Opteron cleankills dual core Itanium 2 (Montecito)
4P 8 core Opteron, SpecInt_rate2000: 156; 4P 8 core Itanium 2 9050 (cost per chip: $3692) with total 96MB cache (24 MB per CPU) , SpecInt_rate2000: 134. Opteron leads by 16.4%.
Previously, we reported that the 8P Sun x4600 smashes the 16P HP Integrity Superdome with 16 Itanium 2 1.6GHZ.
In terms of floating point performance, Opteron ruins Itanium 2. the SUN x4600 got a SpecFP_2000 score of 3538. A Itanium 2 9050 got 3017. A 2.2GHZ Power 5+ got 3513.
Intel's Woodcrest Xeon 5160 got a SpecFP_2000 score of 3056 -- Opteron is 16% faster than Woodcrest and 17% faster than Montecito.
In SpecFP_rate2000, a 4P Opteron 885 (2.6GHZ) Blade got a score of 182. A 4P Montecito got 186, a statistical tie with Opteron 885. The best score for a Woodcrest server is 85.9. Of course, we know Woodcrest is only for low end 2P market, it's incapable of > 2 SMP. But you can see 2P Xeon 5160 is 47% of 4P Opteron.
posted by Sharikou, Ph. D. @ 7/30/2006 01:09:00 PM
121 Comments:
NOt disputing the results but: Would using the 24 MB L3 Cache version make a difference as the version they used does not have hyperthreading and has much less cache? Would using 16*2GB Ram instead of 32*1GB sticks in the intel system make a difference?
I thought one of your biggest complaints about server testing was when the don't use the sam OS?
Why are you comparing a 2.6Ghz AMD to a 1.6 Ghz Intel?
I also noticed that the Intel system is using 32x1GB memory, vs AMD 16x2GB memory. I can not compare this to a server but when I switched from 4x512MB to 2x1GB sticks in my own computer it picked a little performance.
If you, or someone could explain it would be much appreciated.
Thanks
I thought one of your biggest complaints about server testing was when the don't use the sam OS?
No. My complaint against Intel frauds were
1) they compare 32 bit Opteron result with 64 bit Woodcrest result, and fraudulently report that both were 64 bit.
2) they compare 2.2GHZ opteron results to 3GHZ woodcrest results when 2.6GHZ Opteron results were available.
Lol, everyone on this earth knows that Itanium works best with SpecFP. For servers, only SecFP has meanings.
when I switched from 4x512MB to 2x1GB sticks in my own computer it picked a little performance
Assuming that you use K8 CPU then that came from the ability to set the memory command-rate from 2 to 1. K8 is rather picky about the ram it uses (number of sticks, timings, single/double sided).
Unfortunately I can't comment on Itanium in that matter.
Btw, those Itaniums were the cheapest ones availiable costing around $700[1] a piece or ~$11.2k all together. The 885 Opterons cost ~$2.3k a piece[2] or ~$18.4k for eight CPU's. Unfortunately I don't know how much the rest of the machines cost.
[1] http://en.wikipedia.org/wiki/Montecito_(processor)#Features_and_Attrib
[2]
http://froogle.google.com/froogle?q=opteron+885&btnG=Search+Froogle
they compare 2.2GHZ opteron results to 3GHZ woodcrest results when 2.6GHZ Opteron results were available
Would you be so kind and provide some links to benchmarks for those Woodchrests and 2.6GHz Opterons?
those Itaniums were the cheapest ones availiable costing around $700
Dude, pay attention, according to your link, the Itainum 2 9050 costs over $3500.
Floating point performance:
HP ProLiant DL380 G5, Xeon 5160 (3056)
CPU enabled: 1 core, 1 chip, 2 cores/chip.
Sun Fire X4600, Opteron 856 (3538)
CPU enabled: 4 cores, 4 chips, 1 core/chip
Did I miss something ?
Did I miss something ?
Do you think 2 woodcrap is faster than 8 Itanium cores? Use your brain, dude.
Dude you got a brain? Or are you really this stupid.
No one cares about Itanium! It is a side show for the computing world and INTEL. What matters is x86 world and he who rules the benchmarks is KING
Woodcrest Rocks
Conroe Rocks
Merom Rocks.
AMD brings up the back of the bus
You can continue to beat your meat but..
Get over it, go find a real job.
What matters is x86 world and he who rules the benchmarks is KING
Clearly, on FP performance, Opteron smashes Woodcrap, the data shows it.
In any case, Woodcrap is for low end 2P server only.
emotions, emotions, emotions sharikou.. Your fanboys may get raddled by your potty mouth..
Here goes the last respectable thing you thought you had, a non emotional viewpoint!
But I got to admit, you are creative in finding childish wording.. After all, no technical mojo!
Hey, I understand the emotions you are in.. Do you feel blowted? after all, this past week has been hard on you and Hector
Your fanboys may get raddled by your potty mouth..
I never have much patience with retarded Intel fanboys. Have some IQ, and don't waste my precious time.
Doctor Doctor...
Pray tell me how man 4 way servers are sold every year vs. 2 ways?
and pray tell me how many single CDPU dual cores?
Tally the total revenues for me please.
Tell me what happened to Sun, DEC, HP, and IBM with their high end CPU sales trying to save the company.
In the end 4 way won't save AMD any more then the top end saved DEC and won't save HP, IBM and SUN.
Sorry you are the losing side...
But it is fun teasing you..
Where did you get your Piled High and Deep? Clearly not a rigirous technical school.
Do you think 2 woodcrap is faster than 8 Itanium cores? Use your brain, dude.
That's his point genius, those numbers don't make any sense. Something is wrong with your thinking if you think a current gen Opteron is going to beat a current gen IA64 CPU in Floating Point. It's simply not possible. In integer, sure. But it will not happen in FP, the idea is laughable.
tell me how man 4 way servers are sold every year vs. 2 ways?
In revenue terms, for AMD, one 4P dual core machine generates about $8000 revenue (4 x $2000) and $7000 profit. One 2P machine generates $800 revenue and $600 profit. Selling one 4P makes 10x the money of a 2P. That's why HP sells about the same number of units of x86 server as DELL, but makes 3x the revenue.
That's his point genius, those numbers don't make any sense.
Intel fanboys always have low IQ.
They just don't get it.
genious or tard?
Sure 8K sounds like a lot fanboys. Noticed the PhD didn't bother to tell you how many they sold. Take that number and distribute it across the total cost of running the company.
What happenend to those CPU and minicomputer companies that thought high ASP and high profit margins would save them; DEC, data general, Bull, Unisysi, Sperry, SUN, even IBM can't afford the silicon development for their Power.
Thats is the simple facts.. all these companies with the high ASP and profits per machines eventually went bankrupt or turned to CPUs from INTEL.
Where did you get your PhD...? It will tell us what how your predgree is...
Sure 8K sounds like a lot fanboys. Noticed the PhD didn't bother to tell you how many they sold. Take that number and distribute it across the total cost of running the company.
Moron. I gave you 3 equations
price_4p = 10 x price_2p
unit_hp = unit_dell
revenue_hp = 3x unit_dell
Try solve and find out how many 4way machines HP sell.
Also, in a previous post, I estimated that Google builds 100K 4P servers per quarter.
Dude, pay attention, according to your link, the Itainum 2 9050 costs over $3500.
What has that got to do with anything? They didn't use those in that machine, they used the 9010 version, 16 of them to be percise. Didn't you read the system information or what?
Btw, how much do those special dualcore 3GHz versions of 8P Opterons cost?
Also, I'm still waiting for those links about Woodchrest vs 2.6G Opteron. You did say those results vere availiable so please show them.
Sharikou,
I am very curious, you always make a lot of comments and analysis on the CPU pros and cons. I was wondering if you could tell us what your credentials are? I mean what kind of education did you get? Do you indeed have a PhD? If so in what and from where? And what is your publication record?
I am sure many people would take you seriously once you prove to them that you are a credible PhD with good publications (and maybe a good GPA too).
Of course you don't have to have a PhD to be taken seriously, but you are sharikou... phd... arent you?
Thank you!!
price_4p = 10 x price_2p
unit_hp = unit_dell
revenue_hp = 3x unit_dell
Isn't that 3 equations and 5 unknowns??
