activations, respectively. We also introduce a highly parallelizable pre-processing algorithm that uses the neuron activation phases to simplify the neural network verification problems. An extensive experimental evaluation shows the benefit of these techniques on both existing benchmarks and new benchmarks from the aviation domain. A preliminary experiment with ultra-scaling our algorithm using a large distributed cloud-based platform also shows promising results.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
