Kim McCall. The Smalltalk-80 Benchmarks. In Glenn Krasner, editor, Smalltalk-80: Bits of History, Words of Advice, pages 251--270. Addison-Wesley, 1981.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....of software. 2) returns the image on CCD instantly. In order to make the digital camera controllable from Squeak, several primitives are added. The primitives added are shown on table 1. 5 Evaluation Table 2 shows the result of the Smalltalk 80 benchmarks presented in the Green Book[3] 1 . Note that a couple of parameters were modified in order to get a meaningful results on the modern implementation: firstly, the number of the repetition of the benchmarks were modified so that each test is kept running at least one second long. Secondly, test16bitArith was replaced with ....

Kim McCall. The Smalltalk-80 Benchmarks. In Glenn Krasner, editor, Smalltalk-80: Bits of History, Words of Advice, pages 251--270. Addison-Wesley, 1981.

....benchmarks to evaluate garbage collection performance. The first is a synthetic benchmark of our own devising based on tree creation. The second consists of several iterations through the standard macro benchmark suite that is used to compare the relative performance of Smalltalk implementations [7]. Our benchmarks have the following characteristics: ffl Destroy trees with destructive updates: A large initial tree ( 2M bytes) is repeatedly mutated by randomly choosing a subtree to be replaced and fully recreated. The effect is to generate large amounts of garbage, since the subtree that ....

K. McCall. The Smalltalk-80 benchmarks. In G. Krasner, editor, Smalltalk-80: Bits of History, Words of Advice, chapter 9, pages 153--173. Addison-Wesley, 1983.

....benchmarks to evaluate garbage collection performance. The first is a synthetic benchmark of our own devising based on tree creation. The second consists of several iterations through the standard macro benchmark suite that is used to compare the relative performance of Smalltalk implementations [16]. Our benchmarks have the following characteristics: ffl Destroy trees with destructive updates: A large initial tree ( 2M bytes) is repeatedly mutated by randomly choosing a subtree to be replaced and fully recreated. The effect is to generate large amounts of garbage, since the subtree that ....

K. McCall. The Smalltalk-80 benchmarks. In Krasner

....a run is about 24 megabytes. This benchmark explores the cost of applications that generate garbage rapidly. Interactive the macro benchmarks: For this benchmark we iterate 10 times through the full set of macro benchmarks. These benchmarks are part of the standard suite of benchmarks [6] used to compare the relative performance of different Smalltalk implementations. They measure system support for the programming activities that constitute typical interaction with the Smalltalk system, such as keyboard activity, compilation of methods to bytecodes, and browsing. 5 Experiments ....

K. McCall. The Smalltalk-80 benchmarks. In G. Krasner, editor, Smalltalk-80: Bits of History, Words of Advice, chapter 9, pages 153--173. Addison-Wesley, 1983.

....includes the DeutschSchiffman techniques described earlier. We compare transliterations from C into Smalltalk and SELF of the Stanford integer benchmarks [13] and the Richards operating system simulation benchmark [10] as well as the following small benchmarks, adapted from Smalltalk 80 systems [18]: sumToTest = 1 sumTo: 10000 ) sumTo: arg = total 0 to: arg Do: index total: total index. total ) recurseTest = 14 recurse ) recurse = 0 ifFalse: 1) recurse. 1) recurse. We also rewrote most of the Stanford integer benchmarks in a more SELFish ....

McCall, K. The Smalltalk-80 Benchmarks. In Krasner, G., editor, Smalltalk80: Bits of History, Words of Advice, Addison-Wesley, Reading, MA (1983) 153-174.

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