I see you have updated the post. Let's see what systems are you comparing:
K8: AMD Opteron (TM) 85g 3000MHz 4 cores, 4 chips, 1 core/chip
Itanium2: Itanium 2 9050 1600MHz 4 cores, 4 chips, 1 core/chip
Woodcrest: 3.0GHz, Intel Xeon processor 5160 1 core, 1 chip, 2 cores/chip
Power5: 2200 MHz, 1 CPU 1 core, 8 chips, 1 core/chip (SMT off)
In short we have 4 K8 vs 4 Itanium2 vs 2 Woodcrest vs 8 Power5
Now lets do some fun math and divide the results to cores and now we have this:
K8: 884.5
Itanium2: 754.25
Woodcrest: 1528
Power5: 439.125
So according to your data, Woodcrest is about 72.8% faster than K8. Is my math correct?
I just noticed another update you made. You got rid of the old Itanium and K8 systems (16x9010 vs 8x885) and replaced them with new ones. That happened soon after I pointed out the price advantage Itanium system had over the Opteron system. That's right, the old and superexpensive Itaniums were actually considerably cheaper than the fancu Opterons.
That kind of selective infromation makes me think about creating my own blog just for pointing out "mistakes" like that in the couple of blogs I sometimes read.
Wirmish said...
Floating point performance:
HP ProLiant DL380 G5, Xeon 5160 (3056)
CPU enabled: 1 core, 1 chip, 2 cores/chip.
Sun Fire X4600, Opteron 856 (3538)
CPU enabled: 4 cores, 4 chips, 1 core/chip
Did I miss something ?
Really something strange I see with Woody Spec's benchmarks.
SpecFP2000:
1) ProLiant DL380 G5 (3.0GHz, Intel Xeon processor 5160, one chip) - 3056;
2) PowerEdge 1950 (Intel Xeon processor 5160, 3.00GHz, two chips) - 2818.
All results were published in July.
Did I miss something?
Yurif.
The Memory Controller on the K8 is sensitive to timings as we know, and switching from 4x 512 to 2x 1GB lowers the command rate (the time it takes to access a RAM chip) from 2T to 1T, and improves performance by up to 6%.
The problem with todays Memory is that it is more healthy to Intel than AMD. You are hard-pressed to find 1GB sticks of RAM that have less than CL3 (DDR1) and DDR2-800 that has less than CL4. What I would love to see is a FX-62 @ 3.0GHz w/ 2GB (1GB x2) of PC2-6400 w/ the Extended Voltage series RAM w/ timings @ 2-2-2-5 800MHz.
Obviously the price will be ridiculously high, but the performance would no doubt be worth it.
"What happenend to those CPU and minicomputer companies that thought high ASP and high profit margins would save them; DEC, data general, Bull, Unisysi, Sperry, SUN, even IBM can't afford the silicon development for their Power."
AMD should scrap K8L and HT3.0 all together so. 4P is actually NOT the high-end but is mid-range for AMD-based systems for sure. HORUS, and some other third-party Opteron-based configurations up to 32P are the tops.
Like Sharikou said, use your brain's logic side a little. If 1-2P is low-end for servers, what makes 4P high-end so much out of the wazoo? The game changed, AMD turned the tables up, period. When K8L is out, 8P will be commodity servers.
wow =) good information sites..thanks for the effort =)
This blog always reminds me of this website:
http://www.kcna.co.jp/index-e.htm
If there is an opinion on this blog by the writer, it is always anti-Intel and pro AMD. There are no concessions.
Much like North Korea cuts down the US. US is evil, can do no good, and Kim Jong Il is the king, baby!
Sharikou, can you comment? Has Intel ever done anything GOOD? Or will you admit that this blog is nothing more than propaganda.
I could care less about Intel or AMD, but I keep coming to see how crazy it can be over here.
Sharikou you are such a genius, for a decade you are the first one I've ever seen who cmpares SpecInt between Itanium and Opteron. You are the live proof of how stupid/dumb AMD fanbois are
SpecInt? LoL, maybe next time compare x86 performance. Opteron would definately "frag" Itanium 2.
(Too bad K8 is very dead in SpecFP...)
They didn't use those in that machine, they used the 9010 version, 16 of them to be percise. Didn't you read the system information or what?
System info here:
Hardware Vendor: Bull
Model Name: NovaScale 3045 (1600MHz)
CPU: Itanium 2 Processor 9050 1600 MHz FSB 533MHz
CPU MHz: 1600
FPU: Integrated
CPU(s) enabled: 8 cores, 4 chips, 2 cores/chip
Intel retards are wasting world's resources.
Sharikou,
I am very curious, you always make a lot of comments and analysis on the CPU pros and cons. I was wondering if you could tell us what your credentials are? I mean what kind of education did you get? Do you indeed have a PhD? If so in what and from where? And what is your publication record?
I am sure many people would take you seriously once you prove to them that you are a credible PhD with good publications (and maybe a good GPA too).
Of course you don't have to have a PhD to be taken seriously, but you are sharikou... phd... arent you?
Thank you!!
You will NEVER get Shari-fraud to answer this. He hides behind his made-up PhD and slanders what he doesn't like and cherry picks facts and benchmarks to suit his case. He is almost as bad as GWB going to war in Iraq!
What I would love to see is a FX-62 @ 3.0GHz w/ 2GB (1GB x2) of PC2-6400 w/ the Extended Voltage series RAM w/ timings @ 2-2-2-5 800MHz.
Too bad that this kind of timings are physically impossible with DDR2 :)
System info here:
How nice of you to bend the truth like that.
I was talking about the benches that you posted before you edited the entry and replaced the benches with cheapest parts with benches on the the most expensive ones.
As can be seen from the first couple of posts the original entry didn't have the super-expensive Itanum in it, why else there are talks about "why you didn't have benches of the high-end CPU's". You added it after my first comment on the prices. You even added the link to Wikipedia I gave in your entry :)
I do like that you included the FP benchmarks that show that Core2 absolutely obliterates everything else having nearly the same performance as systems with twice the core numbers and CPU's costing ~3x as much :)
Intel retards are wasting world's resources.
You know, retards are also people and we have special placed for them where they can live their peaceful life. I do agree that pointless bashing is a waste of recources.
May I ask with what I made you think I'm a retard? For me it seems that I only interpreted the information you gave.
am very curious, you always make a lot of comments and analysis on the CPU pros and cons. I was wondering if you could tell us what your credentials are? I mean what kind of education did you get? Do you indeed have a PhD?
People have the tendency to believe in authoritative figures. A PhD is some accomplishment. For instance, Pat Gelsinger told a story that his mother always asked him to get a Ph.D, which he didn't. But it's still not too late for him to do it...
Many people with PhDs are visiting this blog for sure, do they ask the question whether I have a PhD for real? No. For those who have gone through and done it, it's just natural for someone with certain level of intelligence to end up getting the highest level of education. Those who are asking this question are those without a PhD. But then, I don't have to show my PhD to someone with lower degree to make me more authoritative. Call me Dr. Sharikou if you want, but I feel more comfortable if we are discussing the merits of matters without such formality.
"You will NEVER get Shari-fraud to answer this. He hides behind his made-up PhD and slanders what he doesn't like and cherry picks facts and benchmarks to suit his case. He is almost as bad as GWB going to war in Iraq!"
Just like you'll never bother countering his ideas with hard facts instead of posting your incessant dribblespeak? PLEASE. Until you can prove your points with some cold hard facts of your own, do the rest of us a favor and shut your own pie hole, thanks.
Just because you don't like the man's ideas doesn't mean he's not accurate. So far he's got more evidence to back up his claims than YOU DO.
[wayyyy offtopic, soapbox]
You're totally clueless about the full scope of the Iraq situation,(like most people unfortunately, take liberals for example) I take it you're not from the US. This goes far beyond GWB as well, the Middle East has been brewing a conflict of this scale for over the past 30-40 years and it all really started with the Carter administration and possibly Nixon's term.(basically Carter acting like a wuss to the Islamic fanatics thinking the US is a paper tiger) This war is against Islamic fundamentalist extremists,(terrorists, not Iraqis) and rooting them out away from US soil.(and rebuilding a new Iraq for the Iraqis as well in the process) In other words, we're fighting in the Middle East to draw our enemy to Iraqi against fundamentalists who say "Covert to Islam or DIE!!!!" This means EVERYONE else who isn't a follower of Islam!(YOU and me!) Make no mistake about this, this involves us all and this is not a war any of us can afford to lose. We will win or we all secede to the eventual brutality of Sharia law, effectively Islamo-fascism.
