Ok so huh... what about if I have used Windows, Corel Linux, SuSE, Slackware, Linux From Scratch, FreeBSD, NetBSD, Debian, OpenBSD, Gentoo and Ubuntu over past few years ??
This is Slashdot. The GP, I and probably a lot of other readerswho are not average users care about performance in 64-bit mode. See, for example, I write 64-bit assembly code optimized for AMD processors. So far I have never had the chance to evaluate a Core 2 CPU. So, like the GP, I would like to see 64-bit benchmarks of Core 2 CPUs. Is it so hard to understand ?
These benchmarks measure the number of processor cycles. Therefore the clock speed is mostly irrelevant. FYI it is rather common to count cycles when speaking about the efficiency of crypto/hashing algorithms.
Every time I hear someone like you complaining that his voice wasn't being heard, or that his edits have been reversed without valid reason, I ask him to show me the article(s) where this has happened, so I can check the history log myself and see what really happened.
And you know what ? I never get an answer...
This tends to make me think that such complains about wikipedia are generally exaggerated and/or groundless. In other words, the edits have probably been reversed for a valid reason.
So, Metasquares, show us the articles as well as the history log entries corresponding to your edits, so that we can all see whether what you say is true or not.
It is true that Dell has been rumoring this for years. But today it's different: Dell has promised to their customers and shareholders that they will start selling AMD-based desktops in September 2006, and AMD-based servers by the end of 2006, as you can read on one of the many dell.com's webpages about the subject.
You can't say "it's still a rumor". This is wrong. If they don't do it now, they will get severly sanctionned by their shareholders this time.
simply because formatting it would take almost 18 hours.
Wrong. For most recent Linux filesystems, the formatting step is actually very fast (notable exceptions are: ext2, ext3.) See XFS. Formatting a 100,000 MB partition requires writing only 50 MB of data.
Ok. For the year 2006, the total amount of Uranium required by all world nuclear power reactors combined will be about
65,000 tons, assuming a density of 19.1 g.cm^-3, this is equivalent to a cubic container measuring 40 feet x 40 feet x 40 feet (40 feet =~ 15 meters). Of course Uranium waste is never stored like that in huge containers but even assuming it is stored while filling only 10% of the raw volume of a container, it is still much better than releasing millions of tons of polluting gases in the atmosphere, better than polluting rivers and seas, etc.
Nuclear? Too much radioactive waste. Yes I realize we've made some vast improvements in nuclear tech.
Do you have any idea of the volume of radioactive waste produced by the whole world over a 1-year period ? It is much smaller than what you think. Search for it and I'll promise you will change your mind after knowing it. No I won't give you the response here, I want to make you change your mind by yourself;-)
Very nicely done. But if I were you I would have done it slightly differently. I would not have used the internal disk bay (faster interventions, no need to open the case when replacing a disk). I would have only put 12 disks in the hot-swap bays (11 for RAID, and 1 hot spare). I would have used software RAID (hw and sw RAID are both capable of saturating a GigE pipe --I suppose you were doing backups over the network). Software RAID would have allowed me to create a small 100 MB RAID1 partition for booting (/boot), a 1-100 GB RAID5 partition for your root fs (/), and the remaining space for your backups (/space). This way you would have ended up spending less money (12 HDDs instead of 13 HDDs, no hardware RAID card), having more usable space for your backups (the equivalent of 10 HDDs minus 1-100 GB for / --compared to only 9 HDDs in your case) and having easier maintenance procedures (no case opening when replacing any disk).
Exactly ! This is what I wanted you to say. They can afford it in part *because* they maintain their infrastructure
internally, don't buy expensive hardware/parts and save money. Look at my company B example, they have $40k left to buy backup servers, spare parts, etc.
What your analysis also fails to take into account is any sort of failure.
I agree that correctly handling failure is a very important point in order to build a reliable infrastructure out of cheap hardware. It requires time & effort. But IMHO there are plenty of cases where people say it is very hard where in fact it's perfectly doable, and would be cheaper in the long run. I think this way of designing complex IT architectures is going to explode in the near future (10 years). This is also why the debate is so intense in this thread. This is a new way of doing things, this scares people who were used to buy trusted/expensive servers up to now. Google is the clear pioneer in this field.
