🚀 go-pugleaf

Michael Wong <nospam_mikwon@ebtech.net> wrote in article
<3498a3df.3709664@news.ebtech.net>...
> >Nice try attempting to change the subject.  You seem to spend a lot of
time
> >trying to say what I've said or think and most of what you attribute to
me
> >is totally false.  
> 
> I notice how you make no attempt whatsoever to address my specific
> points here, preferring to simply dismiss my entire post as a single
> block with your pithy statement above.  Interesting tactic.

Then where did the following come from?

> 
> >Depending on the drive model,  the track one speed holds from between
> >15%-33% of the drive's capacity.  All it takes for a workstation's hard
> >drive performance to be transfer rate dominated is that the average
> >read/write be somewhere around 150K.  What do you think the read/write
> >ratio is on an average workstation usage excluding the swap file?  
> 
> #1) Where do you get your evidence that the track-one speed is
> rock-solid for the first 15% to 33% of the drive's capacity?  Does it
> suddenly drop down for the next 33%, and then suddenly drop down again
> for the last 33%?

Exactly stairsteps but from 4-10 usually.  Frequently the first notch(zone)
is big and the rest smaller.  Have you ever researched the concept of zone
recording and how it's implemented?

> #2) All it takes is that the average read/write is arond 150K?  Prove
> this.  Suppose I read twelve 150K blocks simultaneously.  Will that be
> transfer rate dominated?

If the OS and drive are properly setup for workstation usage, yes.

>  And what's this nonsense about excluding the
> swap file?

Excluded from consideration from writes to the slow part of the disk as the
swap file is generally preallocated and writes of new files on your "full"
disks tend to go towards the end where the disk is slower.  Swap file
writes on a "full" disk are not necessarily slow.


>  Do I get to exclude the swap file when I run a huge
> compile?  Do I get to exclude the swap file when I'm crunching
> toolpaths?  Tell me, when somebody wants to run two simultaneous
> simulation operations on his dual-CPU workstation, where do you find
> the "no simultaneous I/O operation" button, and the "don't touch
> swapfile" button?  I would love to know, since you seem to have found
> both of those buttons for YOUR theoretical workstations.  Oh wait- I
> forgot, none of your theories are based on real-world experience.
>

Obviously you're missing the important issues as the above indicates.


Regards,

Ron Reaugh