[/wayyyy offtopic, soapbox]
Sharikou said:
"Moron. I gave you 3 equations
price_4p = 10 x price_2p
unit_hp = unit_dell
revenue_hp = 3x unit_dell
Try solve and find out how many 4way machines HP sell. "
Like some poster above pointed out, perhaps you can explain in your "Perverse journal of 64-bit computing" how to solve for 5 unknowns through 3 equations? I'm sure its some PhD trick....
[wayyyy offtopic, soapbox]
You're totally clueless about the full scope of the Iraq situation,(like most people unfortunately, take liberals for example) I take it you're not from the US. This goes far beyond GWB as well, the Middle East has been brewing a conflict of this scale for over the past 30-40 years and it all really started with the Carter administration and possibly Nixon's term.(basically Carter acting like a wuss to the Islamic fanatics thinking the US is a paper tiger) This war is against Islamic fundamentalist extremists,(terrorists, not Iraqis) and rooting them out away from US soil.(and rebuilding a new Iraq for the Iraqis as well in the process) In other words, we're fighting in the Middle East to draw our enemy to Iraqi against fundamentalists who say "Covert to Islam or DIE!!!!" This means EVERYONE else who isn't a follower of Islam!(YOU and me!) Make no mistake about this, this involves us all and this is not a war any of us can afford to lose. We will win or we all secede to the eventual brutality of Sharia law, effectively Islamo-fascism.
[/wayyyy offtopic, soapbox]
well. if the war in iraq is about getting rid of islamists, then it's a extremly stupid thing to replace a more or less stable (non islamistic) dictatorship with a extremly fragile democracy. isn't it?
but little do i know... i'm sure the thousands of us soldiers and the ten thousands of civilans that died and that will die in future are worth whatever purpose this war was started.
People have the tendency to believe in authoritative figures. A PhD is some accomplishment. For instance, Pat Gelsinger told a story that his mother always asked him to get a Ph.D, which he didn't. But it's still not too late for him to do it...
your writing style makes it very difficult for a lot of your readers to believe you really aquired the claimed ph.d.
Many people with PhDs are visiting this blog for sure, do they ask the question whether I have a PhD for real? No. For those who have gone through and done it, it's just natural for someone with certain level of intelligence to end up getting the highest level of education. Those who are asking this question are those without a PhD. But then, I don't have to show my PhD to someone with lower degree to make me more authoritative. Call me Dr. Sharikou if you want, but I feel more comfortable if we are discussing the merits of matters without such formality.
we will call you dr. shakirou, when the arguments you make in your blogs have reached a certain kind of discussion level.
if you like to stop the discussion about your "title" get rid of it in your description or bring prove for it.
your writing style makes it very difficult for a lot of your readers to believe you really aquired the claimed ph.d.
So, you expect someone with a PhD to talk like a priest with big words and a rigid face and look down at you?
You know, I kinda enjoy this discussion. It proves again that my degree is actually a valuable asset. I think all the PhDs visiting this site will agree.
"if you like to stop the discussion about your "title" get rid of it in your description or bring prove for it."
I think his title is something that doesn't need to be discussed or hated upon by others who think its necessary. This blog is posting Sharikou's opinion on matters and others respond with their own opinion. Sharikou obviously knows something about computers, you can't make this sh*t up. The people who constantly demand proof of his PhD or anything else need to shut up.
Give an opinion on a matter and stop insulting. We can go all day with "you started it first" which will wind us up back here, so from now on how about we all give facts based upon our views and what we see with our eyes and hear with our ears, instead of calling names like morons lets debate like men and women.
So, you expect someone with a PhD to talk like a priest with big words and a rigid face and look down at you?
Not that but calling names, selective information and modifying own words to make ones point correct is certainly not a usual thing you see from highly educaded person. Usually I have seen such behavior amongst kids.
But then again, having high education does not neccesarily mean that one is not stupid ...
Can anybody explain this line to me :
"Core 2 Duo fastest ramping chip in Intel's history, Otellini says "
hmm I dont remember reading about ramping when the P4 came out .. :))
shakirou said ...
You know, I kinda enjoy this discussion. It proves again that my degree is actually a valuable asset. I think all the PhDs visiting this site will agree.
i doubt there are any phds reading or writing on your blog. except some psychologists. for them it's an interesting field study.
so go ahead. make my day.
"You know, I kinda enjoy this discussion. It proves again that my degree is actually a valuable asset. I think all the PhDs visiting this site will agree."
I don't have a degree and I probably never will, but I guarantee I can hold my own against the best of them.
"So, you expect someone with a PhD to talk like a priest with big words and a rigid face and look down at you?"
People look at companies and expect everybody to act like them, simply because that has become the "norm" in America; something that is horribly G4Y.
Not that but calling names, selective information and modifying own words to make ones point correct is certainly not a usual thing you see from highly educaded person.
You Intel kids are allowed to run free here. Everytime you run out of ammo, you start digging mud on personal matters. I have presented solid benchmarks on SPEC.org showing Opteron frags the fastest/largest cache dual core Itanium 2 on both floating point and integer performance. You Intel kids can't accept the numbers?
"we will call you dr. shakirou, when the arguments you make in your blogs have reached a certain kind of discussion level.
if you like to stop the discussion about your "title" get rid of it in your description or bring prove for it."
We? do you represent a group poeple or what? Who decide if the discussion level is good or not ? YOU? Who the hell are you?
It silly to expect people to prove themself simply you don't agree with others.
For following the your logic? please prove yourself is not mad, otherwise, shut up!
Well, Intel is on the offensive and its time for AMD to fold up....
AMD Fanbois... the writing is on the wall... please dont waste time trying to find some means of bashing intel.
Q3 will be the turning point. just look at SEMICO shipment data for this quarter. AMD's ASP's will come crashing down and will result in a net operating loss...
I have presented solid benchmarks on SPEC.org showing Opteron frags the fastest/largest cache dual core Itanium 2 on both floating point and integer performance
Indeed you did. Also you yet again modified the benches and removed the specFP results of the 1P Core2 Xeon. If that is not selective information then I don't know what is.
Here[1] is an image of my post where I have the results and basic configurations of the systems. There I have a Woodchrest. Did I put it there in random?
Any good explanation of why you removed it? Should I look that bench up and post it here or will you add it into the comparison?
You said:
You Intel kids can't accept the numbers?
Oh, the irony :)
Let me calculate some more numbers on the specFP benches you posted here :)
4x 865 @ $1.6k[1] = $6.4k
1x Core2 Xeon 5160 @ $0.9k=$0.9k
Overall Core2 system was ~15% slower but it's cpu cost ~7x less than the Opterons.
Now where is that uptimate FP power K8 was supposed to have over Core2?
[1]
http://img516.imageshack.us/img516/7320/modifiedlz6.png
[2]
http://froogle.google.com/froogle?q=opteron+856&btnG=Search+Froogle
[3]
http://froogle.google.com/froogle?q=Intel+Xeon+processor+5160&btnG=Search&lmode=online&scoring=p
For me it seems that you will bend the facts however you can. Only thing that matters for you is that AMD must look great compared to Intel.
You say that intel fans talk rubbish and don't give facts. You did give facts and when they didn't suite you you removed them.
If I don't get this post publshed as-is it would be a perfect example of how low you can go just to prove your "facts".
People have the tendency to believe in authoritative figures. A PhD is some accomplishment. For instance, Pat Gelsinger told a story that his mother always asked him to get a Ph.D, which he didn't. But it's still not too late for him to do it..."
Many people with PhDs are visiting this blog for sure, do they ask the question whether I have a PhD for real? No. For those who have gone through and done it, it's just natural for someone with certain level of intelligence to end up getting the highest level of education. Those who are asking this question are those without a PhD. But then, I don't have to show my PhD to someone with lower degree to make me more authoritative. Call me Dr. Sharikou if you want, but I feel more comfortable if we are discussing the merits of matters without such formality.
Sharikou, I place very little importance on whether or not people have degrees. George Bush is a moron and he's got letters behind his name, likewise, some of my friends are morons and have got degrees in this and that.
Conversely, a lot of brilliant people have degrees, and far more DON'T. So I would argue that "degrees" are no measure of intelligence or that one naturally gets a degree in something because they are intelligent. Degrees are received because a person is determined (or well-connected), not because they are intelligent.
However, to proclaim that one has a degree, yet truly doesn't is shameful and disgraceful and devalues the efforts that people have gone to in order to get their own degrees.
