It is both amusing and frustrating to hear all of the "armchair computer scientists" discussing the reasons this technology is a bad idea. As if they knew more about the subject than the many PhD's who have dedicated their careers to this subject based on the knowledge gleamed from the one Computer Architecture class the poster took as an undergraduate.
I was invited to work on a team at the University of Utah (Sutherland's old school) where they were researching this very topic. This is old news; they have been working on it for years. And as some people have correctly pointed out, there are both good and bad points to sync or async logic.
There are two major reasons to work on async logic: clock skew and power savings. The reason for power savings alone is a good one. People here have been complaining that it "is not worth it for only a 20% power savings".
Yes it is! In a modern office, computers end up taking a lot of power. Imagine your local server room. Don't you think they would like a 20% decrese in their power bill?
That means instead of building five power plants, you only need four (on a grand scale; please no newbie replies like doodz, thiz guy thinkz you n33d five pawer pl3nts to run a box). That is significant. And with today's high-MHz CPUs this means even more. Some think >50% savings, and even more during low cycle time.
The clock skew issue has been covered somewhat here. One of the major hurdles in solving the design problem is the development of new design tools, which is what many people at Utah are currently working on.
The way to move forward is not to argue for the limitations of systems of the past. Don't make me pull out Ken Olson quotes here.
That's the whole point of many years of resarch on the subject. This is nothing new.
The CPU hits bottlenecks in slow components now. The difference is that right now everything keeps running at full power during a bottleneck while it does nothing at all. With async design, when components are waiting they take a fraction of their power as there is not clock cycling them off-and-on.
Also, the layout and design of a synchronous clock is a major limiting factor in CPU design.
I am a non-Mormon who grew up in Utah, and I am
writing this from my desk at work in Salt Lake City.
Almost everything said in the preceeding message
is a pile of horseshit.
When my brother lived in UT and worked for Novell, he lived in a suburb full of child-bearing age couples with many children, but
none of them played outside where they might meet a gentile
When I was a child there was one family who
would not let their children play with me as I was
not a mormon. Yes, it was terrible. But they
were allowed to play outside with all the mormon
kids, and the other 50 mormon families had no
problem with me.
If you order a Coke at a fast food joint, it's flat and nasty tasting, because it's the first one they've sold in a week.
Okay, several things:
1: Mormons drink coke. Cold caffine-containing
beverages are okay, just not hot ones. It makes no sense to me either, but that's the rule. In fact, the local coke bottler (Swire Coca-Cola) is
owned in part by the Mormon church. The mormon
university "BYU" does forbid students from drinking caffine, but they are an exception.
2: Most mormons I know are addicted to
coke even more than hackers. A mormon friend of
mine goes into physical withdrawl without at least
2l of Mountain Dew per day.
3: Since fast food establishments mix CO2 and the
syrup at the dispenser, it is impossible for the
drink to be "flat and nasty tasting", unless all the drinks are flat.
I suppose it is possible you did have a
brother who lived in Utah and told you these
things, but he is either a liar or a fool. Also,
Novell is located in Provo, which is so boring
and full of Mormons that most of us stay away
from the place.
There are fifteen people working here with me at a new Utah startup company, and only one mormon. My last company had 100 engineers and
probably 20 non-engineers and was very mormon; about 60-70 percent (as they had hired most of their staff from an old company that was very
mormon-dominated). But we had beer at are monthly
picnics, and we all got along.
Salt Lake City is only about 40% Mormon. Yeah, the liquor laws are a pain, but hardly the end
of the world.
The United States also tried using nukes for
civil purposes. The "plowshares" program tried
using atomic weapons for excavation at the
Nuclear Testing site in Groom Lake, Nevada.
The craters--intended for water storage or
for construction--were so radioactive that they
never went forward with the program beyond the
test site.
Apparently the Soviets actually made some
reservoirs with their bombs.
Regarding your last statement...Slashdot is going downhill. There are a lot
of kiddies here who are quick to fly off the
handle, or take some minor point that is insignifigant to the argument of a
discussion and use it to flame a poster.
It's funny to see Hennessy and Patterson talk like that after they pushed that silly benchmark so hard in the first place. Can you imagine Donald Knuth pulling a 180 like that?
There have been a lot of good comments made by
everyone, so I will only add a couple of my own.
As others have said, speed is best-calculated as whole system speed. Not only the speed of the components (hard drive RPM, for example) but the technology and implementation of that technology is a factor.
For example, is the hard drive 5400 or 10000?
