home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.os.os2.misc:29255 comp.sys.ibm.pc.hardware:23384 comp.benchmarks:1367
- Newsgroups: comp.os.os2.misc,comp.sys.ibm.pc.hardware,comp.benchmarks
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!usenet.coe.montana.edu!news.u.washington.edu!uw-beaver!ubc-cs!unixg.ubc.ca!ochealth
- From: ochealth@unixg.ubc.ca (ochealth)
- Subject: Re: Text scroll benchmark
- Message-ID: <1992Sep4.025108.24826@unixg.ubc.ca>
- Keywords: text scroll benchmark window
- Sender: news@unixg.ubc.ca (Usenet News Maintenance)
- Nntp-Posting-Host: unixg.ubc.ca
- Organization: University of British Columbia, Vancouver, B.C., Canada
- References: <14gnjd+.jps@netcom.com>
- Date: Fri, 4 Sep 1992 02:51:08 GMT
- Lines: 50
-
- In article <14gnjd+.jps@netcom.com> jps@netcom.com (John Serafin) writes:
- >
- >I am proposing a text scrolling benchmark. Although most computers can
- >scroll text much faster than it is possible to read, scrolling delay can
- >be quite objectionable during interactive modem sessions. Scrolling text
- >in a window in a GUI environment can take up so much CPU time that full
- >modem throughput is not realized. I have made some measurements and my
- >results are at the end of this article. My motherboard is an ISA 486
- >from Intel with an Intel chipset. For some reason, vector oriented
- >graphics run faster on Taiwanese motherboards with the same SVGA card.
- >Therefor other people with ProDesigner IIs may get better results.
- >
- >I would be particularly interested in seeing how various 8514, XGA, and TIGA
- What surprises me about OS/2 is how poorly screen writes are done.
- Perhaps it's the Trident driver that's poorly coded. I am running
- 1024x768 256 mode.
-
- The Trident card is quite good hardware, and their driver support is good
- (just wish they'd HURRY UP with fixing these beta drivers!)
- Normal VGA mode is very quick when scrolling, but 1024x768 sucks.
- Everything else about 1024x768 is quick. I'm also using a 486 33MHz 12 meg
- machine, so that's not causing performance problems.
-
- The reason scrolling is slow is that OS/2 seems to write one character at
- a time, or maybe a line at a time. This slows things to a crawl in 1024x768
- mode. *However* screen updates for a large area of text are quite fast.
- In a windowed TE/2 session, 2400 bps, the line by line scrolling sends
- Pulse up to 100% CPU. But if I click the mouse in the window for a few seconds
- the scrolling stops, and when I release the mouse, the whole window updates
- in a flash. Pulse shows much reduced CPU usage.
-
- I notice that fat pig X Windows on a Sun 3/60 is much faster(screen) than my
- 486. X is in C, the 16 bit PM stuff is all assembler. What gives?
- Well if you watch an Xterm closely, it updates a terminal window in
- a kind of "burst mode", kind of "lazy writes" for the screen.
- (BTW, the 3/60s have unaccelerated framebuffers)
-
- I'll wait for IBM to tweak the video some more, before I think of
- a grahics accelerator.
-
-
-
- >jps@netcom.com | Operating an automobile is more like riding than driving.
-
-
- --
- ______________________________________________________________________________
- jpm: ochealth@unixg.ubc.ca
- Happily using OS/2 2.0 because MS Windows isNT ___
- Insert VapourFeature ^^^
-