I am NOT saying that you aren't a doctor. However, your response to this person's valid question was a non commital answer. You did not say whether you are truly a doctor or not.
I feel that by giving such a ambiguous response almost implies that you are NOT a doctor.
You have no reason to respond to this inquiry of course, after all this is YOUR blog and we are your guests, whether you are a doctor or a farmer or a terrorist.
However, proclaiming that you are a doctor when you may not be would be an outright lie. I DO think you owe it to your readers (some of which are doctors, of course) to tell the truth.
I know that by providing some sort of proof it would only add value to your statements and prove unconvincingly that you are not a liar.
Do you understand? You are very honest and let people who are crazy or against you post in this forum, and this makes me believe that you are who you say your are as you don't often stoop to their level. I rarely agree with your opinions, but by proving at the very least that you are not a liar you would add substance to your claims.
If this was a big hoax, well, simply take the PH.D off your account and apologize. Nothing more needs to be done!
Please say yes or no, and if 'yes' maybe some of us would enjoy reading your published material on topics other than computing?
Thanks,
A reader.
I didn't fully understand what SPECfp, and SPECint were, so I looked it up... SPECint_rate2000, and SPECfp2000.
After looking it up it would seem that Itanium has not been optimised.
If you look at the test results both SPECfp_base2000, and SPECfp2000 have the same scores and also with SPECint_rate_base2000, and SPECint_rate2000.
Here is my question... Would it be fair to say that due to the lower frequency and lack of optimizations that the scores will get better over time for Itanium 2?
Something I noticed is that on the scores before optimisation, Intel seems to be very close on SPECint_rate_base2000 Intel 134, AMD 138 and actually ahead in SPECfp_base2000 Intel 3017, AMD 2851.
If someone could elaborate on this it would be much appreciated.
Thanks.
Also you yet again modified the benches and removed the specFP results of the 1P Core2 Xeon. If that is not selective information then I don't know what is.
You Intel retards are beyond education. Why don't you use your brain a bit instead of wasting my time? You thought 2 Woodcrap is as fast as 4 Itanium 2 dual core? Crap heads. Someone posted a result showing 2P woodcrest gets lower SpecFP score than 1P, didn't that ring something in your head?
The SpecFP_2000 is about running one copy of the test programs at a time. Whether you have 8 cores doesn't matter--only one copy was running at a time. It is test for one core-- and Opteron is faster. You can see that from the description of the result page.
I removed the Woodcrest so you crap heads won't waste my time any more explaining this, and at the end, you keep asking such stupid stuff.
You Intel retards are beyond education. Why don't you use your brain a bit instead of wasting my time?
Is this how you treat guests to your blog?
Calling them retards?
PhD and alike topics.
For some a PhD represents a achievement which ones earns by years of independent research culiminating in a discover/undestanding/furthering the art in a certain field. Of course the institution and professor you earn it from has a huge impact on is it really something signficant or barely worthy of the value of the sheep skin the diploma is printed on.
Most people I know that have earned a PhD are open, willing and excited to share their personal trials/tribulations and the achivements and wisdom gained in this endeavor. FOr some there is cynism as it doesn't necessarily guarantee nor predict future fame/fortune/success. But it is an entry ticket to many areas; R&D in many companies, Professor, etc.
The fact Sharikou refuse to divulge where and what he earned his PhD is enlighting itself.
Sharikou you can post or not this.. but you claim it is about mudd slinging. Nothing willl quiet the mudd more then steping forward and bravely show the man behind the title you so proudly bear on you signature.
In my experiece I've found those that throw around their PhDs on business cards, titles etc. are those that are most insecure in their own state.
The Doctor
PS.. PhD has little reflection on ones typing, grammer, nor future prediction of success. Look only here to see what happens to one PhD.
"Is this how you treat guests to your blog?
Calling them retards?"
When the guests shit in the pool knowing they did something stupid how do you think the host is gonna react? Simple, fast, hard rules of the house by the host, it's his damn house and he does whatever the hell he wants!
Like the man said, when they don't bother to use their brains continuously asking stupid questions that are easily researched with Google do you blame him for calling them retards??? I sure as hell don't!
It's real simple Intel fanboys, you don't like it here? He doesn't seem fair??? DON'T LET THE DOOR KNOB HIT YOU ON THE ASS ON THE WAY OUT!
Most people I know that have earned a PhD are open, willing and excited to share their personal trials/tribulations and the achivements and wisdom gained in this endeavor.
Most of my friends are PhDs, and none of them like to be obssessed with the past. For them and myself, getting a PhD is just natural. You are talking about people who grew up to be the top of their classes. No big deal there.
The SpecFP_2000 is about running one copy of the test programs at a time.
Hm, seems like my memory is failing me then. For some reason I remembered that those were run in parallel. Seems as I mixed them up with CFP2000 Rates.
Btw, as those tests measeure the speed of only one core then why am I seeing so majorly different results on other systems with same CPU?
http://www.spec.org/osg/cpu2000/results/res2006q3/cpu2000-20060707-06419.html
http://www.spec.org/osg/cpu2000/results/res2006q2/cpu2000-20060410-05842.html
It seems that one of the AMD systems has about 50% more performance than the others. That X4600 seems to be the only one that has so high score. I wonder what they did with that system to get it working so efficiently and why haven't others done that.
What do you think, is it absolutely impossible to get any more speed out of that Xeon? If so then what is stopping it?
I would imagine that giving it some DDR2-800 memory with a bit better timings might help a bit. AMD has already rather good memory. After all it has been shown that it takes DDR2-800 with relatively tight timings to get the performance of DDR1-400 with tight timings.
Too bad I don't have Core2 and K8 myself. If I would I could perform all sorts of test on it. Anyone wants to give me some so they could be put side-by-side ?
:)
Sharikou said:
"Moron. I gave you 3 equations
price_4p = 10 x price_2p
unit_hp = unit_dell
revenue_hp = 3x unit_dell
Try solve and find out how many 4way machines HP sell. "
Like some poster above pointed out, perhaps you can explain in your "Perverse journal of 64-bit computing" how to solve for 5 unknowns through 3 equations? I'm sure its some PhD trick....
Any update on this yet Shakira?
"s this how you treat guests to your blog?
Calling them retards?"
"Those to whom evil is done, do evil in return." - Roger L. Simon
Seriously, getting a phd, unless it is in a science or mathmatics is not that difficult. It is time consuming and very expensive. Most of the non science phd's I know english, history, linguistics, economics, religion, education are the poor tortured artist types that got their ph.d. because they could. It doesn't take a genius, except math and science. Law degrees are no better with the exception that self respecting lawyers choose not to put their title in their name esq., j.d. or worse yet since it is a terminal degree dr. Just a bunch of people with time and money. I am one of the above so I have had the opportunity to observe many of individuals for which I speak.
I am curious to know what shakirou's ph.d. is in from a reference standpoint to get an idea of his basis of knowledge. In addition, whether it is from an american university or not.
Most of my friends are PhDs, and none of them like to be obssessed with the past. For them and myself, getting a PhD is just natural. You are talking about people who grew up to be the top of their classes. No big deal there.
And for some reason you obsess over how you feel Intel behaved in the past?
There is no arguing that you're intelligent, you're adept at wielding all sorts of data like a weapon, effecively building traffic to your website whether you are right or wrong.
However, someone with more intelligence would understand that being a little (just a LITTLE!) would help promote their ideas better. You are not effective (apparently) at giving credit where credit may be due.
Like I said, however, there's no doubt in my mind you're generating thousands of hits on this blog a day, but the retention is very little because aside from your ardent supporters, not many people (including industry players) put much merit into what you say.
Why not? The words you choose and the data that represents your ideas is selective. The comments you make are dismissive when someone comes forth with pertinent data. You display no quarter when confronted. Every thought that you put forth is designed to devalue Intel.
By the way, people can see through your ambiguous statements and you still haven't given a straight answer to whether you are a doctor or not.
Would you like to address this question?
But then, I don't have to show my PhD to someone with lower degree to make me more authoritative.
-written ambigulously, as if to say "Someone with a degree doesn't have to show it off..."
Call me Dr. Sharikou if you want, but I feel more comfortable if we are discussing the merits of matters without such formality.
-again, ambiguous. One could say, "Call me Einstein if you want, but I feel more comfortable being known as Bob."
You know, I kinda enjoy this discussion. It proves again that my degree is actually a valuable asset. I think all the PhDs visiting this site will agree.