What technicians do you know who can re-work a motherboard ?
Not many. But look at what Google would do: trash it and buy a new one. This is the advantage of commodity parts, just identify the failing component (mobo, memory, cpu, NIC, switch...) and replace it.
Not to mention, but why would they buy more sun boxes when they hold more data than the smaller vendors (16TB vs 24TB source: vendors), take up less space, use less power per GB and so on. The sun boxes are more efficient (a fucking HUGE reason for an enterprise that would buy dozens or even hundreds of these).
No. All you need is a 48-bay 4U chassis similar to Sun's to match their storage density. So technically speaking a high-end Sun server would have the same storage density and watt/GB consumption than a DIY server.
[Google has] by some estimates in excess of 50k computers.
Actually the estimates are between 100k and 500k. But my point still holds: when they started, with only 100 machines, 1k, 10k, etc they always kept using commodity parts, no matter how large their datacenters were.
An expensive warranty that smaller vendors COULD provide (but most DON'T,
my point was to show the price difference).
And they don't because clients who buy hardware from small vendors
know that even the most advanced and expensive vendor warranties
and technical support contracts cannot match the efficiency, cost and
speed of interventions performed by technicians hired by a company to maintain
its own infrastructure.
Real world examples: UPS don't use external services
to fix their vehicles, they hire mechanics and run their own repair shops.
Google don't buy support contracts from Sun/HP/Dell, they hire technicians
and engineers and maintain their own servers.
Imaginary example: company A buys 10 of those Sun servers with a 2-year
warranty (so $20k each),
company B buys 5 cheap servers with twice the storage capacity ($16k each)
and hires a technician for 2 years ($40k/year). Over 2 years company B
has saved $40k (do the math), plus it has the advantage of having the
technician on-site (faster interventions than Sun), and he would be
able to maintain other servers at no additional cost for you
(a full-time technician maintaining only 5 servers would have a lot of spare time).
Is it so hard to understand ? Profitable companies prove that it is
the right way to do it (UPS, Google...).
If you think that an on-site 2-year warranty on a 4U box and a
top-of-the-line chassis are worth Sun's price ($33k, or $20k assuming a 40%
discount), then you are gullible and uninformed of the real value of such
things. Proof ? This
guy bought a similar server from a smaller vendor, with twice the
storage capacity at half the price of Sun. Guess that you could throw
in any expensive warranty and still be well below Sun's prices.
The disks would go in the chassis (see my itemized list). You may not
know it but Sun is not the 1st company to use a chassis with vertical bays.
Here is one example
among many. The price would be more likely around 2 or 3 grands by the way,
instead of 1 grand. But anyway this doesn't change the fact
that this Sun box is way overpriced, even with a good 40% discount.
Regarding the mobo, just pick one with two AMD 8131 or 8132 PCI-X
bridges. This will give you 4 independent PCI-X busses. The two PCI-X
bridges would have to be on 2 different HT links in order to not
dangerously approach the theoretical one-way data throughput limit of 3.2
GB/s of one 1600 MT/s 16 bits HT link. The two PCI-X bridges could be
either connected to different CPUs or to the same CPU because the Opteron
XBAR _can_ easily handle the ~3 GB/s you speak about,
it has been designed to support 19.2 GB/s of HT traffic and even more with the recent
upgrade to 2000 MT/s ccHT links. Now with the 4 independent PCI-X
busses, you could put 4 SATA HBAs on the 1st and 2nd busses, and 2 HBAs on
the 3rd and 4th busses. This way the first 2 busses will run at 100 MHz
and the 2 others will run at 133 MHz, giving a practical throughput of
3.4 GB/s (2 * (100 MHz * 64 bits / 8) + 2 * (133 MHz * 64 bits / 8),
and assuming a 90% efficiency as found on most PCI/PCI-X busses), this is
enough to handle the 3 GB/s you are speaking about. There are plenty of
single AMD 8131 mobo on the market right now starting at $250. I am sure
you can find one with two AMD 813x for $500 max.
