home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!stanford.edu!agate!dog.ee.lbl.gov!lbl.gov!vxwexplo
- From: Roeber <fjr@ssd.ray.com>
- Newsgroups: comp.os.vxworks
- Subject: re: performance of fprintf() for floating point
- Date: Tue, 12 Jan 1993 08:58:24 -0500
- Organization: Lawrence Berkeley Laboratory, Berkeley CA
- Lines: 62
- Sender: vxwexplo@lbl.gov
- Message-ID: <199301121358.AA21035@pike.ssd.ray.com>
- NNTP-Posting-Host: 128.3.112.16
- Originator: daemon@vxw.ee.lbl.gov
-
- Joe Van Andel writes:
-
- > On a MVME133 (68020 @ 20Mhz), running under VxWorks 5.02b, release 1, I
- > can only write 3-4 lines/second, even if I'm writing to "/null". The
- > MVME133 has a co-processor, and I can't figure out what could possibly
- > be taking so long . .
-
- Below is an old message on the subject:
-
- ---------- Begin Forwarded Message ----------
-
- I did some grinding and got the cvtb (2) routine to go much faster.
-
-
- The Results:
-
- Hardware:
- mvme133xt, 25MHz, 68020
-
- Test:
- timexN sprintf, y, "%7.2lf", x
-
- Old cvtb (2) without floating point CP: 33 ms
- Old cvtb (2) with floating point CP: 16 ms
-
- New cvtb (2) without floating point CP: 9 ms
- New cvtb (2) with floating point CP: 2 ms
-
-
- That is about as good as I'm going to get it without spending much more
- time. But is this fast enough for you? Your whole conversion will still
- take 10-15 ms. (15 times faster) That's well above your motor latency.
-
- So tell me what you want...
-
- I can give you a new floatLib.o for your library.
- Or if you want to try to grind this code some more I could even
- send you source to floatLib.c.
-
- The moral of the story is floating point conversions to ascii are very
- expensive operations. Not too much I can do about that.
-
-
- John Fogelin
-
- Wind River Systems
-
- ---------- Begin Forwarded Message ----------
-
- At a later point Alan Biocca posted a routine for fast printing of floating
- point numbers on 680X0 processors. I am not sure but the routine might
- be in the VxWorks archive. Fred
-
- _______________________________________________________
- | Fred J Roeber, Raytheon Submarine Signal Division |
- | 1847 West Main Road, Mail Stop 188 |
- | Portsmouth, RI 02871-1087 (401) 847-8000 (X4205) |
- | |
- | fjr@ssd.ray.com |
- |_______________________________________________________|
-
-
-