Fudzilla: Bulldozer performance figures are in

[RussianSensation]

A very tiny segment of AMD's or Intel's clients.

That may be true but in marketing, a LOT has to do with perception. You know the reason Ferrari participates in Formula 1 or Audi in LeMans? Being inadequate for gamers means these same users tell all of their friends and parents that Intel makes "better chips". For the average Joes, Intel has the perception of making better CPUs, and in the last 5 years, it was pretty much true, exacerbating the image problem for AMD.

XP+, A64 and X2 were all great gaming CPUs. Phenom I/II were not. So history has shown us that CPUs that excel in gaming, also tend to do well in common office tasks and generally have superior performance / watt and IPC. If we examine Pentium 4 and Pentium D, when Intel didn't focus on IPC, performance / watt, performance per core, then we have a situation where a CPU that didn't follow those criteria also didn't do as well in the marketplace.

So contrary to your view that IPC, performance / watt and performance per core doesn't matter, all 3 AMD generations that excelled at these metrics sold better than Phenom I/II did, while when Intel didn't follow those 3 metrics, their own CPUs were losing market share.

Let me guess, you have never been to Best Buy and have never observed people going for laptops ... right? The vast majority of these people couldn't tell the difference between a keyboard and a mouse and you think they care for 'performance', 'mobility' or other silly things like that? Just between you and me, don't tell anyone: they care about green pieces of paper with dead presidents on them, specifically the lack of these when shopping for laptops. Some are also looking for 'cool' when they buy products, specifically Apple 'cool'.

Apple has shown that many users are willing to pay for the experience. They don't necessarily care about specs on paper as long as the product delivers - Intel CPUs do deliver.

Notice why Apple carries Intel processors? (1) Better performance / watt (2) Better efficiency --> better battery life (3) Faster performance for most tasks a typical user cares about (4) Ability to meet manufacturing demand (but another reason why Intel has more manufacturing capacity is because there is more demand for its products).

Most people would agree that it's much easier to market a great product (i.e., Intel CPUs) than to have great marketing for a mediocre product (i.e., AMD's Phenom I/II). As a side note Centrino was a brilliant marketing move by Intel because when the consumer got the product, it was superior to AMD's offerings. AMD couldn't do anything of the sort because to claim that their CPU was superior in performance / watt or battery life would have simply been a misrepresentation of the truth.

Notice how the Fusion is taking off for AMD, because it really is the better gaming platform than HD3000. Great product doesn't require revolutionary marketing because it almost sells itself.

How do you explain Apple then? They never made Macs even close to performance to HP/Dell PCs/laptops yet their sales are going up and up and up while everyone else sinks.

What are you talking about? Apple really came into its own when they switched to using Core 2 Duo/ Core i3/i5/i7 and Sandy Bridge processors. The company's focus is on user experience, which has A LOT to do with mobility / portability and that is partially related to excellent battery life of their products (esp. laptops). Steve Jobs ditched PowerPC chips for Intel since Intel offered far superior performance / watt. You are telling me no one cares about performance / watt, IPC and power efficiency? I got news for you -- manufacturers and designers who strive to create a killer mobile device/laptops/tablet care because it's a lot easier to sell "10 hour battery life notebook" with a fast processor than to slap a 2 lbs of protruding battery to a laptop and try to sell that as a premium product. Apple didn't choose AMD because AMD's current architecture doesn't meet any of the criteria that Apple wants in order to design a successful product.

Another sign that Apple does care about efficiency and performance / watt is illustrated by them abandoning NV for AMD this generation for GPUs. Lower efficiency implies either slower performance or more heat, which are undesirable compromises. Are you are still going to deny "no one cares about performance / watt" even after Intel adopted this as their key design strategy for implementing new features into their CPUs? :sneaky:

Could it be that the majority of consumer don't know/care about performance if a computing product offers good speed already? Apple is the perfect example where computer speed is irrelevant as long as the machine can do what most people need (read: they don't care about enthusiasts). And even a 4 year old Phenom II or Core 2 Duo or Quad can do that. So why do these people need a super duper Bulldozer again?

If your logic was true, no one would even bother upgrading to faster chips. Laptops would only add features, better aesthetics and slimmer form factors. According to you, Intel could have easily produced Core 2 Duo since 2006 and not done a single thing to increase its performance / watt. Yes, they did exactly that...oh wait.