Now when I think about it you could even use SATA port multipliers in
order to reduce the number of HBAs, allowing all busses to run at 133 MHz.
I am aware of 12-port and 24-port SATA HBAs (Areca comes to mind) but those
are outrageously expensive and are not necessary to handle all that
throughput. My experience and those of my friends playing with high-end
enterprise gear prove that _very_ simple and inexpensive PCI-X SATA chips
such as the SII3124 or Marvell 88SXxxxx are way sufficient to handle the
max combined read throughput of any number of disks attached to their SATA
ports. The reason being that the designers of such chips have come up with
a simple and performant hardware interface optimized to reduce the CPU
load. I know for a fact that the SII3124 design is somewhat close to the
AHCI spec which is the best example of a performant SATA hw interface.
So I _do_ believe that it is possible to build a $13-14k server with 48
SATA disks in 4U offering ~3 GB/s of raw read throughput. I don't
understand why so many people refuse to believe that, especially since other
posters in this thread have pointed out that some vendors are already
selling similarly priced servers !
Those who work 70-hour weeks for the first 6-8 years
of their professional life, building custom-designed IT infrastructures
out of cheap hardware parts, designing reliable software assuming
servers will fail, saving tons of money for their company,
accomplishing projects everyone else thought were impossible before,
and then retiring in their early 30s, rich and most importantly
ready to fully enjoy the
rest of their life with friends & family members.
Those who work 9-to-5 in a cubicle, wasting the company's money
by making poor decisions and buying expensive servers from common
hardware vendors,
constantly looking for the least amount of technical responsibilities,
blaming everybody else for the failure
(or lack of success) of their company, and finally retiring in
their 60s, tired by years of a boring life.
I choosed to belong to 1st one.
Sorry to hear you saying I made a crazy choice.
The 12 TB config is sold at $33k, or $2.75/GB, but assembling such a server yourself
is possible and can be done today for 1/3rd of this price:
1 x dual-Opteron mobo = 1 x $500
2 x Opteron 285 = 2 x $1100
8 x 2 GB DDR400 registred DIMM = 8 x $300
6 x 8-port PCI-X Marvell SATA card = 6 x $100
48 x 250 GB 7.2kRPM SATA disks = 48 x $110
1 x Chassis+PSU+Rails = 1 x $1000
Total = $11980 or $1.00/GB
(I have actually slightly overestimated the above prices.)
Of course people are going to say that such a server
is not be as reliable as a Sun server, that it does
not come with technical support, etc. But in most cases
such arguments are invalid because you save so much money
that you can afford assembling/maintaining the server
and replacing faulty hardware parts yourself. Time is money,
but by having saved money you can now afford time;-)
The living proof that such a model would be successful
is Google: instead of buying Sun servers like most startups in their
time, they built their servers themselves to save money.
Haha ! Welcome to the world of Phreaking... You might not know it but the telephone network is as easily hackable, vulnerable and exploitable as the Internet is today. Good luck tracing the bad guy who impersonated your credit card company you supposedly called on 1-800-XXX-YYYY, when he might have penetrated voicemail systems, set up temporary forwarding, hacked telephone switches, etc...
This is why he got caught, because he did not care about the safeness of his delivery
procedure (whatever the reason was, laziness, etc).
For every man like him in jail, there are hundreds of others who are more
careful, who use safe drop mailboxes, who are currently enjoying their scaming
activities. And a lot of those people will probably never get caught.
Just to recap things, the Xeon 5100-series, aka "Woodcrest", is the very first released processor
family that is based on the new 8th generation, Intel Core Microarchitecture, technically inspired
from the 6th generation (PPro, PII, PIII), instead of the 7th generation (P4). As a side note, Intel
has been using the "Core Solo" and "Core Duo" denominations for some processors but this is just
a marketing usage of the term "Core", because such processors are NOT based on the Intel Core
Microarchitecture. Anyway, Woodcrest is the first to represent this all-new Intel Core Microarchitecture
that is supposed to save Intel from the very competitive K8 design (Opteron, Athlon64...).