-"...my degree..." in? Basket weaving? Charlatanry?
Everytime you run out of ammo, you start digging mud on personal matters.
Well, then simply show us that you're a doctor! Nothing could quash this controversy quicker than a little bit of proof! We all promise to shut the heck up if you finally put the issue to rest, RIGHT EVERYONE?
Dr S.
I would advise you to not post your personal information for the edification of computer geeks. If you have ever been to a hacker’s convention you’ll know what I mean. (MS 2004) You have a lot of enemies. Only a cynical idiot would even ask you to. If you were to post such info I would have to say you are not a PhD. This is just a blog site; you do not owe anyone anything. That’s why people who post don’t leave there email address and other personal info, DA. Keep doing what you’re doing for all those that are interested.
" Anonymous said...
You Intel retards are beyond education. Why don't you use your brain a bit instead of wasting my time?
Is this how you treat guests to your blog?
Calling them retards?
1:53 PM, July 31, 2006 "
considering most guests here has come from anandtech powered intel-spawned retards ready to insult shakirou insteath of debating about computing stuff ( that includes insults to his tittle, his persona, his intelligence.. )
I wouldnt be surprised if he as been getting annoyed lately with the constant insults..
most guests here are more of trolls rather than people interested in a healty debate.
accept it guys, intel fanboys cant make a goodamn point.. and thus they get desesperate, moving to insults..
and that guys, means YOU ALREADY LOST.
"The Doctor
PS.. PhD has little reflection on ones typing, grammer, nor future prediction of success. Look only here to see what happens to one PhD."
dude, the irony must be making your ass very very itchy.
I mean you talking about titles, where lately the only thing you have been doing is, INSuLTING, not giving any aceptable post.
accept it "Doctor"
you're just a kid triying to act the man with acknowledge by insulting others who have real ideas, real predictions, real opinions, and not licking the shoes of a monopoly-controlling corporation wich seems to have sucesfully brainwashed your brain.
When the guests shit in the pool knowing they did something stupid how do you think the host is gonna react? Simple, fast, hard rules of the house by the host, it's his damn house and he does whatever the hell he wants!
Like the man said, when they don't bother to use their brains continuously asking stupid questions that are easily researched with Google do you blame him for calling them retards??? I sure as hell don't!
It's real simple Intel fanboys, you don't like it here? He doesn't seem fair??? DON'T LET THE DOOR KNOB HIT YOU ON THE ASS ON THE WAY OUT!
Who said I was an Intel fanboy? I could really care less about CPU's and corporations. I come here for the drama. I come to see people make asses out of themselves by blind assumptions.
If I want a respectable unbiased opinion, this definitely isn't the place to go! It's just fun here.
Especially when the cage-keeper loses his cool and calls everyone dummies and idiots and twists everything around to his convenience, while his idol-worshipping lemming followers accept what the cage-keeper says without question.
The truth is that this blog isn't a good scientific forum. It's based on other people's benchmarks and tests and all the opinions here based on these third-party evaluations.
The stock market speculation found by the author here is downright ludicrous. If anyone is basing their life savings on the opinions here, God help them. More importantly, the opinions expressed by the author/cage-keeper are completely one-sided, much like propaganda from various nations in the past and present.
Sharikou is smart. He knows how to develop a cult of personality. He has capitalized on the controversy his ideas generate and has ample spare time to post around the net to draw people in. I'm surprised that he hasn't rolled out advertising yet.
His theories and opinions can be validated by outside information if read out of context, and sometimes his opinions can even be fully factually backed up. Sometimes he's absolutely justified in saying that Intel was a tyrannical corporation in the past. I don't discount everything he says just because HE said it.
However, the pure unadulterated hatred for anything Intel should always be questioned by his followers and people who have a vested interest in his topics.
Yes, it's his blog and he can do whatever the hell he wants. Door hit me in the ass, yah yah. But I must give it to Sharikou, he allows everyone to speak freely and he deserves a ton of credit for that. I don't think I would be as generous as Sharikou in that respect, and I'm pretty certain that my own personal intelligence level far exceeds that of the good (?) doctor. (Note: I just said that because I like tooting my own horn. Like putting a PH.D at the end of my name, I guess!)
Not everyone knows everything about processors and benchmarks, and when someone is insulted as such it seems that those people that often choose to side with Sharikou are far more interested in ridicule than explanation, which goes against their own motives in a way. Motives being "looking intelligent".
So...before you go out and defend your leader, remember this: You look like a farthead and a noony-noo-noo-pooface.
http://www.tigerdirect.com/applications/Category/category_tlc.asp?CatId=22
tigerdirect shows AMD in the top 5 again leading in sales too.
"You are talking about people who grew up to be the top of their classes. No big deal there."
Lol, you are getting funnier everyday...
Sharikou said:
"Moron. I gave you 3 equations
price_4p = 10 x price_2p
unit_hp = unit_dell
revenue_hp = 3x unit_dell
Try solve and find out how many 4way machines HP sell. "
Like some poster above pointed out, perhaps you can explain in your "Perverse journal of 64-bit computing" how to solve for 5 unknowns through 3 equations? I'm sure its some PhD trick....
Any update on this yet Shakira?
Does anyone know how to do this? ;)
he allows everyone to speak freely and he deserves a ton of credit for that. I don't think I would be as generous as Sharikou in that respect
I allow this, because I look at everything from a more philosophical perspective. We are only at the beginning of a computing revolution. AMD64 will take us to another level, but that's not the end. Intel is history -- if you look back in two years.
One thing everyone should learn is to focus on issues. I found Intel execs particularly uncivilized and under educated. Intel got a crap architecrure, so fix it, copy it, whatever, make a change. No need to be angry at AMD and live in anger every day. AMD is gonna rule, so accept the inevitable. When you are in a hole, stop digging. Look at what Intel execs are doing, they are digging their own graves. Unless they change their course, you will see Intel BK in 5 to 7 quarters.
I am curious to know what shakirou's ph.d. is in from a reference standpoint to get an idea of his basis of knowledge.
I was in this meeting with doctoral candidates from other fields, it was something about getting a fellowship from the university. There was some guy who was in a doctor of education program, doing some research on kids. mostly doing some statistics and then waving hands, saying this factor is important. That kinda of stuff. I pretended to show some interest. But the guy told me that the stuff we did was true science, his was just liberal arts stuff...
My PhD thesis? I doubt anyone at Intel can understand it. Maybe someone at IBM can.
I've read your comments over and over Sharikou. I cannot fathom how Intel will be out of game anytime in the future aside from legal complications.
Assuming that Intel is not doing anything illegal that would cause the corporation to cease its functions, how can you possibly say such a huge and important company (regardless of their ethics or products) will be out of business within a year or two?
I mean, most of the consumer market can identify Intel and has some sort of confidence in the brand. AMD is not so fortunate and cannot hope to achieve this same comfort - in the near future anyways.
Brand alone could keep Intel going for years, even if the corporation was helmed by chimpanzees.
I think that you forget that people in general do not feel the same as you do towards Intel. To John Q. Public, Intel is a perfectly fine company, and for all they know sells products that are on par or at least more recognizable.
For the John Q. Public (who has a modicum of computer understanding), an "AMD 64FX" means very little. On the other hand, a Pentium 4 3.0 Ghz clearly illustrates a generational gap from a Pentium 3 (which they may have) and 3.2 Ghz would appear to be double a 1.6Ghz machine.
AMD portrays their products confusingly to the consumer. Even if they revamped their whole nomenclature system today it would take at least longer than you predict for Intel's demise to change consumer opinion.
This alone would be enough to keep the chimpanzees rolling in bananas.
Look Sharikou, I really like reading your blog and have learned ALOT from it but the name-calling just detracts from the validity of your arguements.
It's just unprofessional and quite frankly, makes you look bad and perhaps even worse than the intel fanboys.
Using words like "woodcrap" also weakens your arguement against it.
Finally, reading through the name-calling and crap is just noise and totally does not contribute to a worthwhile read.
just my 2 cents.
I really like reading your blog and have learned ALOT from it but the name-calling just detracts from the validity of your arguements.
I guess Intel folks achieved their goal. Any way, the numbers speak for themselves. Opteron smashes Itanium and Woodcrest in FP performance.
My PhD thesis? I doubt anyone at Intel can understand it. Maybe someone at IBM can.
