home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!auvm!CCM.HF.INTEL.COM!THOMAS_M_SKINNER
- Message-ID: <930122093353_8@ccm.hf.intel.com>
- Newsgroups: bit.listserv.sas-l
- Date: Fri, 22 Jan 1993 09:33:53 PST
- Reply-To: Thomas M Skinner <Thomas_M_Skinner@CCM.HF.INTEL.COM>
- Sender: "SAS(r) Discussion" <SAS-L@UGA.BITNET>
- From: Thomas M Skinner <Thomas_M_Skinner@CCM.HF.INTEL.COM>
- Subject: Re: updated benchmark scores...iCOMP index
- Comments: To: SAS-L@uga.cc.uga.edu
- Lines: 60
-
- Folsom, California
- January 22, 1992
- Clearing Skies,50's
- Dear SAS-Lers:
-
- In regard to the recent postings of performance on various hardware/OS
- configurations:
-
- It might be useful to ellucidate on which release of SAS the
- benchmarks were run on. When refering to specific X86 systems, it
- might be useful to specify the manufacturer, and in the case of intel
- cpu's, the exact designation (SX, SL, DX, DX2, OverDrive, Coprocessor)
- as performance will vary... which opens up a door of opportunity for
- me to present the following:
-
- For purely marketing purposes, and as a tool for non-technical
- end-user consumers to judge system configurations, Intel has
- formulated am index of performance for their chips. The "iCOMP"
- index, was formulated using the chips' frequency (Mhz), the number of
- instructions per second processed, and a series of existing technical
- benchmarks. The result is a single number that can be used to compare
- one processor against another.
-
- The 25MHz Intel486 SX was arbitrarily used as the baseline for the
- index and set at 100. This index includes a 32 bit performance
- component, to better reflect the demands placed on it over the next
- 3-5 years in a typical lifespan of a system purchased today.
-
- This is not a "system" performance factor, but strictly rates the
- relative performance of Intel CPU's. Applied to the previously posted
- benchmark results, the influence of disk drive access speed, bus
- configurations, OS, etc. become apparent. This index is not intended
- for comparisons of non-intel cpus.
-
- i386 SX-16 ** 22
- i386 SX-20 *** 32
- i386 SX-25 **** 39
- i386 SL-25 **** 41
- i386 DX-25 ***** 49
- i386 DX-33 ******* 68
- i486 SX-20 ******** 78
- i486 SX-25 *Baseline* 100
- i486 DX-25 ************ 122
- i486 SX-33 ************** 136
- i486 DX-33 ***************** 166
- i486 DX2-50 *********************** 231
- i486 DX-50 ************************* 249
- i486 DX2-66 ****************************** 297
-
-
- Thanks to list for their indulgence in this message, any inquiries
- about the index may be directly addressed to myself.
-
- Intel-Insider,
- Tom Skinner
- ======================================================================
- T.M. Skinner Internet: Thomas_M_Skinner@ccm.hf.intel.com
- Sr. SAS Tech. Support Mail: 1900 Prairie City Rd. MS FM1-102
- Intel Corporation Folsom, CA 95630
- ======================================================================
-