Oh gosh, someone get on the batphone and call AMD's new CEO and warn them of their mistakes. I mean, AMD focusing on the professional and server market instead of the continually dwindling 'enthusiast' market ... why do they do that?!

Except that more cores alone isn't sufficient to win the server market share. You need an extremely power efficient processor. Throwing more cores won't solve the problem. AMD threw a 12-core Magny Cours against a 6-core Xeon and it didn't do squat for them. But you can go ahead and deny that IPC, Performance / watt, Performance per core don't matter.

You can be sure that when Bulldozer launches, its power efficiency will be scrutinized by many people in the server and consumer markets because it directly impacts operational costs and every day usability for mobile devices.

Critical to whom? You? Part of a minuscule market segment?

Pretty much to anyone who understands where the future of computing is heading - mobile, tablet, smartphone space where (1) performance / watt (2) IPC (3) performance per core is critical. I think only you are being in denial.

Are you insinuating that if Bulldozer was simply an 8 core Phenom II, AMD would be back on track? We know it wouldn't be enough, therefore, successes in the 3 areas I mentioned are critical to win back performance.

Just a silly question: what if AMD doesn't care about recapturing its 'glory'.

You obviously haven't met any engineers who design CPUs then. They are very passionate and want their firm to succeed.

What if all they want to make processors that their clients like and want to buy? What if AMD doesn't give a rat's ass about some 'enthusiasts' on some random interwebs forum? You are deluding yourself if you think that the 'enthusiast' market is all that. It's nothing more than a tiny drop in the bucket for either Intel or AMD.

I never said that enthusiast market is the only one that matters. But somehow you keep missing the point of what is necessary for a modern processor to be successful - Excellent power efficiency, great performance / watt, strong IPC and more cores. If you only focus on more cores, you will fail miserably.

 

[kevinsbane]

no, he was saying that a strong salt makes a weak password strong even from brute force attacks. Yes, salt kills rainbow tables, but it doesn't affect brute force/dictionary attacks.

I think this may stem from a differing understanding of the term "salt" here, but why does salting a password not affect a dictionary attack?

I mean, I understand the concept of a dictionary attack against a password, but if I salt my password, even if it is a weak password to begin with, then how does a dictionary attack work against a salted, "weak" password?

For example, I take my weak password:

another1

That is, a weakly salted password, easily attacked via a dictionary attack. I add a salt to my password that is unknown to an attacker. The salt is

"************"

The above consists (maybe) of letters, numbers, symbols, upper and lower case and is of an unknown length and is combined with the weak password in an unknown way. I don't see how a dictionary attack would work against this.

Unless you are speaking of the the *system* salting whatever the user actually uses as a password to change the hash. Since the system must be able to perform the salting functions, it must store that salt (and the functions involving that salt) and thus you can access the salt for yourself.

 

[Voo]

I add a salt to my password that is unknown to an attacker.

And as at least three people have several times already said, the whole idea of a salt is that it isn't secret.. the whole concept wouldn't technically work if it were different.

 

[Tuna-Fish]

That is, a weakly salted password, easily attacked via a dictionary attack. I add a salt to my password that is unknown to an attacker.

Salts are by definition available to the attacker with physical access, because the machine must be able to decrypt the data when given a correct password. If something is not available it is not a salt. And depending on how you intend to use it, it would either be useless to security or impossible to implement.

 

[Abwx]

Another punk bench...

The more he post, the more IPC increase...:biggrin:

 

[kevinsbane]

And as at least three people have several times already said, the whole idea of a salt is that it isn't secret.. the whole concept wouldn't technically work if it were different.

Ok, then we are understanding the term differently. Maybe I was thinking of padding, instead?

Salting as part of the encryption process to ensure a hard to crack hash, as opposed to adding data to a password as part of the password to be entered itself. IE, another1 is what I actually enter in my password field; it is then salted and hashed by the system. Contrast to me using another1ADGjl' as what I actually enter into my password field, which is then salted and hashed by the system.

I was just wondering since since Cogman said this, and I couldn't understand why adding a salt to a weak password would not make it a strong password. If an end-user were to "salt"/pad his password in the way I describe in my previous post, then he would definitely be able to transform a weak password into a strong password - conditional on him choosing a decent "salt"/pad that is not trivial.