Is it IDE? If so, how large is the cache? What
is the transfer rate?
Even the fastest IDE will slow down a system with
overhead, so lets talk SCSI. Which SCSI? SCSI, SCSI-II, Fast SCSI-II, Wide or Ultra or Ultrawide or U2W or UW160 or Fiber Channel or....
Now, is it an elcheapo controller from CompUSA
or is it a high-end busmastering controller? The difference is huge.
Okay, that's a lot of data about the disk
system. For a server or a workstation it will
make a huge difference. But if I am just cracking
SETI packets it will make almost no difference.
See where I am going with this? There are too
many variables to fit into any standard benchmark. The only decent way to test any computer is with your own software; the application you intend to be running.
Failing that, the best way for me--sadly--is the MHz of the chip. Why? If it is a decent (Intel, High-end Asus, SuperMicro) MB, the
speed will be about the same from system-to-system. Cheap computers all use IDE
disks that I will just turn into.MP3 archives,
and good computers usually all use the latest-and-greatest SCSI, often nice Adaptec
controllers. The MHz of the chip
gives me something to compare them with; two
similar configuration machines will be about the same speed cycle-to-cycle. As long as it is the same brand and model CPU! This does
not hold true for comparing a Laptop Pentium with a Pentium Pro, and comparing a Mac to an IBM is like comparing Apples to....you get the idea.
For real applications (servers) or across
different platforms (HP700 vs UltraSparc vs Origin vs Alpha), most buyers know what they are getting
into and run decent tests on the boxen.
Finally, for home users you can forget it; you can talk till you're blue in the face about why a Compaq DeskPro or Proliant will be faster than
the $600 Compaq of the same MHz down at the local
superstore, and they'll end up buying the $399
MediaGX E-machine anyway. Most home users only
care about price. Change the 600Mhz sticker to
1200Mhz and they will swear it is faster...
I know what you are saying. I am a Slackware fan myself, and it can be hard when programs specify "Red Hat 6.2 or better"--what the hell does that mean???SuSE has a neater logo.
The poster is probably stuck buying from Dell due to corporate supidity. We had to replace all
our functional VT320 terminals with PCs that
were always crashing and locking up. And all we ran on them was a terminal emulator the
pretended to be a VT320.
The new computers took more power, more space,
and did not work nearly as well. So why did
we have them? There was a contract to replace
all terminals, ignoring the fact that we did not NEED the ability to use a word processor next
to the CT scanner.
I am not the poster of the topic. The poster was not asking for some home-made kludge as
many people here have assumed, there are commerical companies that make enclosed (front,
back and side door) racks with forced-air blowing
throughout the entire case. Some even have
a dedicated air-conditioner. Throughout the
rack are mounted pull-out shelves that can hold
a motherboard. This is what I am talking about. Some have one standard PC-type power supply
mounted on each shelf, others have a single large supply. The ones I have seen advertised in the Linux mags seemed to include
the motherboards and drives. You'll note he mentions that he has "seen them advertised" in his post.
experiencing a 20% failure rate over a period of 4 years, well, I think you may need to ask yourself why in the world you got
such cheap junk to power your cluster. Don't expect to buy a good power supply for under $100[US], and even that much will
only get you a mediocre one.
Oh, I agree 100%. But most users here will buy
a whole case and power supply for $50. I purchased the case/supply for $100 at my old job and got chewed out for wasting money. This is
where the problem lies. And I worry that the
supplies in the racks may be substandard; the
only decent PC-type supplies I have found are the
ones from California PC (calpc) and the Linear
supplies from Integrand Research. A single high-end supply is better than 24 cheap Chinese ones. It is hard to find a good PC power supply...like most parts.
So, what you are saying is that you want a 'rack' with...
The ones I have seen are in standard relay
racks.
As far as your comments about not needing fans, I hope you are kidding. Say your single system uses about 200W...
The rack has huge blowers the cool the whole
rack. All I said (or meant to say) is in
the designs where there is one PC supply per
shelf (and therefore per motherboard), that PC supply (which only needs to cool
itself, not anything else) probably does not need
its fan, as the rack itself is kept so cool. I mentioned this during my rant on cheap PC supplies
with Magical Frying Fans. Like I mentioned
in another message, one rack here has a built-in McLean air condtioner. The supply itself would be kept cool by convection; chilled air would be blowing around it, and it would be mounted
to a metal shelf in a metal rack in a computer room. I was trying to give credit to your many-supplies viewpoint....
but you must air-condition the room, ideally to
below 25C.
