Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin AT sources DOT redhat DOT com Message-Id: <5.1.0.14.2.20010713193521.025c85e8@pop3.cris.com> X-Sender: rrschulz AT pop3 DOT cris DOT com X-Mailer: QUALCOMM Windows Eudora Version 5.1 Date: Fri, 13 Jul 2001 19:49:15 -0700 To: jorgens AT coho DOT net, cygwin AT cygwin DOT com From: Randall R Schulz Subject: RE: Curious - why Cygwin Bash so much faster in rxvt than Windows CLI? In-Reply-To: <01C10BD2.A811EFB0.jorgens@coho.net> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Steve, I ran another test with a different font in RXVT, and got very different results: New RXVT test: ...\rxvt.exe -g 120x91+853+0 -bg #000055 -fg #eeeeee -fn "Andale Mono-12" -sr -sl 5000 -vb -e /bin/bash --login -i % time ls -lR / real 0m32.422s user 0m5.920s sys 0m21.781s % tsize rows 92; columns 120 (Note that I invoke rxvt requesting a 91-line-high window and get a 92 line-high window. This off-by-one always happens to me with RXVT, for whatever reason--see below.) Previous RXVT test: ...\rxvt.exe -g 120x99+732+0 -bg #000055 -fg #eeeeee -fn "Lucida Console-11" -sr -sl 5000 -vb -e /bin/bash --login -i % time ls -lR / real 1m33.360s user 0m6.765s sys 0m31.561s % tsize rows 100; columns 120 Previous Windows Character Subsystem window: ...\bash.exe --login -i % time ls -lR / real 1m27.782s user 0m7.296s sys 0m23.608s % tsize rows 93; columns 120 In the previous test, the windows were nearly the same size (in pixels). In the second case the choice of font yielded a closer match to the character cell dimensions of the Windows Character Subsystem window, but the RXVT window size in pixels was about 15% narrower and the Character window. In all cases, the windows very nearly (or exactly) filled the screen's 1200 vertical pixels less a double-high "bar" (single-height taskbar + single-height quick-launch bar). Performance measurement is _never_ simple or straightforward! Randall At 19:33 2001-07-13, Steve Jorgensen wrote: >There may not be a real performance effect, but there is a definitely >perceptible latency. You see it, right? I'm glad to know it's not >actually interfering with processes, though. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/