home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.intel
- Path: sparky!uunet!spool.mu.edu!agate!boulder!ucsu!gonzaled
- From: gonzaled@ucsu.Colorado.EDU (LGV/MC)
- Subject: Re: Intel's new performance test iCOMP
- Message-ID: <1992Dec19.031237.4870@ucsu.Colorado.EDU>
- Organization: University of Colorado, Boulder
- References: <92353.39992.J056600@LMSC5.IS.LMSC.LOCKHEED.COM>
- Date: Sat, 19 Dec 1992 03:12:37 GMT
- Lines: 51
-
- J056600@LMSC5.IS.LMSC.LOCKHEED.COM writes:
-
- >In <1992Dec18.112919.22234@fwi.uva.nl>, Stefan Sinnige writes:
-
-
- >>I read in our local newspaper about a new performance test, to compare
- >>Intel processors. The performance test is developed at Intel itself and is
- >>called the iCOMP (tm) Rating. The results are as follows:
-
- >> i486 DX2 66 MHz 297
- >> i486 DX 50 MHz 249
- >> i486 DX2 50 MHz 231
- >> i486 DX 33 MHz 166
- >> i486 SX 33 MHz 136
-
- i486 DX 25 MHz 122
-
- >> i486 SX 25 MHz 100
- >> i486 SX 20 MHz 78
- >> i386 DX 33 MHz 68
- >> i386 DX 25 MHz 49
- >> i386 SL 25 MHz 41
- >> i386 SX 25 MHZ 39
-
- i386 SX 20 MHz 32
-
- >>The test is developed to show which processor is more powerfull. But they did
- >>not tell, how the test is performed, so I do not know what these numbers stand
- >>for.
-
- >>Can someone bring some light in the darkness .....
-
- >I can tell one thing about their benchmarks--they clearly use the co-processor.
- >Otherwise, the 486SX and 486DX numbers would be identical. That's just come-
- >thing to keep in mid if you don't anticipate using any applications which
- >use the FPU. For apps which don't use it (well over 90% don't), the SX is
- >every bit as fast as the DX.
-
- From PC Magazine, December 8, 1992:
- "The current iCOMP index is made up of four weighted components: 16-bit
- integer (67 percent), 16-bit floating point (3 percent), 32-bit integer
- (25 percent), and 32-bit floating point (5 percent). Those weightings
- reflect Intel's best guess about the importance of each component in
- applications over the next three to five years (the average system
- lifetime). But Intel expects to update the formula as new CPU
- capabilities emerge (graphics and video are two likely candidates), and
- as applications change (as more of them take advantage of 32-bit
- technology, for example)."
-
- The full article appears on page 31.
-
-