I think these systems are almost always in
proper computer rooms.
I suspect your first effort had so many power supply failures because of improper venting and cooling.
I was just discussing failures due to cheap
power supplies on user's desks. I've never
designed one of these things; I would buy one off
of the shelf. In my own designs I like standard
5-unit rackmount cases--from Integrand (a.k.a. Tri-Mag) if I can afford it.
This is not a trivial
design, and I would suggest you either learn a lot more about thermal system design, or use a product which is from a
company with a lot of experience in this area.
The companies who sell the Linux systems based
on this design do--or at least claim to. The
rack we have here does not contain any computers;
they have been stripped out. But it was made
in a professional fab shop and they did a nice
job. I would photograph it if I had a digital
camera.
Okay, I know what you are saying, but I still
like larger supplies. The cheap supplies fail
a lot %-wise compared to "real" power supplies. In particlar components like the fan.
To be fair, I don't think you really need a
fan in most of these racks; the PS would cool
well enough through convection. But most of the
supplies use cheap components that fail at a
pretty alarming rate. The only large PC network
I set up lost a lot of power supplies, and I
bought the better ones.
Though a larger supply may be no more efficient, the sort of supplies used in large
computers tend to fail pretty infrequently, and
uptime may be more important than power use. Plue, if 20% of the cheap supplies fail in five years, and you have 24 of them chances are you
will lose five computers over that time. A single
hot-swap box (see below) probably has a failure rate in the low single digits...even if it is
as high as 5%, 2 x 0.05 is a lot lower than 24 x 0.2, if you see what I mean.
Plus, with a single supply you can mount the
supply in the unenclosed portion of the rack and
blow the heat into the room, not mix it into the
air circulating in the computer-portion of the
case. Also, this allows you to keep the AC even
farther from the computer--which is nice if
noise is a problem. I use one computer that not
only has the PS in a shielded portion of the
case, it is a linear power supply. Very heavy
and a lot less efficient, but the +5V line is
rock-stable on my scope.
As far as failure tolerance, while a single
supply does provide a single point of failure
most high-end systems use multiple hot-swap supplies. The network gear I help design uses
four supplies but can run on two (or one with
minimal blades). With more than one supply
you get load-balancing, but it lets you hot-swap
without a problem.
Now, the voltage drop is a good point. I've
run +5V quite a ways, and frankly with multiple 16Ga wires the drop is not even worth talking
about, even in a double wide rack (one machine of mine runs 5V through a four-wide rack with no
significant drop). Mini- and mainframe computers do this all the time. While newer machines use a PS-per-cage, look in the back of a pdp11/60 or
a big HP9000/800 sometime. There is a 5V supply you
can spot weld with! If it is a design that
powers a single cage, you can use sense lines
to regulate the voltage at the backplane...although not in this case. Remember that 5V can wander +/- 250mV with no problem for many designs.
Now, with 3.3 volts the voltage drop could be a concern. I don't know how the rack companies do it; one way
would be to put an "industrial brick" type
DC/DC converter from 5-3.3V on each shelf. You
could heatsink it to the shelf and not generate
too much heat. I don't know how they deal with
that issue in the commerical products.
Please people, don't chew out the guy unless
you know what he is talking about. Many of you clearly don't; these are commerical products
used in high-end computing applications. Remember that there are many products that really exist, even though you can't buy them at CompUSA.
This is a very good way of building a server.
The ElCheapo power supplies that come in PC cases are not only inefficent, they fail frequently and
take out everything downstream.
The single power supplies used in these racks
are overengineered to the point that they don't
fail in this manner. The racks usually have front
and back doors and a blower...one we have where
I work has a side-mounted McLean air conditioner! They are better cooled than any little PC box.
As far as manufactures go, I don't know. Some
used to advertise in Linux Journal (which I have
not taken for a couple years) but many of the
designs used one-supply-per-rack, which I really
don't care for. I like a single, decent-quality
supply more than 24 crummy ones. I recall
that 24-per-rack seemed to be the maximum you
could fit into a rack...this would make each
one about two rack units high, which seems to be
the smallest I can imagine. They were using
Intel all-in-one MBs with integrated 10/100 and VGA, with Black Box KBM switches. I would use
the Belkin switches myself due to $$$.
Most seem to have rolling ventilated shelves
with a IDE drive on each shelf.
Ours here at work has no identifing marks except "McLean" on the air conditioner, but as
McLean is a famous rack blower manufacturer I would think they just made the air conditioner,
not the rack. The rack was just shipped in from
a company called "NetVantage" that we bought, so
I can't tell you anything else.
