[rescue] Sunburst

gsm at mendelson.com gsm at mendelson.com
Fri May 21 07:56:59 CDT 2010

On Fri, May 21, 2010 at 03:53:59AM -0700, Skeezics Boondoggle wrote:
>But that has always been my experience with Intel boxes.  I've seen
>100+ interactive users working away comfortably on a SS5, when a then
>"beefy" Pentium running the same job mix would fall over dead with 20
>or 30 users.

Was that with Linux or UNIX? Over the years there have been purpose built
X86 computers that had a decent I/O architecture and a decent version of

One customer I had in 1995 had an 8 processor pentium 90 machine from
NCR (with microchannel I/O) that was for the time fast. It ran a version
of UNIX. 

>  I've seen this repeated through multiple generations,
>from the older SuperSPARCs to the US-IV+: they degrade gracefully
>under extremely heavy loads, while x86 boxes fall over badly.

Up until recently Linux had a scheduler that really performed poorly under
heavy loads. In the 2.6 kernel (or around that time anyway) a new scheduler
was implemented that supposedly fix those problems, but I have no way of
verifying the hype.

In some ways I'm surprised by MacOS, the Leopard scheduler degrades nicely
under load, even though it was designed for a single user workstation.

>Normalized for clock rate, the SPARCs actually _could_ usually compete
>with the Intel chips on single-threaded tasks, but if you could get
>2-3x the clock rate at 1/2-1/3 the sticker price, it was always an
>extremely hard sell the get management to believe that the
>cheaper/"faster" box wasn't the best decision.

This was the creation of the DEC marketing department. They used to push
more MIPS per Dollar (although it may have been KIPS) in those days. It's
ironic because the more MIPS with no consideration about what each
instruction did is what eventually brought down DEC.

>  But now, with Oracle
>apparently doing everything they can to make Solaris and SPARC even
>more irrelevant and obscure than they already were, it's likely moot
>anyway.  Soon not many people outside of lists like this will remember
>or even care about this stuff.

That's IMHO FUD. IBM did that for a long time for mainframes, and all it
did was keep staff availibity low, and salaries high. One of the things that
has destroyed the employment market is the profusion of Linux script kiddies
who take low paying sysadmin jobs. 

Sun failed because they tried to give too much away, and in good times
they did well. In bad times, they needed more profit, but could not get it.
Ellison, for all of his faults, is going to make a profit off of SUN.

Having been a benefactor of SUN's largess since the mid 1990's, I thank them
for it, and appreciate it. On the other hand I was not a stockholder,
and probably happy for it.

>OBTW,  that U1E I ran all those builds on ten years ago is still
>running (albeit with a dead NVRAM battery!) as a
>console/DNS/NIS/NTP/paging server here at home:

When you do shut it down, you can remove the NVRAM chip, open it up and
replace the battery. If it's the usual Dallas chip, it just requires a pair
of cutters and a battery with wires. These days, you can buy a socket for
a lithium coin cell cheaply and use that. 

I did that a lot around 1998 (has it really been that long?) when I had 
3 SUNs running in a cluster with obviously fake serial numbers/MAC
addresses. As I remember they were 0xcoffee 0xcoffed and 0xcoffef. :-)


Geoffrey S. Mendelson, Jerusalem, Israel gsm at mendelson.com  N3OWJ/4X1GM
New word I coined 12/13/09, "Sub-Wikipedia" adj, describing knowledge or
understanding, as in he has a sub-wikipedia understanding of the situation. 
i.e possessing less facts or information than can be found in the Wikipedia.

More information about the rescue mailing list