So, Woodcrest seems indeed to be a very good processor, as shown in this
preview (the less-biased, more technically accurate I have been able
to find up to this day). Intel claims that Woodcrest is "80% more performant
at 35% less power" compared to the original dual-core Xeon processor, and most
benchmarks seem to confirm this claim. It may seem technically impressive, but
in fact considering the very poor design of the original dual-core Xeon processor,
such an improvement HAD to be expected and was almost a prerequisite for Intel
to even start thinking about taking back Opteron's market share.
Here is a quick fact list I have assembled from my own research and from the review linked above:
At equal clock frequencies, Woodcrest is about 5-15% more powerful than Opteron on traditional workloads (common x86 and arithmetic instructions), and much more powerful (30% and more) than Opteron on multimedia workloads (mostly SSE, SSE2, maybe FPU I am not sure).
At equal clock frequencies, Opteron is still much more powerful (30% and more) than Woodcrest on memory-intensive workloads due to its integrated memory controller (leading to better latency) and ccHT links in SMP cases (where memory throughput increases with the number of ccHT links).
At equal clock frequencies, Woodcrest consumes less power than Opteron, but Woodcrest's memory (FB-DIMM) requires more power than Opteron's memory (DDR400). So overall, a Woodcrest-based system consumes about as much power as an Opteron-based system (as shown in page 3 of the review).
At equal clock frequencies, Woodcrest is cheaper than Opteron, but Woodcrest motherboards (socket 771) are more expensive than Opteron motherboards (socket 939 and 940) and FB-DIMM memory is twice the price of DDR400. These pricing differences are so large that Opteron is still preferable to Woodcrest in most cases: Opteron is cheaper for any single or dual-cpu server config with 4 GB or more of memory, Opteron is cheaper for any entry-level server config (about $1500 and below) whatever the amount of memory is, Woodcrest seems to only make sense when the high-end processors (Xeon 5140, 5150 and 5160) are used with NO MORE than 4 GB of memory (else Opteron's cheaper memory has a price advantage).
Of course, in the high-end server market (4, 8 or more processors), Opteron is still the clear technical leader because Intel STILL hasn't switched to a CPU interconnect similar to HT and STILL isn't using an integrated memory controller.
In conclusion, I would say that when comparing only the processors,
Woodcrest is superior to Opteron in many aspects (such as instruction
throughput), and Opteron beats Woodcrest in other aspects
(such as memory accesses). But when comparing a whole Woodcrest-based
system versus an Opteron-based system, other factors come into play
(such as price and scalibility), which make Opteron superior to
Woodcrest in a lot of cases.
Slashdot Mirror
Ok so huh... what about if I have used Windows, Corel Linux, SuSE, Slackware, Linux From Scratch, FreeBSD, NetBSD, Debian, OpenBSD, Gentoo and Ubuntu over past few years ??
Even faster !
This is Slashdot. The GP, I and probably a lot of other readers who are not average users care about performance in 64-bit mode. See, for example, I write 64-bit assembly code optimized for AMD processors. So far I have never had the chance to evaluate a Core 2 CPU. So, like the GP, I would like to see 64-bit benchmarks of Core 2 CPUs. Is it so hard to understand ?
These benchmarks measure the number of processor cycles. Therefore the clock speed is mostly irrelevant. FYI it is rather common to count cycles when speaking about the efficiency of crypto/hashing algorithms.
Remind me of an old joke...
Windows 95: comes with built-in support for long filena~1.
Ok, sorry about that, I misunderstood your original post. I thought you were speaking about article edits.
Every time I hear someone like you complaining that his voice wasn't being heard, or that his edits have been reversed without valid reason, I ask him to show me the article(s) where this has happened, so I can check the history log myself and see what really happened. And you know what ? I never get an answer...
This tends to make me think that such complains about wikipedia are generally exaggerated and/or groundless. In other words, the edits have probably been reversed for a valid reason.
So, Metasquares, show us the articles as well as the history log entries corresponding to your edits, so that we can all see whether what you say is true or not.