I will check when I get home, I seem to recall
I ordered brochures from some of those companies.
The Athlon must be the 500Mhz then. There
is one K-6 and one Athlon and they are 50Mhz
apart. They are on the same desk. I don't
pay a lot of attention the the fine details of PC hardware; I'm tired of dealing with junk.
That was my point, to show that you just can't
do a generic comparison based on the MHz of the
CPU. The kiddie who posted the first message (the
one to which I replied) claimed that any x86
chip would blow the UltraSparc away 'cause "its faster d00dz".
My comment is only relevant when read in
context. It should not have been moderated to
+4. But the mods are often clueless....
You might want to tone down your inflammatory
language since you don't know anything about
me. Just a hint.
I am working for a small company with needs
for reasonable computing power and we just
purchased a top-of-the-line HP workstation.
I am more of a hardware guy. This machine
is very well built; not like the cheap 712/60's
I've used in the past. The electrical and
mechanical construction of this computer reminds
me of the big iron machines of the past.
Plus, it's faster than hell.
So, from a hardware standpoint I love HP; it
is sweet. But as others have remarked, HPUX is
a pain to admin--at least for a HPUX newbie. I
have admin'd Linux and BSD boxen for seven years,
but HPUX is a new animal. If it ever gets even
a tiny bit confused, it throws it's hands up
in the air and vomits. This annoys me.
As far as SGI systems go, they have never
struck me as very well built, even though I
like them. A few years ago the University I was
attending (getting a MS in EE) purchased
what was then the fastest graphics computer
in the world, a many (128?) processor SGI/Cray
in nine racks. The dust level in our machine
room (which was pretty high by computer room
standards--it was installed by Univac for the 1108) caused it to fail within a month. Even with the dust problems fixed (the FS guys took the entire thing apart and vacuumed each board--plus replaced a lot) the machine has been a constant battle, and
Irix issues cause it to reboot frequently.
An 1986 vintage HP9000 series 800 mini (the
fastest box in '86) has been running nonstop in
the same room for going on fifteen years with
almost no downtime. Yeah, it's old and is more
dust-tolerant, but it also seems better built.
Clearly I'm not an expert, but I would vote
for the HP hardware over the SGI. Too bad you
can't run your apps on the HP box with Linux
installed...
Please! As the English language has evolved over the last 934 years, "men" has always been
used to mean "humans".
Now suddenly in the last 20 years people have
decided that this means "just males" and we
have to change our entire manner of speaking
to keep from offending somebody. Women should
not be offended by this, and would not be had
somebody not drilled it into them that this was
a "slight" and therefore offensive.
Eventually we will have to say "Women, individuals-of-unknown-but-preferable-to-either-ma le-or-female-gender, transexuals, victims or
Turner's syndrome, and men".
Our NanoProbes are able to (benignly) penetrate a user's stealth firewall to verify the presence of the system hidden behind. Since our NanoProbes
are able to bypass stealthing
He seems to claim that his packets cannot be
blocked....watch me (or anyone else) block them. Seems high on
ego, low on content.
But a Sun UltraSparc 60 running 360Mhz machine
completes a setiathome packet in four hours. An
Athlon 450 "faster" machine completes the same
packet in 20 hours.
I work for a group that could actually use this
box. But, alas, while we could afford to buy and
power it, I don't think we could afford to pay
for the setup costs.
And I know for a fact that I can't
administer it!
Re:Sounds interesting, but I'm annoyed by the US c
on
Ash: A Secret History
·
· Score: 2
Gads, I missed a whole paragraph of the article. Caffeine time.
Sounds interesting, but I'm annoyed by the US cuts
on
Ash: A Secret History
·
· Score: 2
As there is a US version, I can only assume
that nobody will be importing this title so we'll
all have to order them from amazon.co.uk or
someplace. I hate buying an incomplete verion of anything.
One wonders who decided to make the cuts. Was
it the US publisher, or did somebody in the UK
decide that "those dumb yanks would never be able
to understand this".
As an aside, an amusing British view of the
US can be found at The American Adventure theme park (warning: site requires flash). Did you know
all Americans speak with southern-Ozark accents?
Slashdot Mirror
It is both amusing and frustrating to hear all of the "armchair computer scientists" discussing the reasons this technology is a bad idea. As if they knew more about the subject than the many PhD's who have dedicated their careers to this subject based on the knowledge gleamed from the one Computer Architecture class the poster took as an undergraduate.