 

[sandorski]

Y'all arteries are hardening...

 

[Voo]

Ok, then we are understanding the term differently. Maybe I was thinking of padding, instead?

No "padding" in the context of modern cryptography makes only sense when looking at block ciphers (basically you can encrypt data only in blocks of size N, so if the data isn't a multiple of N you have to pad it).

A hash's primary goal is to make rainbow tables infeasible, by increasing the size of those by 2^N (N is the bitsize of the hash). That's especially important for server databases where otherwise you could hash one word and compare it with the complete database to find matches (also any user with the same password would have the same hash, which is problematic as well).

So there's no reason for the salt to be secret in the first place, since it fulfills its goal just as well without it. Also there are some much more basic problems: If the "salt" is secret it's semantically part of the password and has to be provided by the user or someone else - now that obviously makes it completely useless for local or server side encryption. So basically the local salt is stored at a well known, documented location of the hard drive (because the client software - say truecrypt - needs it as well to encrypt the drive) and for servers it's stored in plaintext besides the password.

 

[frozentundra123456]

That may be true but in marketing, a LOT has to do with perception. You know the reason Ferrari participates in Formula 1 or Audi in LeMans? Being inadequate for gamers means these same users tell all of their friends and parents that Intel makes "better chips". For the average Joes, Intel has the perception of making better CPUs, and in the last 5 years, it was pretty much true, exacerbating the image problem for AMD.

XP+, A64 and X2 were all great gaming CPUs. Phenom I/II were not. So history has shown us that CPUs that excel in gaming, also tend to do well in common office tasks and generally have superior performance / watt and IPC. If we examine Pentium 4 and Pentium D, when Intel didn't focus on IPC, performance / watt, performance per core, then we have a situation where a CPU that didn't follow those criteria also didn't do as well in the marketplace.

So contrary to your view that IPC, performance / watt and performance per core doesn't matter, all 3 AMD generations that excelled at these metrics sold better than Phenom I/II did, while when Intel didn't follow those 3 metrics, their own CPUs were losing market share.



Apple has shown that many users are willing to pay for the experience. They don't necessarily care about specs on paper as long as the product delivers - Intel CPUs do deliver.

Notice why Apple carries Intel processors? (1) Better performance / watt (2) Better efficiency --> better battery life (3) Faster performance for most tasks a typical user cares about (4) Ability to meet manufacturing demand (but another reason why Intel has more manufacturing capacity is because there is more demand for its products).

Most people would agree that it's much easier to market a great product (i.e., Intel CPUs) than to have great marketing for a mediocre product (i.e., AMD's Phenom I/II). As a side note Centrino was a brilliant marketing move by Intel because when the consumer got the product, it was superior to AMD's offerings. AMD couldn't do anything of the sort because to claim that their CPU was superior in performance / watt or battery life would have simply been a misrepresentation of the truth.

Notice how the Fusion is taking off for AMD, because it really is the better gaming platform than HD3000. Great product doesn't require revolutionary marketing because it almost sells itself.



What are you talking about? Apple really came into its own when they switched to using Core 2 Duo/ Core i3/i5/i7 and Sandy Bridge processors. The company's focus is on user experience, which has A LOT to do with mobility / portability and that is partially related to excellent battery life of their products (esp. laptops). Steve Jobs ditched PowerPC chips for Intel since Intel offered far superior performance / watt. You are telling me no one cares about performance / watt, IPC and power efficiency? I got news for you -- manufacturers and designers who strive to create a killer mobile device/laptops/tablet care because it's a lot easier to sell "10 hour battery life notebook" with a fast processor than to slap a 2 lbs of protruding battery to a laptop and try to sell that as a premium product. Apple didn't choose AMD because AMD's current architecture doesn't meet any of the criteria that Apple wants in order to design a successful product.

Another sign that Apple does care about efficiency and performance / watt is illustrated by them abandoning NV for AMD this generation for GPUs. Lower efficiency implies either slower performance or more heat, which are undesirable compromises. Are you are still going to deny "no one cares about performance / watt" even after Intel adopted this as their key design strategy for implementing new features into their CPUs? :sneaky:



If your logic was true, no one would even bother upgrading to faster chips. Laptops would only add features, better aesthetics and slimmer form factors. According to you, Intel could have easily produced Core 2 Duo since 2006 and not done a single thing to increase its performance / watt. Yes, they did exactly that...oh wait.