More smoke and mirrors from Sharikou. Seriously, do you even know what PhD qualifiers are Sharikou? Go google that first, then talk about thesis.
So what is the harm in telling us what your PhD qual subjects were? And what U you are from and what area your thesis is in? These are very general questions so you need not worry about personal information.
And at least IBM folks who read this can comment (or are you suggesting IBM folks don't even bother to read your blog, which is why you refuse to tell the others about your thesis?)
"Unless they change their course, you will see Intel BK in 5 to 7 quarters."
You need to update yourself, this prediction was made last quarter (and the 7 quarter prediction I believe was made as early Q1).
I think you mean 4-6 quarters now! (Clock's ticking...)
As for the PhD thing - folks cut him some slack if he/she doesn't want to reveal the info, that is his/her perogative. Although as to the whole privacy/potential ID theft thing, I don't think stating the field his/her PhD is in would be an issue.
"I don't have a degree and I probably never will"
What a freek##* shock!
Look Sharikou, I really like reading your blog and have learned ALOT from it but the name-calling just detracts from the validity of your arguements.
"It's just unprofessional and quite frankly, makes you look bad and perhaps even worse than the intel fanboys.
Using words like "woodcrap" also weakens your arguement against it.
Finally, reading through the name-calling and crap is just noise and totally does not contribute to a worthwhile read.
just my 2 cents. "
That is the kind of language they teach at university of crap.. Graduates: Sharikou!
Well Sharikou, you know all these Intel Fanbois are really bashing you is because they themselves has yet to get hold of any Conroes for their systems.
Otherwise they will bombard you with their own benchmarks against similar AMD systems.
Lets face it, Conroe can be almost considered semi-vapourware. Even the low-end 6300 no one here can get unless they happen to be a IT E-Journalist for a website or magazine. And probrably it will be a B1 stepping chip.
Lets face facts. Anytime right now a person can get an AMD chip that will outperform any of Intel's P4 chips in the same market range in power efficiency, temperature and benchmark scores.
Yet there is any such access to the Conroe. So when it comes down to it for the average Joe that wants to buy a PC now, AMD has won.
Of course you Intel Fanbois can prove me wrong right now by buying straight away one of those P4s and scanning the receipt. But you won't. Because no matter cheap those blazing turds are you don't have enough faith in your company's product to buy one of them for your personal use (as in not to give or make for anyone other than youself).
Basically speaking, Intel Fanbois have a tremandous lack of faith and a fatal streak of egoism in them. You all would sell your parents down the river for a Conroe but you'd not even pass water over a P4.
That's the reason why AMD Supporters are proud of their company and products. We are just as happy with A64s and X2s and will still buy them even if we can't get the Opterons and FXs.
So when you Intel Fanbois can show that much faith and loyalty to Intel then you come here and shake your balls at us.
Sharikou Said: "I was in this meeting with doctoral candidates from other fields, it was something about getting a fellowship from the university. There was some guy who was in a doctor of education program, doing some research on kids. mostly doing some statistics and then waving hands, saying this factor is important. That kinda of stuff. I pretended to show some interest. But the guy told me that the stuff we did was true science, his was just liberal arts stuff..."
Translation: I have no PhD and I will never admit it. The only phD meeting I was in was that of my friend who I freakin envy!
Most of my friends are PhDs, and none of them like to be obssessed with the past. For them and myself, getting a PhD is just natural. You are talking about people who grew up to be the top of their classes. No big deal there...
No big deal? Why do you use the title than?
My PhD thesis? I doubt anyone at Intel can understand it. Maybe someone at IBM can.
Try us.. that is the most imature thing I've heard. Like in grade school when the bigger kids pretended to know something by but not sharing, but in the end they knew nothing more.
Open the robe Sharikou.. are you really so humble. You don't act like that on matters related to INTEL nor AMD. Why with your education/background. It only shows you are a total pretender with no educational or educationl expertise.
I will say Sharikou has NO Doctorate, no higher education. He is a layman with a blog trying to get his little 2 cents of fame. In the end he gets his spotlight, but looks like a fool. Right up there with Paris Hilton.
It is funny to debate a retard like you pushing such a one sided picture of things. There is much that AMD did that was amazing, and much that INTEL did that was stupid. To not recognize the mistakes/weakeness of AMD nor the recent advancemtns by INTEL only solidifies your position as number one one eyed king leading the blind amd fanboys.
You are pure entertainment as to how stupid the AMD fanboy can get deluded by his passion for AMD and hatred for INTEL.
The doctor.
PS.. I don't need to act like a man. I have a PhD from a top 10 EE school in undergraduate and a top 10 in MS and another one for PhD. And was offered jobs by who's who. Be assured I've touched you personally... and all you readers too..
My PhD thesis? I doubt anyone at Intel can understand it. Maybe someone at IBM can.
Try us.. that is the most imature thing I've heard.
Dude, I was making an objective assessment. At the time my papers were written, there were about 8 people doing research in that area in the whole world. Even my advisor didn't know how the calculations were done, but you can check the resulting equation, it was all correct. So he asked me to put an appendix in my thesis with some intermediate steps of the calculation. I used to brag to people about how I solved it, but I was too young then. But I am quite sure it's above the IQ of 99.99% of Intel dudes.
t is funny to debate a retard like you pushing such a one sided picture of things. There is much that AMD did that was amazing, and much that INTEL did that was stupid.
I don't you why you Intel fanbois are like mad dogs. Come on, everyone knows Itanic is crap, and Opteron happens to beat it and Woodcrest hands down.
Most of the Intel 'fanboys' ..I know of are Intel share holders who are holding on to soon to be worthless piece of papers..who barely know to use a computer ..forget about CPU arch..
Most AMD fanboys are hard core hardware addicts ..
Does anyone have a similar observation ??
Why show SpecFP and not SpecFP_rate? Doesn't it seem like the benchmark that actually shows system level performance (sort of) is the more important one, not the one that shows one core?
Just like you'll never bother countering his ideas with hard facts instead of posting your incessant dribblespeak? PLEASE. Until you can prove your points with some cold hard facts of your own, do the rest of us a favor and shut your own pie hole, thanks.
I have many times posted links to items that countered Sharifrauds assertions. However, there is just so much verbal diarrhea here that I can't counter it all. I don't have the kind of time Sharifraud has to do this as I actually have a job and a life and a family.
By the way, GWB did invade Iraq based on cooked intelligence. If you don't believe it, watch the Frontline special "The Dark Side". Chock full of facts, quotes and other misdeeds just like Sharikou's blog. :-)
We will win or we all secede to the eventual brutality of Sharia law, effectively Islamo-fascism.
<<< My own soapbox >>>
WTF??? I don't think anyone actually believes that if we didn't invade Iraq, we would all be wearing burkas and long beards in ten years. How do you reach your concusions? Iraq could have never invaded the west and subjugated use to Sharia law. That is even less likely than Intel going bankrupt in 5 quarters. Iraq was soooo much further from having a nuclear weapon than Iran is now and N. Korea already has one. Iraq didn't even have uranium enrichment capabilities. That was all destroyed by Isreal a long time ago.
<<< end soapbox >>>
Fat Mod Mike says:
I think his title is something that doesn't need to be discussed or hated upon by others who think its necessary. This blog is posting Sharikou's opinion on matters and others respond with their own opinion. Sharikou obviously knows something about computers, you can't make this sh*t up. The people who constantly demand proof of his PhD or anything else need to shut up.
I can't speak for others who question SHarifraud's PhD but if he got his degree in some sort of hard science field it should be apparent that he didn't learn much. One of the first, most basic rules of science is that your facts lead to conclusions not the other way around. For example, Sharifraud says Intel will go bankrupt in 5 quarters with some sort of authority. HOwever , all the facts that he chooses to are cherry picked and based on very big IFs. C'mon even you AMD fanboys, do you really believe Sharifraud when he says Intel will be bankrupt in 5 quarters??? Really??? Think before you drink the Kool-aid.
You Intel kids are allowed to run free here. Everytime you run out of ammo, you start digging mud on personal matters. I have presented solid benchmarks on SPEC.org showing Opteron frags the fastest/largest cache dual core Itanium 2 on both floating point and integer performance. You Intel kids can't accept the numbers?
