iCOMP

iCOMP, short for Intel Comparative Microprocessor Performance, was an index created and published by Intel to communicate the relative performance of its microprocessors. It was designed during an era when personal-computer buyers and technology marketing commonly treated processor clock speed, expressed in megahertz, as a simple proxy for overall computing performance. Intel’s research indicated that many buyers assumed a higher MHz rating automatically meant a faster processor, even when comparing different processor architectures or processor families. The iCOMP concept addressed that problem by combining results from several standardized or widely recognized benchmarks into a composite performance score. Rather than measuring only clock frequency, the index attempted to represent several aspects of processor behavior. Its calculation used an integer-performance component derived from Ziff-Davis Labs’ ZDbenchCPU results, themselves related to earlier PC Labs benchmarks. Floating-point performance was represented through Whetstone as implemented in PowerMeter, while 32-bit integer and floating-point components were drawn from SPECint92 and SPECfp92. The resulting score was intended to provide a more meaningful basis for comparing Intel processors with different designs and operating frequencies. iCOMP was revised several times as Intel introduced newer processor generations and as benchmark conditions changed. Version 1.0, introduced in 1992, used the Intel 486SX running at 25 MHz as its reference point. Version 2.0, introduced in 1996, used the Pentium 120 as its baseline. Version 3.0, introduced in 1999, used a Pentium II running at 350 MHz as its reference. These reference changes reflected the rapid development of Intel’s microprocessor range and the need to recalibrate the index for newer performance levels. The index was both a technical measurement framework and a product-communication device. By presenting a composite number, Intel could explain performance differences that were not obvious from clock speed alone and could position newer processors in relation to earlier models. At the same time, because the index was associated with Intel and relied partly on benchmark selections and weighting, it was not a universal or vendor-neutral measure of every aspect of computing performance. Its historical importance lies in showing how processor manufacturers attempted to make benchmark-based comparisons accessible to consumers during the transition from the 486 era to Pentium and Pentium II generations. The available reference material documents the index’s purpose, benchmark composition, and three published revisions, but does not establish a separate consumer-facing business, standalone corporate structure, current website, or continuing publication schedule. iCOMP is therefore best treated as a historical Intel performance index and trademarked or branded measurement system rather than as an independent operating company.

Intel developed iCOMP in response to a communication problem in the personal-computer market. Buyers often interpreted processor clock speed as a direct measure of performance, even though processors with different architectures could deliver different results at the same frequency. Intel’s research suggested that the MHz figure had become overly influential in purchasing decisions. The company therefore promoted a composite index intended to reflect a broader range of processor capabilities. The first documented release, iCOMP 1.0, appeared in 1992 and used the 25 MHz Intel 486SX as its reference processor. The score was calculated from several benchmark categories rather than from frequency alone. The largest component came from an integer CPU benchmark associated with Ziff-Davis Labs’ ZDbenchCPU, which was derived from earlier PC Labs testing. Whetstone, as implemented in PowerMeter, supplied a 16-bit floating-point component. The index also incorporated SPECint92 and SPECfp92 to represent 32-bit integer and floating-point performance. As Intel’s processor portfolio developed, the original reference point became outdated. Intel revised the index to iCOMP 2.0 in 1996, benchmarking the system against a Pentium 120. This revision maintained the basic idea of combining different benchmark dimensions while updating the comparison scale for the Pentium era. In 1999, Intel introduced iCOMP 3.0 with a Pentium II running at 350 MHz as its reference. The successive versions show that iCOMP was not a fixed, timeless rating: its baseline and relevance were tied to the processor generation and benchmark environment of the period. Historically, iCOMP occupied a middle ground between technical benchmarking and consumer marketing. Its use of multiple test categories offered a more nuanced comparison than clock frequency, while the single composite score made the results easier to communicate in product literature and competitive comparisons. The approach also had limitations. Any composite index depends on the selected benchmarks, their implementation, and the weighting or calculation method used. Consequently, an iCOMP score was useful for the comparisons Intel intended to illustrate, but it should not be read as a complete measure of every workload or as a universal ranking of all processors. The available material identifies iCOMP as an Intel-published index and records three revisions through 1999. It does not provide evidence of an independent company, a separate management team, a current product line, or a later version. The present operating status of the index is therefore not established in the supplied references. Its documented legacy is as a historical Intel framework for explaining comparative microprocessor performance during the 486, Pentium, and Pentium II periods.

• 1999

  Introduction of iCOMP 3.0

  [1]iCOMP (index)en.wikipedia.org

  The third documented revision used a Pentium II running at 350 MHz as the benchmark reference.

• 1996

  Introduction of iCOMP 2.0

  [1]iCOMP (index)en.wikipedia.org

  Intel revised the index and changed its reference point to the Pentium 120, aligning the rating with the company’s newer processor generation.

• 1992

  Introduction of iCOMP 1.0

  [1]iCOMP (index)en.wikipedia.org

  Intel introduced the first documented iCOMP version, using the 25 MHz 486SX as its reference processor and combining several benchmark categories into a composite score.

A composite, benchmark-based alternative to comparing processors by MHz alone.

• Processor performance index · 1992

  iCOMP 1.0

  The original documented iCOMP release, introduced in 1992. It used the Intel 486SX at 25 MHz as its reference point and combined integer, floating-point, and other benchmark results to create a comparative processor-performance score.

• Processor performance index · 1996

  iCOMP 2.0

  A 1996 revision of the index benchmarked against the Pentium 120. It updated the reference scale for Intel’s Pentium-era processors while retaining the broader goal of measuring performance through multiple benchmark dimensions.

• Processor performance index · 1999

  iCOMP 3.0

  The 1999 revision used a Pentium II running at 350 MHz as its reference. It represented the index’s documented evolution into the Pentium II generation and continued the use of composite benchmark-based comparison.

1. 1iCOMP (index)en.wikipedia.org