Except that more cores alone isn't sufficient to win the server market share. You need an extremely power efficient processor. Throwing more cores won't solve the problem. AMD threw a 12-core Magny Cours against a 6-core Xeon and it didn't do squat for them. But you can go ahead and deny that IPC, Performance / watt, Performance per core don't matter.

You can be sure that when Bulldozer launches, its power efficiency will be scrutinized by many people in the server and consumer markets because it directly impacts operational costs and every day usability for mobile devices.



Pretty much to anyone who understands where the future of computing is heading - mobile, tablet, smartphone space where (1) performance / watt (2) IPC (3) performance per core is critical. I think only you are being in denial.

Are you insinuating that if Bulldozer was simply an 8 core Phenom II, AMD would be back on track? We know it wouldn't be enough, therefore, successes in the 3 areas I mentioned are critical to win back performance.



You obviously haven't met any engineers who design CPUs then. They are very passionate and want their firm to succeed.



I never said that enthusiast market is the only one that matters. But somehow you keep missing the point of what is necessary for a modern processor to be successful - Excellent power efficiency, great performance / watt, strong IPC and more cores. If you only focus on more cores, you will fail miserably.

Excellent post Russian, I pretty much agree with every point you made.

 

[smartpatrol]

XP+, A64 and X2 were all great gaming CPUs. Phenom I/II were not. So history has shown us that CPUs that excel in gaming, also tend to do well in common office tasks and generally have superior performance / watt and IPC. If we examine Pentium 4 and Pentium D, when Intel didn't focus on IPC, performance / watt, performance per core, then we have a situation where a CPU that didn't follow those criteria also didn't do as well in the marketplace.

And don't forget, this was while Intel was effectively paying OEMs not to use AMD processors.

 

[Arzachel]

I never said that enthusiast market is the only one that matters. But somehow you keep missing the point of what is necessary for a modern processor to be successful - Excellent power efficiency, great performance / watt, strong IPC and more cores. If you only focus on more cores, you will fail miserably.

So many words to say so little.
Power efficiency is actually the only way to increase performance due to having to cap TDP, so you're basically saying that to make a cpu perform better you have to make it perform better. And raw performance/W is useless, because ARM A9 is far better at it than Sandy Bridge, but scale up the former to the performance level of the later and its perf/W becomes comperativly abysmal.

Also, as far as I know, currently its FAR more cost effective to focus on scaling out instead of up. Ever wondered why Nvidia put 4 cores on their next Tegra cpu instead of pushing for higher IPC and clockspeed? The only drawback of more cores vs stronger cores is software adoption and Amdahl's law.

And I'd argue, that currently, IPC is farther down diminishing returns than clockspeed, so focusing on clockspeed should get AMD more bang for their buck.

 

[RussianSensation]

And don't forget, this was while Intel was effectively paying OEMs not to use AMD processors.

I forgot to mention another point. If you have a fast/efficient core to start with, the CPU is then much more flexible for use on the desktop/mobile and server space.

Servers / Desktop = as you start to scale to 4-6-8-12 cores, etc., your yields starts to fall off significantly. On top of that, it's a lot easier to get 4 working cores on 1 die at 4.0ghz than to do the same with 8 of the same cores at 4.0ghz. If you have a very efficient and powerful core to begin with, you don't require very high clock speeds to deliver very fast performance. This allows you to improve your yields (i.e., since yields will increase at lower clocks) and directly flows to your bottom line.

Mobile = Higher frequency at the same voltage increases power consumption, but more so, to achieve higher frequencies, you often require a LOT more voltage at the same node process. This can really be a negative for mobile space where lower voltage/wattage CPUs are favoured.

So while it may be possible to release a 4.5ghz chip at 1.30V, if you can achieve the same performance with a 3.3ghz chip at 1.20V, it's much preferable. But not only that, assuming you are manufacturing at the same 32nm process, the 2nd chip still has headroom to scale should you want a higher end SKU offerings, while the 1st chip is already approaching the magical 5.0ghz wall at 32nm, after which you are exponentially increasing leakage/voltage.

