4

I've documented the Ahl benchmark's history, and now I would like to do the same for Byte's version of the sieve. Does anyone have a pointer to the first appearance of this code as a benchmarking tool in that magazine?

Maury Markowitz
  • 19,803
  • 1
  • 47
  • 138
  • Ok, the article is up: https://en.wikipedia.org/wiki/Byte_Sieve Any suggestions on what language to use as the canonical example? JS? C? – Maury Markowitz May 02 '19 at 19:02
  • The canonical example of Byte Sieve should be in ISO Minimal Basic. – hotpaw2 May 02 '19 at 19:32
  • There is such a thing? Google fails to turn anything up. I recall the original ANSI/ISO effort but I seem to recall that was rendered moot and basically just faded away? – Maury Markowitz May 03 '19 at 16:23
  • Out-of-print: ISO 6373-1984 (E) or ECMA-55 Minimal BASIC. See: http://www.ecma-international.org/publications/files/ECMA-ST-WITHDRAWN/ECMA-55,%201st%20Edition,%20January%201978.pdf – hotpaw2 May 03 '19 at 17:06
  • Well I don't think it's ISO, but BASIC and C versions have been provided. – Maury Markowitz May 09 '19 at 16:59

2 Answers2

10

The earliest use, as far as I can determine, of the sieve of Eratosthenes as a benchmark in Byte’s editorial content is in the September 1981 issue, page 180: “A High-Level Language Benchmark”. The article introduces the algorithm and a number of implementations, and uses it to compare various interpreters and compilers on different platforms.

This approach was revisited in the January 1983 issue, page 283: “Eratosthenes Revisited — Once More through the Sieve”, and again in the August 1984 issue, page 132, when the Byte Unix benchmark was introduced.

Stephen Kitt
  • 121,835
  • 17
  • 505
  • 462
5

Not exactly what you are looking for, but meets the technical statement of "a pointer to the first appearance of this code as a benchmarking tool in that magazine" and is earlier than Stephen Kitt's answer. November 1980:

https://archive.org/details/byte-magazine-1980-11-rescan/page/n255?q=sieve

This is in an ad for Digital Research PL/I, and specifically refers to "Erastothenes Sieve" as a benchmark. Though I suppose that leaves open the question of how Digital Research's Sieve compares to the BYTE Magazine Sieve.

manassehkatz-Moving 2 Codidact
  • 14,358
  • 2
  • 37
  • 65
  • 3
    It all depends on whether “this code” refers to any implementation of the sieve, or Byte’s in particular — I assumed the latter. – Stephen Kitt May 01 '19 at 17:51
  • 1
    Ahh, I think this is an interesting point! The original article mentions that Gary Kildal personally supplied a version in PL/1 to be included in the article, and it also mentions that it was six months from start to end, so in fact I think this version IS the same one as September 1981, just way out in front due to publishing timelines! – Maury Markowitz May 02 '19 at 19:04