I was invited to work on a team at the University of Utah (Sutherland's old school) where they were researching this very topic. This is old news; they have been working on it for years. And as some people have correctly pointed out, there are both good and bad points to sync or async logic.
There are two major reasons to work on async logic: clock skew and power savings. The reason for power savings alone is a good one. People here have been complaining that it "is not worth it for only a 20% power savings".
Yes it is! In a modern office, computers end up taking a lot of power. Imagine your local server room. Don't you think they would like a 20% decrese in their power bill?
That means instead of building five power plants, you only need four (on a grand scale; please no newbie replies like doodz, thiz guy thinkz you n33d five pawer pl3nts to run a box). That is significant. And with today's high-MHz CPUs this means even more. Some think >50% savings, and even more during low cycle time.
The clock skew issue has been covered somewhat here. One of the major hurdles in solving the design problem is the development of new design tools, which is what many people at Utah are currently working on.
The way to move forward is not to argue for the limitations of systems of the past. Don't make me pull out Ken Olson quotes here.
That's the whole point of many years of resarch on the subject. This is nothing new.
The CPU hits bottlenecks in slow components now. The difference is that right now everything keeps running at full power during a bottleneck while it does nothing at all. With async design, when components are waiting they take a fraction of their power as there is not clock cycling them off-and-on.
Also, the layout and design of a synchronous clock is a major limiting factor in CPU design.
The biggest point is that this will save enormous amounts of electrical power that can be used elsewhere.
Sutherland's work is nothing new. The Computer Science department at the (a department he founded) has been working on this for years as well. They have made significant progress.
I know, I worked with one of the professors for a while before I went into electrical engineering.
Just 'cuase the British government can't make it work does not mean it is impossible. Many inventions we use every day were considered "impossible".
Remember that since you don't know about the research, perhaps there is something you don't know.
I am a non-Mormon who grew up in Utah, and I am writing this from my desk at work in Salt Lake City.
Almost everything said in the preceeding message is a pile of horseshit.
When my brother lived in UT and worked for Novell, he lived in a suburb full of child-bearing age couples with many children, but none of them played outside where they might meet a gentile
When I was a child there was one family who would not let their children play with me as I was not a mormon. Yes, it was terrible. But they were allowed to play outside with all the mormon kids, and the other 50 mormon families had no problem with me.
If you order a Coke at a fast food joint, it's flat and nasty tasting, because it's the first one they've sold in a week.
Okay, several things:
1: Mormons drink coke. Cold caffine-containing beverages are okay, just not hot ones. It makes no sense to me either, but that's the rule. In fact, the local coke bottler (Swire Coca-Cola) is owned in part by the Mormon church. The mormon university "BYU" does forbid students from drinking caffine, but they are an exception.
2: Most mormons I know are addicted to coke even more than hackers. A mormon friend of mine goes into physical withdrawl without at least 2l of Mountain Dew per day.
3: Since fast food establishments mix CO2 and the syrup at the dispenser, it is impossible for the drink to be "flat and nasty tasting", unless all the drinks are flat.
I suppose it is possible you did have a brother who lived in Utah and told you these things, but he is either a liar or a fool. Also, Novell is located in Provo, which is so boring and full of Mormons that most of us stay away from the place.
There are fifteen people working here with me at a new Utah startup company, and only one mormon. My last company had 100 engineers and probably 20 non-engineers and was very mormon; about 60-70 percent (as they had hired most of their staff from an old company that was very mormon-dominated). But we had beer at are monthly picnics, and we all got along.
Salt Lake City is only about 40% Mormon. Yeah, the liquor laws are a pain, but hardly the end of the world.
The United States also tried using nukes for civil purposes. The "plowshares" program tried using atomic weapons for excavation at the Nuclear Testing site in Groom Lake, Nevada.
The craters--intended for water storage or for construction--were so radioactive that they never went forward with the program beyond the test site.
Apparently the Soviets actually made some reservoirs with their bombs.
Regarding your last statement...Slashdot is going downhill. There are a lot of kiddies here who are quick to fly off the handle, or take some minor point that is insignifigant to the argument of a discussion and use it to flame a poster.
It's funny to see Hennessy and Patterson talk like that after they pushed that silly benchmark so hard in the first place. Can you imagine Donald Knuth pulling a 180 like that?
There have been a lot of good comments made by everyone, so I will only add a couple of my own.
