Date: Fri, 14 Apr 95 18:20:44 -0700 From: Bakul Shah <bakul@netcom.com> To: rgrimes@gndrsh.aac.dev.com, freebsd-scsi@FreeBSD.org Subject: Re: SCSI command overhead Message-ID: <199504150120.SAA21458@netcom10.netcom.com> In-Reply-To: Your message of "Fri, 14 Apr 95 19:48:39 EDT." <199504142348.TAA03880@hda.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Folks, thanks for the data. The 50 ~ 100 us I saw was entirely the SCSI chip overhead (WD33C93A), measured with a logic analyzer and confirmed by looking at cycle counts data wormed out of Western Digital. No disk or memory access was involved in this measurement. This was for an embedded system where SCSI was to be used for some `real time' communication (don't ask why). I was hoping that some of the new chips used simpler state machines and got out of the way of data transfer ASAP. Part of the problem *is* those more complicated state transitions in the target mode. Add to that time spent by the drive processor to map a block number to hd/sec/trk, look up any block forwarding, deal with variable number of sectors/trk etc.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199504150120.SAA21458>