[rescue] SMP on intel wasteful?

[David Passmore]

On Tue, Jun 25, 2002 at 11:49:45AM -0400, Chris Hedemark wrote:

> For the archive handicapped, I'll summarize.

Wow. You really /are/ an asshole. I'll try to keep this short.

> A took a <$2,000 PeeCee running Linux (a machine that would now be
> chipped according to the standards of Mr. McGuire's recycling friend)
> and a fully populated Sun E450.  I ran a number of EDA tools from
> Synopsys, most notably VCS and Design Compiler.  The PeeCee averaged

[snip]

Saw that. How you're drawing a generalization of 'PCs whip UNIX boxes' ass',
I don't know.

> Ah, I see.  I am the guy showing real world examples of why x86 is a
> viable business platform for UNIX workstations.  The other guy says
> "bite my arse hairs" and I am the one making an ass out of myself.
> 
> That's about as clear as mud.

I will be the first to say that Dave sometimes (often) makes an ass of
himself. I've worked with Dave in the past and I give him that leeway
because I respect him. You haven't earned my respect, and thus you don't get
the same treatment.

> PC hardware was not a viable platform for UNIX workstations back then,
> based on its own merits.  Today, I contend that it is, based on its own
> merits.  The load balancing systems are irrelevant.

I would challenge you to produce evidence of a large PC server installation
at a $major company that is not behind a load balancer.

> Don't take things out of context.  Indeed, in the example that I gave,
> performance *is* key.

It's just an example. It doesn't allow you to draw gross generalizations
about how great PCs are. I'm the only one in this thread who has agreed with
you that they have a place, and you're still arguing with me.

> Also, the Sun server farm had *worse* uptime than the PeeCee farm that
> replaced it.  These were hardware failures here.  In one case, a Sun
> E450 self combusted and forced me to clear the second floor to ventilate
> the noxious burning plastic fumes.

Can you produce uptime numbers?

> That is a rather sweeping generalization.  Yes, the majority of PC's are
> made of haphazardly slapped together components.  Then again, I've seen
> some Sun clones made in much the same way.  One of them actually
> electrocuted one of our engineers when he tried to hit the ON button.

I don't buy Sun clones. I've had several Marathon boxes go up in smoke. When
I want reliability, I buy from the best and expect the best. 

> BTW - I used to work for $big_vendor on their load balancing product
> team.  Most of the load balancing customers were actually running AIX
> and Solaris on their balanced servers.  Not PC.

Okay, I'll stick my penis out there; I've worked with F5, with Excite at Home,
with Digex, and now with AOL. There are lots of dedicated UNIX boxes behind
load-balancers, but in the enterprise, that's the /only/ place you'll find
PC hardware.

> I won't argue with you on that.  My previous postings stated that I
> would prefer critical RDBMS functions on sparc hardware.

Then why the thread pushing PCs as the end-all be-all of computing? I can
only assume you like to piss people off unnecessarily.

> I disagree.
> 
> I did some shopping around last night.  I won't bother quoting prices.
> For the price of one Sun Ultra 60 workstation, which is described as a
> low cost machine, I can have three dual processor PC servers with
> hardware RAID, redundant power supplies, redundant fans, etc.

Quote prices from a major vendor (Dell, HPaq, IBM) and put your money where
your mouth is.

> The bus on some of the higher end RISC boxen may indeed be faster.
> 
> But the DASD is no faster.
> 
> The network adapters are no faster.
> 
> About the only significant area where higher end RISC boxen will make a
> difference is memory speed.
> 
> But then the CPU is the bottleneck.

It is exactly the slower components that will produce the most contention on
a bus, no matter how fast it is. When you have a single bus, when a slow
device is transmitting data, it hogs the bus. With a circuit switched
architecture, other devices are free to transmit at the same time.

If you really want to put your money where your mouth is, name a stateful
application that has more than a terabyte of data set, and I will build you
a non-PC system that will do it with better performance and reliability than
anything you can muster-- of course, you have to build the equivalent PC
system, and will provide documentation and price quotes on everything you
build. We will come up with a joint test plan that exercises the systems
with functionality, stress, fault, and profile testing.

David