It is true that Dell has been rumoring this for years. But today it's different: Dell has promised to their customers and shareholders that they will start selling AMD-based desktops in September 2006, and AMD-based servers by the end of 2006, as you can read on one of the many dell.com's webpages about the subject.
You can't say "it's still a rumor". This is wrong. If they don't do it now, they will get severly sanctionned by their shareholders this time.
Wrong. For most recent Linux filesystems, the formatting step is actually very fast (notable exceptions are: ext2, ext3.) See XFS. Formatting a 100,000 MB partition requires writing only 50 MB of data.
Ok. For the year 2006, the total amount of Uranium required by all world nuclear power reactors combined will be about 65,000 tons, assuming a density of 19.1 g.cm^-3, this is equivalent to a cubic container measuring 40 feet x 40 feet x 40 feet (40 feet =~ 15 meters). Of course Uranium waste is never stored like that in huge containers but even assuming it is stored while filling only 10% of the raw volume of a container, it is still much better than releasing millions of tons of polluting gases in the atmosphere, better than polluting rivers and seas, etc.
Do you have any idea of the volume of radioactive waste produced by the whole world over a 1-year period ? It is much smaller than what you think. Search for it and I'll promise you will change your mind after knowing it. No I won't give you the response here, I want to make you change your mind by yourself ;-)
I know you are probably joking, but just to clarify, this country is France.
The lizard probably figured that the patent would have expired during those past 225 million years anyway.
Has already happened in my home country, which generates 79% of its energy in nuclear power plants. Now can I get my electric car ? ;-)
Very nicely done. But if I were you I would have done it slightly differently. I would not have used the internal disk bay (faster interventions, no need to open the case when replacing a disk). I would have only put 12 disks in the hot-swap bays (11 for RAID, and 1 hot spare). I would have used software RAID (hw and sw RAID are both capable of saturating a GigE pipe --I suppose you were doing backups over the network). Software RAID would have allowed me to create a small 100 MB RAID1 partition for booting (/boot), a 1-100 GB RAID5 partition for your root fs (/), and the remaining space for your backups (/space). This way you would have ended up spending less money (12 HDDs instead of 13 HDDs, no hardware RAID card), having more usable space for your backups (the equivalent of 10 HDDs minus 1-100 GB for / --compared to only 9 HDDs in your case) and having easier maintenance procedures (no case opening when replacing any disk).
Exactly ! This is what I wanted you to say. They can afford it in part *because* they maintain their infrastructure internally, don't buy expensive hardware/parts and save money. Look at my company B example, they have $40k left to buy backup servers, spare parts, etc.
I agree that correctly handling failure is a very important point in order to build a reliable infrastructure out of cheap hardware. It requires time & effort. But IMHO there are plenty of cases where people say it is very hard where in fact it's perfectly doable, and would be cheaper in the long run. I think this way of designing complex IT architectures is going to explode in the near future (10 years). This is also why the debate is so intense in this thread. This is a new way of doing things, this scares people who were used to buy trusted/expensive servers up to now. Google is the clear pioneer in this field.
Not many. But look at what Google would do: trash it and buy a new one. This is the advantage of commodity parts, just identify the failing component (mobo, memory, cpu, NIC, switch...) and replace it.
No. All you need is a 48-bay 4U chassis similar to Sun's to match their storage density. So technically speaking a high-end Sun server would have the same storage density and watt/GB consumption than a DIY server.
Actually the estimates are between 100k and 500k. But my point still holds: when they started, with only 100 machines, 1k, 10k, etc they always kept using commodity parts, no matter how large their datacenters were.
An expensive warranty that smaller vendors COULD provide (but most DON'T, my point was to show the price difference). And they don't because clients who buy hardware from small vendors know that even the most advanced and expensive vendor warranties and technical support contracts cannot match the efficiency, cost and speed of interventions performed by technicians hired by a company to maintain its own infrastructure.
Real world examples: UPS don't use external services to fix their vehicles, they hire mechanics and run their own repair shops. Google don't buy support contracts from Sun/HP/Dell, they hire technicians and engineers and maintain their own servers.
