Date: Wed, 23 Aug 1995 11:08:45 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com> To: se@ZPR.Uni-Koeln.DE (Stefan Esser) Cc: james@miller.cs.uwm.edu, current@freebsd.org Subject: Re: Matrox Meteor Video Capture Card Driver Message-ID: <199508231808.LAA09847@gndrsh.aac.dev.com> In-Reply-To: <199508231549.AA18689@Sysiphos> from "Stefan Esser" at Aug 23, 95 05:49:07 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > On Aug 23, 10:37, Jim Lowe wrote: > } Subject: Re: Matrox Meteor Video Capture Card Driver > } > From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com> > > } I tried lowering the PCI latency timer from 80 to 40, then to 20, then 1, > } then 0. It helped me get a few frames at 640x480 before the machine > } hung. I also tried disabling the cpu to pci burst mode. > > It takes some 10 cycles to arbitrate for the PCI bus > and to send the first data word. Numbers less than > 10 (decimal) are not too useful for this reason. Arbitration cycle counts occur _after_ latency timer expire. Numbers lower than 10 do help (but should not be need unless someone has left a fifo out of a high data rate card). A PCI latency timer of 0 says if someone says they would like to get on the bus and I am on the bus I get off _NOW_. Jim, when we talked on the phone the other day you said this board had a bunch of fifo's on it, how deep are they? Can you program depth and bus requests thresholds? > Is there any error message or other fault indication > than the system freezing ? > > } I have been running the meteor in a machine with and adaptec controller > } for 5 days and it never hung. I am still not sure why the machine > } hangs with the ncr controller. > > I take it the Adaptec above is an 2940 ? > > Could you please try having "ncrconsole -v -p 1" > run in another window. Watch the "pre" and "post" > times, which account for the time to initiate a > transfer and to retrieve the SCSI status. > > The lower limits for these numbers are the NCR and > disk drive SCSI overhead. The command overhead of > the NCR depends on its instruction execution rate > (usually in the low MIPS range). > > Do you have a NCR 53c815 or 825 to compare the 810 > results with ? > > One of the differences between the 8x0 and 8x5 is > the reading of instructions in burst mode, which > might help. If he ran this is Georges box it has an 825 in it. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199508231808.LAA09847>