Plenty of benchmarks out there that show that Conroe is kicking ass on just about any AMD desktop CPU. I'm sure you have seen them so I'm not going to link. So what does that mean. AMD has a faster chip than Itanium. Great. Very low volume market. Intel has a faster chip than AMD now in a VERY, VERY high volume market. Also, Intel's mobile parts are generally better than AMD's mobile parts. The real bottom line is this: Intel Q1 income applicable to common shares $1.3 Billion. AMD Q1 income: $184 million. Roughly 1/8th of Intel. I believe this is what happens when the market speaks. Remember kids, it isn't who can design the fastest chip, you also need to be able to make and sell enough of them to show profit. AMD seems to barely be learning this but are going backwards fast since their Q2 earnings were a measly $88 million. Hell Intel's tax bill is more than that. So, as you can see it doesn't really matter that AMD has one or two faster chips. That isn't always how the game is played. It is really going to suck for AMD now that Conroe, Merom and Woodcrest are the new sherrifs in town... You AMD fanboys can cry all you want about how unfair that sounds but in the end when you cry, you cry alone. When you laugh the world laughs with you. Right now, I'm laughing at AMD. (What WERE they smoking; borrowing $2.5 to buy ATI?) Cheers!
"I will say Sharikou has NO Doctorate, no higher education. He is a layman with a blog trying to get his little 2 cents of fame. In the end he gets his spotlight, but looks like a fool. Right up there with Paris Hilton."
I've been reading Sharikou's blog for a couple months. There have been plenty of technical information and insightful analysis. Please, layman cannot do that.
It takes an intelligent person to see others intelligence. If your IQ is too low, you are going to have a hard time to keep up Sharikou's materials. So please, stay off his materials.
-----------
On a different point, I am not impressed with the name calling "Woodcrap". This processor is definitely crap. Could we just simply refer to it as Woodcrest?
-Longan-
Why should a Man have to say he has or hasn't got a PH.D.
It doesn't matter what your qualifications are just how your knowledge was obtained and how accurate it is.
You could go to University, become a Doctor of Philosophy, but I'd still know what 1 + 1 is :P
Stop worrying about who he is. That's not why he's started this blog. He's presenting information to the public for free and you either add to it, dispute it etc and debate it.
So we all learn something.
It doesn't matter whether or not if he's wrong. We help each other out don't we?
You people need real lives, go and get a Girlfriend or something; and a nice one at that.
So you say he's fake? Who cares, just read it, take it with a pinch of salt, or analyse it further.
He's a bit biased? So what? Even more reason to disprove him isn't it?
So far no one cause caught him with his pants down. Sometimes he jokes around by posting rediculous stuff, which he himself knows is biased, but it's comical isn't it? We all have a laugh at it too.
But by all means, if wasting time is what makes you feel better in your lives, go for it guys.
The World is run by the media and alike, you're all a bunch of monkies. Then when someone posts something different or (in this case) states out the obvious, people point fingers and are suddenly afraid that everything they learnt was wrong.
Wake up Guys.
you either add to it, dispute it etc and debate it.
But the problem is: the Intelers can't dispute it--it's all backed up by SPEC numbers. What can they do? I allow their posts to people can see the way they sling mud...
Some thoughts for you fanboys...
Who has 8 billion in the bank?
Who has 49% gross margin, damm mad and greedy and working hard to get it back to 55% +
Which company has leadership manufacturing, high yields, and volume/cost to supply 200+ million CPUs and supporting Northbridges/southbridges.
Which company is the most profitable semiconductor company of the past 10 years and continues to be?
Which company plows more back into basic silicon R&D into delivering the leading edge manufacturing and technology earlier and better than anyother semiconductor company?
Which company has < billion in the bank
Which company has had its debt downgraded
Which company has not made two a profit for two consecutive years
Which company has ot enter into collective development agreement for silicion develompent and 2nd tier foundrys to make their chips because they can't afford the factories themselves.
Who just borrowed 2.5 Billion and also believe they can fund 65nm build/ramp and 45nm development?
Who is going to go bankrupt?
One of the first, most basic rules of science is that your facts lead to conclusions not the other way around. For example, Sharifraud says Intel will go bankrupt in 5 quarters with some sort of authority.
The most important aspect of science is the ability to predict the outcome given a set of initial conditions. I have predicted many things, most of which have come true. My prediction on Intel's BK time has yet to be proven true. You just have to wait and see. Furthermore, I gave an alternative solution to Intel (instead of going Chapter 11): ramp down, downsize to 30-40% of current size, and play dead and conserve resources.
Doc, have you ever been contacted by Intel legal about possible slander charges? If not, you should expect to soon.
SUBJECT; “Operon cleankills dual core Montecito” can anyone disprove the facts from the blog? Example “Opteron is 16% faster than Woodcrest and 17% faster than Montecito.”
ANSWER; no
INTEL FANBOY SOLUTION; attack the messenger forget about the message.
I mean, most of the consumer market can identify Intel and has some sort of confidence in the brand. AMD is not so fortunate and cannot hope to achieve this same comfort - in the near future anyways.
Brand alone could keep Intel going for years, even if the corporation was helmed by chimpanzees.
Ever heard of Hayes?
Hehe... I was going to give Graham the "Godwin's Law" award (for his insightful comparison between Sharikou and GWB), but I feel that there are too many people here that have done the same, in some way or another...
http://en.wikipedia.org/wiki/Godwin's_law
Congratulations Intel fanboys, keep it up… Keep comparing Sharikou to anyone you feel like. Because every time you do, you continue to prove that you have lost the argument. Sharikou was right to say that Intel fanboys are complete retards. Your bickering only helps validate his point(s).
"It seems that one of the AMD systems has about 50% more performance than the others. That X4600 seems to be the only one that has so high score. I wonder what they did with that system to get it working so efficiently and why haven't others done that."
the reason why x4600 shows so big numbers is because sun compiler used in this test, sun studio 11 for solaris, is able to parallel some spec tests on several cpus. its not prohibited by rules and in some sense can be used in real world apps sometimes, but strictly speaking its not result of single opteron core. its result of system, based on opteron and (this is very important!) specific os and compiler.
both itanium and woodcrest overperform opteron per core, but they are not really compete with each other. woodcrest is 2P only, opteron 2p-8p and itanium scales way to longer than 8P.
Let me add something to this already huge discussion:
1) baseline results are much closer to real world, peak results only show compiler capabilities
2) SPECfp (not rate) will benefit from multiple CPUs and cores, result for 4 cores will be better than for 1 core. Thus 4 CPU Opteron vs 1 core Xeon is not fair.
3) Sun results are obtained under Solaris and using Sun compiler, and is just a result of successful tweak for just 1 subtest (179.art)
It's my impression or no one here know the Spec benchmark?
There are two scores, the base and peak, the base is the score some machine get with default compile, the peak is the score some machine get with custom compiler flags,
And there are two other scores... One is for single-thread, in this case a single-core chip rocks and you can copare 2P to 8P, the other is the multi-thread (identified by the "rate"), in this one mult-processors rocks and you can't compare 2P to 8P.
Ever heard of Hayes?
Nope, must not have had the brand power that Intel does.
"I really like reading your blog and have learned ALOT from it but the name-calling just detracts from the validity of your arguements.
I guess Intel folks achieved their goal. Any way, the numbers speak for themselves. Opteron smashes Itanium and Woodcrest in FP"
I was not referring to any comparison in particular.
All I meant was that your arguements are far stronger when you refrain from name-calling.
This has nothing to do with Intel folks achieving their goal.
In fact, if intel fanboys sling mud at you and you do not resort to name calling, your arguements actually get stronger since you don't need to resort to that level to make a point.
6 months later AMD will be better for afew months, then intel, then amd, then intel.
With K8, AMD maintained the performance lead on desktop for three years, and it still has the performance lead on server. It will take more than 3 years for Intel catch up with K8L.
Conroe/Core2 will be forgotten soon, even before it's ramped up.
"When you can buy a X2 3800+ for $150 compared to a E6300 for $220 bling bling, the price is really performance wise just about even."
I think X2 3800+ outperforms E6300 in most benchmarks.
"In the real world Conore at best may feel just abit faster but won't make you say wow or anything special. Its only about a 10% difference really."
You won't convince Intel fanboys with such statement, really, because they'll tell you that Conroe computers SuperPi so much faster than K8. Also IIRC Conroe encodes DivX video quite a bit faster than K8, too.
Of course, benchmarking is just not the only thing that matters. If I play games, without SLI/Crossfire and two very high-end graphics card, Conroe won't give me any advantage; if I run spreadsheets or Java, memory bandwidth might be more important and thus gives K8 bettr performance.
