home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!kithrup!hoptoad!pacbell.com!mips!darwin.sura.net!wupost!usc!sol.ctr.columbia.edu!ira.uka.de!uka!uka!news
- From: S_JUFFA@iravcl.ira.uka.de (|S| Norbert Juffa)
- Newsgroups: comp.benchmarks
- Subject: Re: Weitek 4167: Speed in MFLOPS?
- Date: 19 Aug 1992 15:13:49 GMT
- Organization: University of Karlsruhe (FRG) - Informatik Rechnerabt.
- Lines: 69
- Distribution: world
- Message-ID: <16tofdINNctu@iraul1.ira.uka.de>
- References: <1992Aug14.091124.304%boot.decnet@edwards-tems.af.mil>
- NNTP-Posting-Host: irav1.ira.uka.de
- X-News-Reader: VMS NEWS 1.23
- In-Reply-To: fantini%boot.decnet@edwards-tems.af.mil's message of 14 Aug 92 09:11:24 -0700
-
- In <1992Aug14.091124.304%boot.decnet@edwards-tems.af.mil> fantini%boot.decnet@edwards-tems.af.mil writes:
-
- > Hello Net,
- >
- > I have a question about the Weitek 4167 math co-processor:
- > I know this much: It costs >$500.00 and it is reported to be faster
- > than the co-processor in the 486DX.
- >
- > Can anyone tell me the advantages/disadvantages of this device for use
- > in a 486DX/33? How fast (in MFLOPS) are double precision calculations?
- > It is sometype of array processor that assist in vector type calculations
- > but not in scalar? What benchmarks have been run on this device
- > and what are the figures? How does it compare to (for example) a VAX 6610?
- >
- > It is worth the cost? Any info and comments (pro and con) are welcome.
- >
- > -Thanks
- > --
- > ---------------------------------------------------------------------------
- > Fantini@edwards-tems.af.mil | 6510 TSWS U.S.A.F. Simulation Lab
- > Mathematican & Loose Cannon | "Zu viel Rechnerkraft? Unmoeglich!" - Me
- > ---------------------------------------------------------------------------
-
-
- Comparison of floating point performance
-
- single precision
-
- Weitek 4167-33 Intel 486-33 Intel 486DX2-66
-
- LINPACK MFLOPS 5.0 1.8 3.5
- kWhetstones/sec 22700 12700 25500
-
-
- double precision
-
- Weitek 4167-33 Intel 486-33 Intel 486DX2-66
-
- LINPACK MFLOPS 3.5 1.6 3.1
- kWhetstones/sec 14000 12300 24700
-
-
-
- References:
-
- Intel486(tm) DX2 Microprocessor Performance Brief. 1992
- Order No. 241254-001
-
- WTL 4167 Floating-Point Coprocessor Data Book. Weitek
- Corporation, July 1989
- DOC No. 8943
-
-
- Note that the Weitek Abacus 4167 has no extended precision, no
- built-in transcendental functions and no support for denormals,
- while the Intel 486 supports all these features. While there are
- about 2100 software packages that support the floating-point
- capabilities of the i486, only high-end products like AutoCAD
- support the Weitek. You also need a high-end compiler such as
- Metaware High-C to generate Weitek code from a HLL, if you do
- not want to code in assembly language using the macros supplied
- by Weitek. If I needed high floating-point performance, I would
- opt for a one of the 486DX2-66 system that have recently be
- introduced to the market.
-
-
- Norbert
- -------------------------------------------------------------------------------
- Norbert Juffa email: S_JUFFA@IRAVCL.IRA.UKA.DE Live and let live!
-