Performance / clock wouldn't be so critical if you could scale a chip to 10ghz on 32nm depending on the design. But it looks like we have been stuck around 4.0ghz-5.0ghz mark for a long time now despite being already at 32nm. I think Intel realized with Netburst that they won't be able to scale easily past 5.0ghz and changed their strategy to Performance / clock / watt as a result. If BD core is a lot more efficient than Phenom II core and we get 8 of those, then AMD has a strong shot at spanking SB.

 

[Darknite39]

I'm getting tired of waiting for BD to come out... thinking about just getting a new AM3+ board and an X6 instead to replace my DDR2 board and 2.6 Propus (I do a lot of x264 encoding). Unfortunately, I'd imagine that the old CPUs will drop a bit if/when BD hits, so I'm waiting either way, I guess.

I really just wish we'd get some straight info re: performance, even if the damn chips won't be out for another month+. That would really help my planning.

 

[darkswordsman17]

What happened in this thread? Now its a debate about cryptography?

When I first got into computers, I was a great fan of AMD. So dont take this post as coming from an intel fanboy.

But I just cant understand the people who keep defending AMDs outdated CPU architecture and delays in bringing new chips to the market. It is amazing how many excuses are made for AMD-- Intel cheated, they are the price/performance leader, their CPUs are "good enough", they are competing in the server market, no they are computing in the low end market, they have better graphics, etc.

Some of these points have some validity, put at some point you have to put out a chip that is competitive on performance per clock and per watt. And it would be nice if they could get it out on time too.

I see hardly anyone defending AMDs CPUs. Likewise about people making excuses. Then again, I don't know why you're even focusing on the frenzied gibberish that this forum has turned into, instead of just thinking rationally about this stuff, which you seem to be trying to do.

I'm completely baffled about your "they might be valid points, but stop bringing them up, AMD just needs to outdo Intel" mentality (to be honest, its not just you, but the way you put it perfectly shows a lot of the stuff people have been saying for months if not years now). Its like you understand but choose to completely ignore it.

That's what I'm seeing more and more of.

Ugh, I think CPUs is actually worse than Video these days, and that's quite a feat considering how bad Video is. Actually the worst threads there have been about CPUs funnily enough.

Another punk bench...

The more he post, the more IPC increase...:biggrin:

Please, just stop.

 

[frozentundra123456]

What happened in this thread? Now its a debate about cryptography?



I see hardly anyone defending AMDs CPUs. Likewise about people making excuses. Then again, I don't know why you're even focusing on the frenzied gibberish that this forum has turned into, instead of just thinking rationally about this stuff, which you seem to be trying to do.

I'm completely baffled about your "they might be valid points, but stop bringing them up, AMD just needs to outdo Intel" mentality (to be honest, its not just you, but the way you put it perfectly shows a lot of the stuff people have been saying for months if not years now). Its like you understand but choose to completely ignore it.

That's what I'm seeing more and more of.

Ugh, I think CPUs is actually worse than Video these days, and that's quite a feat considering how bad Video is. Actually the worst threads there have been about CPUs funnily enough.



Please, just stop.

Please do not put words into my mouth or misquote me. I did not say "stop bringing them up" (I assume you are talking about the points in favor of AMD) nor did I say anywhere that they had to "outdo" Intel. If that is how you interpret what I say, you are entitled to your opinion. However, that is not what I said, and you have no right to put into quotes something that I did not say. I said AMD needs to be competitive, not that they had to outdo Intel. There is a difference.

And I obviously was not ignoring the points in favor of AMD. Otherwise, why would I bring them up. And if you want to talk about denial, I would argue that the most obvious denial of all is that of AMD fans that their CPUs are no longer competitive.

 

[wlee15]

Speaking of cryptography...

http://www.sisoftware.co.uk/rank201...dfe2d7f199a491b7cff2c3e580e5d8e8cebd80b0&l=en

Edit: credit goes to Dresdenboy for finding the benchmark.

 

[Idontcare]

Speaking of cryptography...

http://www.sisoftware.co.uk/rank201...dfe2d7f199a491b7cff2c3e580e5d8e8cebd80b0&l=en

Holy smokes, and that's with only 2.356 GHz cores

 

[Plimogz]

^^^^^^^^
That's incredible!
Either there's something wrong with those readings, or AMD is keeping an absolute monster under tight wraps, for no readily explained reason.