Imaginary example: company A buys 10 of those Sun servers with a 2-year warranty (so $20k each), company B buys 5 cheap servers with twice the storage capacity ($16k each) and hires a technician for 2 years ($40k/year). Over 2 years company B has saved $40k (do the math), plus it has the advantage of having the technician on-site (faster interventions than Sun), and he would be able to maintain other servers at no additional cost for you (a full-time technician maintaining only 5 servers would have a lot of spare time).
Is it so hard to understand ? Profitable companies prove that it is the right way to do it (UPS, Google...).
If you think that an on-site 2-year warranty on a 4U box and a top-of-the-line chassis are worth Sun's price ($33k, or $20k assuming a 40% discount), then you are gullible and uninformed of the real value of such things. Proof ? This guy bought a similar server from a smaller vendor, with twice the storage capacity at half the price of Sun. Guess that you could throw in any expensive warranty and still be well below Sun's prices.
The disks would go in the chassis (see my itemized list). You may not know it but Sun is not the 1st company to use a chassis with vertical bays. Here is one example among many. The price would be more likely around 2 or 3 grands by the way, instead of 1 grand. But anyway this doesn't change the fact that this Sun box is way overpriced, even with a good 40% discount.
Regarding the mobo, just pick one with two AMD 8131 or 8132 PCI-X bridges. This will give you 4 independent PCI-X busses. The two PCI-X bridges would have to be on 2 different HT links in order to not dangerously approach the theoretical one-way data throughput limit of 3.2 GB/s of one 1600 MT/s 16 bits HT link. The two PCI-X bridges could be either connected to different CPUs or to the same CPU because the Opteron XBAR _can_ easily handle the ~3 GB/s you speak about, it has been designed to support 19.2 GB/s of HT traffic and even more with the recent upgrade to 2000 MT/s ccHT links. Now with the 4 independent PCI-X busses, you could put 4 SATA HBAs on the 1st and 2nd busses, and 2 HBAs on the 3rd and 4th busses. This way the first 2 busses will run at 100 MHz and the 2 others will run at 133 MHz, giving a practical throughput of 3.4 GB/s (2 * (100 MHz * 64 bits / 8) + 2 * (133 MHz * 64 bits / 8), and assuming a 90% efficiency as found on most PCI/PCI-X busses), this is enough to handle the 3 GB/s you are speaking about. There are plenty of single AMD 8131 mobo on the market right now starting at $250. I am sure you can find one with two AMD 813x for $500 max.
Now when I think about it you could even use SATA port multipliers in order to reduce the number of HBAs, allowing all busses to run at 133 MHz. I am aware of 12-port and 24-port SATA HBAs (Areca comes to mind) but those are outrageously expensive and are not necessary to handle all that throughput. My experience and those of my friends playing with high-end enterprise gear prove that _very_ simple and inexpensive PCI-X SATA chips such as the SII3124 or Marvell 88SXxxxx are way sufficient to handle the max combined read throughput of any number of disks attached to their SATA ports. The reason being that the designers of such chips have come up with a simple and performant hardware interface optimized to reduce the CPU load. I know for a fact that the SII3124 design is somewhat close to the AHCI spec which is the best example of a performant SATA hw interface.
So I _do_ believe that it is possible to build a $13-14k server with 48 SATA disks in 4U offering ~3 GB/s of raw read throughput. I don't understand why so many people refuse to believe that, especially since other posters in this thread have pointed out that some vendors are already selling similarly priced servers !
There are 2 kinds of IT engineers:
I choosed to belong to 1st one. Sorry to hear you saying I made a crazy choice.
The 12 TB config is sold at $33k, or $2.75/GB, but assembling such a server yourself is possible and can be done today for 1/3rd of this price:
(I have actually slightly overestimated the above prices.) Of course people are going to say that such a server is not be as reliable as a Sun server, that it does not come with technical support, etc. But in most cases such arguments are invalid because you save so much money that you can afford assembling/maintaining the server and replacing faulty hardware parts yourself. Time is money, but by having saved money you can now afford time ;-)
The living proof that such a model would be successful
is Google: instead of buying Sun servers like most startups in their
time, they built their servers themselves to save money.