As others have said, speed is best-calculated as whole system speed. Not only the speed of the components (hard drive RPM, for example) but the technology and implementation of that technology is a factor.
For example, is the hard drive 5400 or 10000?
Is it IDE? If so, how large is the cache? What is the transfer rate?
Even the fastest IDE will slow down a system with overhead, so lets talk SCSI.
Which SCSI? SCSI, SCSI-II, Fast SCSI-II, Wide or Ultra or Ultrawide or U2W or UW160 or Fiber Channel or....
Now, is it an elcheapo controller from CompUSA or is it a high-end busmastering controller? The difference is huge.
Okay, that's a lot of data about the disk system. For a server or a workstation it will make a huge difference. But if I am just cracking SETI packets it will make almost no difference.
See where I am going with this? There are too many variables to fit into any standard benchmark. The only decent way to test any computer is with your own software; the application you intend to be running.
Failing that, the best way for me--sadly--is the MHz of the chip. Why? If it is a decent (Intel, High-end Asus, SuperMicro) MB, the speed will be about the same from system-to-system. Cheap computers all use IDE disks that I will just turn into .MP3 archives,
and good computers usually all use the latest-and-greatest SCSI, often nice Adaptec
controllers. The MHz of the chip
gives me something to compare them with; two
similar configuration machines will be about the same speed cycle-to-cycle. As long as it is the same brand and model CPU! This does
not hold true for comparing a Laptop Pentium with a Pentium Pro, and comparing a Mac to an IBM is like comparing Apples to....you get the idea.
For real applications (servers) or across different platforms (HP700 vs UltraSparc vs Origin vs Alpha), most buyers know what they are getting into and run decent tests on the boxen.
Finally, for home users you can forget it; you can talk till you're blue in the face about why a Compaq DeskPro or Proliant will be faster than the $600 Compaq of the same MHz down at the local superstore, and they'll end up buying the $399 MediaGX E-machine anyway. Most home users only care about price. Change the 600Mhz sticker to 1200Mhz and they will swear it is faster...
I've learned my lesson on that one.....
*sniff*....*sniff*
Hey, it smells like Hennessy and Patterson around here!
(We'll see if the mods get that one)You're correct, of course.
Now I can't tell my P's from by B's...sorry about the accidental shouting.
I know what you are saying. I am a Slackware fan myself, and it can be hard when programs specify "Red Hat 6.2 or better"--what the hell does that mean??? SuSE has a neater logo.
The poster is probably stuck buying from Dell due to corporate supidity. We had to replace all our functional VT320 terminals with PCs that were always crashing and locking up. And all we ran on them was a terminal emulator the pretended to be a VT320.
The new computers took more power, more space, and did not work nearly as well. So why did we have them? There was a contract to replace all terminals, ignoring the fact that we did not NEED the ability to use a word processor next to the CT scanner.
Yes, but a rack with a massive blower is cooler than a crummy PC case fan. This is what he is talking about.
Okay, first some confusion to clear up.
I am not the poster of the topic. The poster was not asking for some home-made kludge as many people here have assumed, there are commerical companies that make enclosed (front, back and side door) racks with forced-air blowing throughout the entire case. Some even have a dedicated air-conditioner. Throughout the rack are mounted pull-out shelves that can hold a motherboard. This is what I am talking about. Some have one standard PC-type power supply mounted on each shelf, others have a single large supply. The ones I have seen advertised in the Linux mags seemed to include the motherboards and drives. You'll note he mentions that he has "seen them advertised" in his post.
experiencing a 20% failure rate over a period of 4 years, well, I think you may need to ask yourself why in the world you got such cheap junk to power your cluster. Don't expect to buy a good power supply for under $100[US], and even that much will only get you a mediocre one.
Oh, I agree 100%. But most users here will buy a whole case and power supply for $50. I purchased the case/supply for $100 at my old job and got chewed out for wasting money. This is where the problem lies. And I worry that the supplies in the racks may be substandard; the only decent PC-type supplies I have found are the ones from California PC (calpc) and the Linear supplies from Integrand Research. A single high-end supply is better than 24 cheap Chinese ones. It is hard to find a good PC power supply...like most parts.
So, what you are saying is that you want a 'rack' with...
The ones I have seen are in standard relay racks.
As far as your comments about not needing fans, I hope you are kidding. Say your single system uses about 200W...
The rack has huge blowers the cool the whole rack. All I said (or meant to say) is in the designs where there is one PC supply per shelf (and therefore per motherboard), that PC supply (which only needs to cool itself, not anything else) probably does not need its fan, as the rack itself is kept so cool. I mentioned this during my rant on cheap PC supplies with Magical Frying Fans. Like I mentioned in another message, one rack here has a built-in McLean air condtioner. The supply itself would be kept cool by convection; chilled air would be blowing around it, and it would be mounted to a metal shelf in a metal rack in a computer room. I was trying to give credit to your many-supplies viewpoint....
but you must air-condition the room, ideally to below 25C.
I think these systems are almost always in proper computer rooms.
I suspect your first effort had so many power supply failures because of improper venting and cooling.
I was just discussing failures due to cheap power supplies on user's desks. I've never designed one of these things; I would buy one off of the shelf. In my own designs I like standard 5-unit rackmount cases--from Integrand (a.k.a. Tri-Mag) if I can afford it.
This is not a trivial design, and I would suggest you either learn a lot more about thermal system design, or use a product which is from a company with a lot of experience in this area.
The companies who sell the Linux systems based on this design do--or at least claim to. The rack we have here does not contain any computers; they have been stripped out. But it was made in a professional fab shop and they did a nice job. I would photograph it if I had a digital camera.
Okay, I know what you are saying, but I still like larger supplies. The cheap supplies fail a lot %-wise compared to "real" power supplies. In particlar components like the fan.
To be fair, I don't think you really need a fan in most of these racks; the PS would cool well enough through convection. But most of the supplies use cheap components that fail at a pretty alarming rate. The only large PC network I set up lost a lot of power supplies, and I bought the better ones.
Though a larger supply may be no more efficient, the sort of supplies used in large computers tend to fail pretty infrequently, and uptime may be more important than power use. Plue, if 20% of the cheap supplies fail in five years, and you have 24 of them chances are you will lose five computers over that time. A single hot-swap box (see below) probably has a failure rate in the low single digits...even if it is as high as 5%, 2 x 0.05 is a lot lower than 24 x 0.2, if you see what I mean.
Plus, with a single supply you can mount the supply in the unenclosed portion of the rack and blow the heat into the room, not mix it into the air circulating in the computer-portion of the case. Also, this allows you to keep the AC even farther from the computer--which is nice if noise is a problem. I use one computer that not only has the PS in a shielded portion of the case, it is a linear power supply. Very heavy and a lot less efficient, but the +5V line is rock-stable on my scope.
As far as failure tolerance, while a single supply does provide a single point of failure most high-end systems use multiple hot-swap supplies. The network gear I help design uses four supplies but can run on two (or one with minimal blades). With more than one supply you get load-balancing, but it lets you hot-swap without a problem.
Now, the voltage drop is a good point. I've run +5V quite a ways, and frankly with multiple 16Ga wires the drop is not even worth talking about, even in a double wide rack (one machine of mine runs 5V through a four-wide rack with no significant drop). Mini- and mainframe computers do this all the time. While newer machines use a PS-per-cage, look in the back of a pdp11/60 or a big HP9000/800 sometime. There is a 5V supply you can spot weld with! If it is a design that powers a single cage, you can use sense lines to regulate the voltage at the backplane...although not in this case. Remember that 5V can wander +/- 250mV with no problem for many designs.
Now, with 3.3 volts the voltage drop could be a concern. I don't know how the rack companies do it; one way would be to put an "industrial brick" type DC/DC converter from 5-3.3V on each shelf. You could heatsink it to the shelf and not generate too much heat. I don't know how they deal with that issue in the commerical products.
Please people, don't chew out the guy unless you know what he is talking about. Many of you clearly don't; these are commerical products used in high-end computing applications. Remember that there are many products that really exist, even though you can't buy them at CompUSA.
This is a very good way of building a server. The ElCheapo power supplies that come in PC cases are not only inefficent, they fail frequently and take out everything downstream.
The single power supplies used in these racks are overengineered to the point that they don't fail in this manner. The racks usually have front and back doors and a blower...one we have where I work has a side-mounted McLean air conditioner! They are better cooled than any little PC box.
As far as manufactures go, I don't know. Some used to advertise in Linux Journal (which I have not taken for a couple years) but many of the designs used one-supply-per-rack, which I really don't care for. I like a single, decent-quality supply more than 24 crummy ones. I recall that 24-per-rack seemed to be the maximum you could fit into a rack...this would make each one about two rack units high, which seems to be the smallest I can imagine. They were using Intel all-in-one MBs with integrated 10/100 and VGA, with Black Box KBM switches. I would use the Belkin switches myself due to $$$.
Most seem to have rolling ventilated shelves with a IDE drive on each shelf.
Ours here at work has no identifing marks except "McLean" on the air conditioner, but as McLean is a famous rack blower manufacturer I would think they just made the air conditioner, not the rack. The rack was just shipped in from a company called "NetVantage" that we bought, so I can't tell you anything else.
I will check when I get home, I seem to recall I ordered brochures from some of those companies.
The athlon 450 doesn't exist
The Athlon must be the 500Mhz then. There is one K-6 and one Athlon and they are 50Mhz apart. They are on the same desk. I don't pay a lot of attention the the fine details of PC hardware; I'm tired of dealing with junk.
Find a clue
A "me too" comment so soon?
That was my point, to show that you just can't do a generic comparison based on the MHz of the CPU. The kiddie who posted the first message (the one to which I replied) claimed that any x86 chip would blow the UltraSparc away 'cause "its faster d00dz".
My comment is only relevant when read in context. It should not have been moderated to +4. But the mods are often clueless....
You might want to tone down your inflammatory language since you don't know anything about me. Just a hint.
I am working for a small company with needs for reasonable computing power and we just purchased a top-of-the-line HP workstation.
I am more of a hardware guy. This machine is very well built; not like the cheap 712/60's I've used in the past. The electrical and mechanical construction of this computer reminds me of the big iron machines of the past.
Plus, it's faster than hell.
So, from a hardware standpoint I love HP; it is sweet. But as others have remarked, HPUX is a pain to admin--at least for a HPUX newbie. I have admin'd Linux and BSD boxen for seven years, but HPUX is a new animal. If it ever gets even a tiny bit confused, it throws it's hands up in the air and vomits. This annoys me.
As far as SGI systems go, they have never struck me as very well built, even though I like them. A few years ago the University I was attending (getting a MS in EE) purchased what was then the fastest graphics computer in the world, a many (128?) processor SGI/Cray in nine racks. The dust level in our machine room (which was pretty high by computer room standards--it was installed by Univac for the 1108) caused it to fail within a month. Even with the dust problems fixed (the FS guys took the entire thing apart and vacuumed each board--plus replaced a lot) the machine has been a constant battle, and Irix issues cause it to reboot frequently.
An 1986 vintage HP9000 series 800 mini (the fastest box in '86) has been running nonstop in the same room for going on fifteen years with almost no downtime. Yeah, it's old and is more dust-tolerant, but it also seems better built.
Clearly I'm not an expert, but I would vote for the HP hardware over the SGI. Too bad you can't run your apps on the HP box with Linux installed...
Please! As the English language has evolved over the last 934 years, "men" has always been used to mean "humans".
Now suddenly in the last 20 years people have decided that this means "just males" and we have to change our entire manner of speaking to keep from offending somebody. Women should not be offended by this, and would not be had somebody not drilled it into them that this was a "slight" and therefore offensive.
Eventually we will have to say "Women, individuals-of-unknown-but-preferable-to-either-ma le-or-female-gender, transexuals, victims or
Turner's syndrome, and men".
Give me a break...
Am I missing something? What is the big deal?
Our NanoProbes are able to (benignly) penetrate a user's stealth firewall to verify the presence of the system hidden behind. Since our NanoProbes are able to bypass stealthing
He seems to claim that his packets cannot be blocked....watch me (or anyone else) block them. Seems high on ego, low on content.
Don't get me wrong, the Athlon is a great chip.
But a Sun UltraSparc 60 running 360Mhz machine completes a setiathome packet in four hours. An Athlon 450 "faster" machine completes the same packet in 20 hours.
Which is faster? (Hint 4
Let me get this straight...
You open the message by calling him an "idiot", and close it by calling him a "dumbass".
Sounds to me like you are mouthing off without thinking.
Your message is rude and in very poor form.
I work for a group that could actually use this box. But, alas, while we could afford to buy and power it, I don't think we could afford to pay for the setup costs.
And I know for a fact that I can't administer it!
Gads, I missed a whole paragraph of the article. Caffeine time.
As there is a US version, I can only assume that nobody will be importing this title so we'll all have to order them from amazon.co.uk or someplace. I hate buying an incomplete verion of anything.
One wonders who decided to make the cuts. Was it the US publisher, or did somebody in the UK decide that "those dumb yanks would never be able to understand this".
As an aside, an amusing British view of the US can be found at The American Adventure theme park (warning: site requires flash). Did you know all Americans speak with southern-Ozark accents?