 

[Abwx]

Speaking of cryptography...

http://www.sisoftware.co.uk/rank201...dfe2d7f199a491b7cff2c3e580e5d8e8cebd80b0&l=en

Details for Result ID 2x AMD Eng Sample, ZS262445TCG45_32/26/20_2/16 (2N 6CU 12C 2.36GHz, 6x 2MB L2, 6MB L3)

2mb L2 is correct , but neither the 6mb L3 , nor the core count
seems adequate , unless it s a legit 12C cache reduced variant.

 

[Idontcare]

2mb L2 is correct , but neither the 6mb L3 , nor the core count
seems adequate , unless it s a legit 12C cache reduced variant.

Don't forget these are multi-socket, so AMD's cache scout reserves some of the L3$.

 

[Abwx]

Don't forget these are multi-socket, so AMD's cache scout reserves some of the L3$.

Quite possible , but if true , the score of this 24C/24T server
is just 55% better than the equivalently clocked 32C/64T
Xeon 7560 , perhaps due to lack of the relevant instructions
in the Intel part.

There are other numbers of this ES in the same site:

http://www.sisoftware.co.uk/rank201...a69bae88e0dde8ceb68bba9cf99ca191b7c4f9c9&l=en

According to the arithmetic bench, it do 3.18 operations/cycle/core.

 

[Plimogz]

If I am getting this straight, and these results are trustworthy, and this is indeed the newly implemented AES in action, can someone with more knowledge in this matter speculate what other domains besides cryptography the new AMD architecture would excel at?

 

[StrangerGuy]

Holy smokes, and that's with only 2.356 GHz cores

I fail to see how is that impressive.

BD
(7GB/s) / (2 x 12 cores) / 2.4GHz = 0.121

980x
(4GB/s) / (1 x 6 cores) / 4GHz = 0.167

And that is only Gultown not SB-E.

 

[Abwx]

I fail to see how is that impressive.

BD
(7GB/s) / (2 x 12 cores) / 2.4GHz = 0.121

980x
(4GB/s) / (1 x 6 cores) / 4GHz = 0.167

And that is only Gultown not SB-E.

This comparison is pointless.
You should have done it with multi socket variants.

Do you think that SBE will have better bandwith that a gulftown?..
As much as 70/80% more , using almost the same IMC, memory
and channel count?...

 

[grimpr]

2mb L2 is correct , but neither the 6mb L3 , nor the core count
seems adequate , unless it s a legit 12C cache reduced variant.

root@bulldozer ~]# dmesg
Copyright (c) 1992-2011 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
The Regents of the University of California. All rights reserved.
FreeBSD is a registered trademark of The FreeBSD Foundation.
FreeBSD 8.2-STABLE #0 r********M: Tue Jun 21 18:37:22 MSD 2011
root@****.yandex.ru:/place/mkdump_workdir/obj/place/GIT-repos/FreeBSD-8-stable/sys/stable8-amd64-dtrace amd64
Timecounter "i8254" frequency 1193182 Hz quality 0
CPU: AMD Eng Sample, ZS262445TCG45_32/26/20_2/16 (2600.01-MHz K8-class CPU)
Origin = "AuthenticAMD" Id = 0x600f11 Family = 15 Model = 1 Stepping = 1
Features=0x178bfbff FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,MMX,FXSR,SSE,SSE2,HTT
Features2=0x1698220b SSE3,PCLMULQDQ,MON,SSSE3,CX16,SSE4.1,SSE4.2,POPCNT,AESNI,XSAVE,b28
AMD Features=0x2e500800 SYSCALL,NX,MMX+,FFXSR,Page1GB,RDTSCP,LM
AMD Features2=0x1c9bfff LAHF,CMP,SVM,ExtAPIC,CR8,ABM,SSE4A,MAS,Prefetch,OSVW,IBS,SSE5,SKINIT,WDT,b15,b16,b19,b22,b23,b24
TSC: P-state invariant
real memory = 68719476736 (65536 MB)
avail memory = 66525257728 (63443 MB)
ACPI APIC Table: <061711 APIC1910>
FreeBSD/SMP: Multiprocessor System Detected: 24 CPUs
FreeBSD/SMP: 2 package(s) x 12 core(s)

http://diybbs.zol.com.cn/10/11_99829.html