Haha ! Welcome to the world of Phreaking... You might not know it but the telephone network is as easily hackable, vulnerable and exploitable as the Internet is today. Good luck tracing the bad guy who impersonated your credit card company you supposedly called on 1-800-XXX-YYYY, when he might have penetrated voicemail systems, set up temporary forwarding, hacked telephone switches, etc...
This is why he got caught, because he did not care about the safeness of his delivery procedure (whatever the reason was, laziness, etc). For every man like him in jail, there are hundreds of others who are more careful, who use safe drop mailboxes, who are currently enjoying their scaming activities. And a lot of those people will probably never get caught.
Just to recap things, the Xeon 5100-series, aka "Woodcrest", is the very first released processor family that is based on the new 8th generation, Intel Core Microarchitecture, technically inspired from the 6th generation (PPro, PII, PIII), instead of the 7th generation (P4). As a side note, Intel has been using the "Core Solo" and "Core Duo" denominations for some processors but this is just a marketing usage of the term "Core", because such processors are NOT based on the Intel Core Microarchitecture. Anyway, Woodcrest is the first to represent this all-new Intel Core Microarchitecture that is supposed to save Intel from the very competitive K8 design (Opteron, Athlon64...).
So, Woodcrest seems indeed to be a very good processor, as shown in this preview (the less-biased, more technically accurate I have been able to find up to this day). Intel claims that Woodcrest is "80% more performant at 35% less power" compared to the original dual-core Xeon processor, and most benchmarks seem to confirm this claim. It may seem technically impressive, but in fact considering the very poor design of the original dual-core Xeon processor, such an improvement HAD to be expected and was almost a prerequisite for Intel to even start thinking about taking back Opteron's market share.
Here is a quick fact list I have assembled from my own research and from the review linked above:
At equal clock frequencies, Woodcrest is about 5-15% more powerful than Opteron on traditional workloads (common x86 and arithmetic instructions), and much more powerful (30% and more) than Opteron on multimedia workloads (mostly SSE, SSE2, maybe FPU I am not sure).
At equal clock frequencies, Opteron is still much more powerful (30% and more) than Woodcrest on memory-intensive workloads due to its integrated memory controller (leading to better latency) and ccHT links in SMP cases (where memory throughput increases with the number of ccHT links).
At equal clock frequencies, Woodcrest consumes less power than Opteron, but Woodcrest's memory (FB-DIMM) requires more power than Opteron's memory (DDR400). So overall, a Woodcrest-based system consumes about as much power as an Opteron-based system (as shown in page 3 of the review).
At equal clock frequencies, Woodcrest is cheaper than Opteron, but Woodcrest motherboards (socket 771) are more expensive than Opteron motherboards (socket 939 and 940) and FB-DIMM memory is twice the price of DDR400. These pricing differences are so large that Opteron is still preferable to Woodcrest in most cases: Opteron is cheaper for any single or dual-cpu server config with 4 GB or more of memory, Opteron is cheaper for any entry-level server config (about $1500 and below) whatever the amount of memory is, Woodcrest seems to only make sense when the high-end processors (Xeon 5140, 5150 and 5160) are used with NO MORE than 4 GB of memory (else Opteron's cheaper memory has a price advantage).
Of course, in the high-end server market (4, 8 or more processors), Opteron is still the clear technical leader because Intel STILL hasn't switched to a CPU interconnect similar to HT and STILL isn't using an integrated memory controller.
In conclusion, I would say that when comparing only the processors, Woodcrest is superior to Opteron in many aspects (such as instruction throughput), and Opteron beats Woodcrest in other aspects (such as memory accesses). But when comparing a whole Woodcrest-based system versus an Opteron-based system, other factors come into play (such as price and scalibility), which make Opteron superior to Woodcrest in a lot of cases.
Are we talking about the corona or Corona ? Because ejections of the second one are disgusting.