H. Massalin. An Efficient Implementation of Fundamental Operating System Services. Ph. D. dissertation, Columbia University, 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....inefficient. There is a folklore argument saying that remapping costs are dominated by basic hardware and OS overhead. This argument is wrong when dealing with new architectures: a system call instruction can be made as efficient as any indirect call and OS overhead can be kept low, as Synthesis [Massalin 1992] and L3 [Liedtke 1993b] show. ffl MPT 64 and GPT enable hierarchies and fast operations on regions. ffl IPT does not. 7 4K 128K 4M 128M 4G 128G 4T 128T 4KT128KT4MT region size 10 10 2 10 3 10 4 10 5 10 6 o p e a t i o n s . ....

Massalin, H. 1992. Efficient implementation of fundamental operating system services. Ph. D. thesis, Columbia University, New York.

....systems, notably Lisp, and usually involved prohibitive interpretive overhead or compilation time. Since then, it has been applied to code optimization [33] efficient dynamic implementation of programming languages [26] optimization of bitmap graphics [15,47] operating system optimization [42] and other tasks specifically suited to RTCG [9,10,39,40] These present only selections; a complete overview of the field would exceed the scope of this paper. Of particular interest is the work relating RTCG to partial evaluation. The Fabius system of Lee and Leone [39,40] performs RTCG of ....

....to a variety of different situations. There is no clear cut border between ordinary compilation and RTCG. On the ordinary end of the spectrum, compilation is offline as in traditional compilers. On the other end of the spectrum, there are online systems like DCG [18] or the Synthesis kernel [42] that compile code for immediate execution. The rest of the spectrum is not empty: programming languages like ML, Scheme, or Smalltalk have a read eval print loop that accepts function definitions that are compiled and the code is immediately available for execution. Hence, they are essentially ....

Henry Massalin. Efficient Implementation of Fundamental Operating System Services. PhD thesis, Columbia University, 1992.

....but increases latency and memory traffic. design time Optimization requires time and planning that are unavailable to exploratory and evolutionary programmers. Run time code generation (RTCG) as exemplified by Common LISP [51] Pike s Blit terminal [43] Masselin s Synthesis operating system [36], researchers at the University of Washington [32] and elsewhere [14] 35] is one way to attack this problem. With these systems, one writes programs that create programs. Generating rare cases and fused, one pass loops as needed directly addresses the program size and latency traceoffs outlined ....

Henry Massalin. Efficient Implementation of Fundamental Operating System Services. Columbia 1992.

....number of instructions executed during each traversal step is small. This technique is especially useful when the work done on each data element is also small compared to the traversal cost. The Synthesis run queue, composed of thread table elements, is an example of an executable data structure [20]. At thread creation time, each element is optimized to reduce context switch cost. The pointer to the next thread, for example, serves as the destination of a jump instruction to eliminate address load overhead. 3.2 Performance Benefits Many of the performance measurements on Synthesis were ....

....programs, measured in seconds. Each program repeatedly executes the specified system calls (the left column) The validation program contains only user level memory location references. Figure 1: Comparison of Synthesis and SUNOS I O Performance time (almost 4 times) A more recent experiment [20] illustrates the relative I O latency for Synthesis and two widely used commercial operating systems. The Synthesis window system on the Sony 1860 NEWS workstation can finish cat etc termcap in 2.9 seconds, while X Windows (BSD Unix) takes 23 seconds and NextStep (a derivative of Mach) with ....

[Article contains additional citation context not shown here]

H. Massalin. Efficient Implementation of Fundamental Operating System Services. PhD thesis, Department of Computer Science, Columbia University, April 1992.

....in assembler, which allows unrestricted access to the machine s architecture, and perhaps discover new coding idioms that have not yet been captured in a higher level language. Surprisingly, we found that there are some things that were actually easier to do using Synthesis assembler than using C [1]. One example is the finite state machines found in various I O drivers and protocol engines. Even the conceptually simple function of terminal emulation gives rise to difficult problems when coded in C. For example, how do we handle the case when the input buffer runs out while in the middle of ....

....on Synthesis. The most time consuming part was implementing the new services: virtual memory, ethernet driver, and window system. They were all implemented from scratch , using all the performance improving ideas used in the rest of Synthesis, such as kernel code generation. Recent measurements [1] show similar performance gains in these areas as well. The ethernet driver, for example, is fast enough to record all the packet traffic of a busy ethernet (400 kilobytes second) into RAM using only 20 of a 68030 CPU time. This is a problem that has been worked on and dismissed as impractical ....

[Article contains additional citation context not shown here]

H. Massalin. Efficient Implementation of Fundamental Operating System Services. PhD thesis, Columbia University, Forthcoming 1992.

No context found.

H. Massalin. An Efficient Implementation of Fundamental Operating System Services. Ph. D. dissertation, Columbia University, 1992.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC