🚀 go-pugleaf

In article <1992Aug18.173248.21572@leland.Stanford.EDU> m@crito.stanford.edu (M Carling) writes:
>You compare real-world tests with the drive manufacturers' claims  
>seemingly in an attempt to show that NeXT's SCSI driver is inferior to  
>that of other vendors. Please, if you want to make such a claim, test  
>NeXT's SCSI performance against someone else's SCSI performance, not  
>against claimed theoretical performance.

I would not exactly call "disk / read" a "real-world" test.  How much more
ideal can you get?  Also, I'm not claiming that other *UNIX* SCSI driver's
are necessarily good--simply that NeXT's could stand some improvement.

>This is correct. NeXT currently supports neither sychronous SCSI nor  
>tagged command queueing. I unsuccessfully lobbied to have them included in  
>3.0.

That makes two of us.

> Before NeXT will support these SCSI features, they will put every  
>type of hard drive that they ever shipped through qualification testing to  
>make sure that implementing them won't "break" anyone's system. I am  
>confident, however, that NeXT will eventually support these features.

I woould have preferred to see them put it in but be disabled by default
so that adventuresome users could have enabled it at their own risk.

--Marc