I agree with the other points you mentioned above, though.
I think X2 3800+ outperforms E6300 in most benchmarks.
oops! i was wrong. ah... it's too early... where's my coffee...??
It will take more than 3 years for Intel catch up with K8L.
As you know that much about K8L, could you make a list of features that will give K8L a major advantage over Core2? I've asked it from couple of other places too but so far I haven't got an ansver. I've studied the slides and core images and compared them with other architectures but haven't found anything. There must be something I'm missing.
Sorry Sharikou, I apologize for the "offtopicness" but some clueless people need to get schooled and get a history lesson.
Congradulations one again, as already stated for proving the Godwin Theory right Graham, you get a gold star for that. *sarcasm* Time to wrap this up though.
"I have many times posted links to items that countered Sharifrauds assertions. However, there is just so much verbal diarrhea here that I can't counter it all. I don't have the kind of time Sharifraud has to do this as I actually have a job and a life and a family."
Haha, okay, I have yet to see any links from you, feel free to post them at your lesiure. So far I have yet to see any from you providing concrete rebuttal to most of Sharikou's claims. Care to try again?
"By the way, GWB did invade Iraq based on cooked intelligence. If you don't believe it, watch the Frontline special "The Dark Side". Chock full of facts, quotes and other misdeeds just like Sharikou's blog. :-)"
Frontline??? HAHAHAHA! You are a fool if you believe anything the Establishment Media has to spout out as being "FACT". I don't trust CNN, CBS, NBC or even Fox for "facts"! You are no more "enlightened" than the conspiracy theorists OR the radical idealists/conservatives during Clinton's "reign".
If you were paying ANY attention to the news during the first Gulf War and up to current events you will realize that the UN was worthless in implementing it's own policies against Iraq, that the US and UN gave Saddam over 10 years since the first Gulf War ceasefire to comply with UN mandates. Saddam NEVER tried to comply with any of the UN mandates that he could get away with, Coalition forces found an entire jet buried in the desert after the invasion for Pete's sake, who knows what the hell else is buried out there! You are making a very black and white clueless observation about a very gray area situation.
Again for all the "Iraq liberal mantra" spouting individuals, Iraq isn't about oil, this wasn't entirely about terrorists, this wasn't about rebuilding Iraq for it's people, this was about Saddam/Iraq not complying with UN backed sanctions that they VIOLATED SINCE THE FIRST GULF WAR CEASE-FIRE STIPULATIONS INSTITUED BY THE UN AT THE CLOSE OF THE FIRST GULF WAR. This means, the war continued in 2002, THIS IS A COLD, HARD UNDENIABLE FACT! Spare me your excuses and fabications of GWB invading on "cooked" evidence, plain and simple, Saddam did not comply with the UN sanctions so the US and Coalitions forces resume the invasion of Iraq started by GHB in the first Gulf War. Everything else surrounding the reasons for the second Gulf War is very moot, although some will never admit or want believe it.
"WTF??? I don't think anyone actually believes that if we didn't invade Iraq, we would all be wearing burkas and long beards in ten years. How do you reach your concusions? Iraq could have never invaded the west and subjugated use to Sharia law. That is even less likely than Intel going bankrupt in 5 quarters. Iraq was soooo much further from having a nuclear weapon than Iran is now and N. Korea already has one. Iraq didn't even have uranium enrichment capabilities. That was all destroyed by Isreal a long time ago.
<<< end soapbox >>>"
GEEEZZZZ. Talk about not seeing the forest through the trees. I am going to explain this all so hopefully you can comprehend:
THE US MILITARY IN IRAQ IS TO CREATE A FOCAL POINT AGAINST THE ISLAMIC MILITANTS IN THE MIDDLE EAST.
It's called bring the war to them and keeping them off US soil. This isn't even about IRAQ anymore it's about IRAN, SYRIA, LIBYA, LEBANON, AFGANISTAN and any other ME country that harbors Islamic militiants and drawing them out to take them down, plain and simple.
You are simply CLUELESS to how vicious these people are and to what lengths they will go to see thier cause fulfilled. Ever heard of a suitcase NUKE? This is one of Al Qaeda's main goals to create! All it is going to take is for one of them to be set off by an Islamic militiant in the mainland of a country and instantly, that country will be brought to it's knees, technologically, militarily and economically.(or with countries small enough like Isreal, wipe them off the map) With a massive enough Electromagnetic Pulse blast,(save the 100 mile radius or more of the nuclear blast destroying anything in it's range) anything using electronics within a EMP's effective range (1000 miles or more) will be destroyed and immediately a country will be thrown back into near stone age technology. Now does that sound so crazy to YOU??? Do you realize what kind of threat these people are NOW??? Do you realize you are rationalizing radical fundamentalist Islamics behaviour when they saw heads off innocent men, pose as civilians while firing at Coalition forces and endangering innocent bystanders, telling us in taped recordings to submit to Islam or DIE???
Sorry Graham but you are truly clueless if you believe they won't all die trying to establish a Islamo-fascist state into a reality. It's kill or be killed, they have given us no choice. Usually I don't care to get so long winded about things, but do I pray that somehow this point gets into yours and any other thick headed individual's skull about this because it concerns and involves us all. It will decide our future depending on the steps we take now. My advice to you is go back and go learn about actual non-revisionist history or stay out of the way and shutup while others protect your freedoms to voice your inaccurate objections.
As you know that much about K8L, could you make a list of features that will give K8L a major advantage
CORE2 is irrelevant. Compared to K8, K8L will offer 50% better integer performance and 3x floating point performance per core. Scalability wise, K8L will have 5.2GT ccHT links.
Ever heard of Hayes?
Nope, must not have had the brand power that Intel does.
sarcasm on: Must have not had the brand power that Intel does.
Hayes was THE modem until U.S. Robotics started getting a foothold challenging Hayes' invention by improving their performance features over Hayes but this is beside the point since they only did that near the end of Hayes' life. Hayes chose the wrong technology to replace its modem business and also the wrong strategy which led to its disappearance.
Intel suffering the same is not inplausible. Intel is trying to lock out others...look at what happened to Apple vs Wintel. Apple almost disappeared. Apple did not want to make friends with others and so they have to be satisfied with a closed niche market.
There were times when nobody would imagine Hayes or Apple disappearing. They looked so permanent. Well Hayes did disappear and Apple narrowly missed the same fate. What makes you think that Intel will be exempt if they also make similar mistakes?
Compared to K8, K8L will offer 50% better integer performance and 3x floating point performance per core. Scalability wise,
IIRC, Core2 had exactly the same number of ALU's :)
Unless you think K8L has much better ALU efficiency I see no reason why should K8L be much (any) faster than Core2 in per-core performance. So far I haven't heard that general-purpouse SIMD units have got much faster.
K8L will have 5.2GT ccHT links.
Of cource Intel has nothing except 1333 or perhaps even 1666MHz FSB to put agains thatI I won't argue against that AMD will keep and extend it's scalability compared to Intel.
Of cource that ccHT helps only with 2P+ systems with software that needs a lot of syncing.
MadModMike said it will be 30-60% faster in per-core performance. Do you agree with that?
Hayes was THE modem until U.S. Robotics started getting a foothold challenging Hayes' invention by improving their performance features over Hayes but this is beside the point since they only did that near the end of Hayes' life. Hayes chose the wrong technology to replace its modem business and also the wrong strategy which led to its disappearance.
Intel suffering the same is not inplausible. Intel is trying to lock out others...look at what happened to Apple vs Wintel. Apple almost disappeared. Apple did not want to make friends with others and so they have to be satisfied with a closed niche market.
There were times when nobody would imagine Hayes or Apple disappearing. They looked so permanent. Well Hayes did disappear and Apple narrowly missed the same fate. What makes you think that Intel will be exempt if they also make similar mistakes?
Oh THAT Hayes!
Wow, what a memory refresher. Back in the bbs days (I used to run one) I remember Hayes being popular. I just checked around, they're still alive actually, but bought out back in the 90's.
If I recall, there was a division over protocols and USR's protocol finally won out - much like is happening today with high def DVD.
This was in the days before operating systems made it possible for incompatible communications hardware to talk to each other.
Software designed for current processors will linger on for eons, long after multiple processors are commonplace in the home PC. Therefore Intel's old stock will remain useful long into the future, regardless. This makes it less likely that Intel will go belly up. As long as it can keep up to or surpass AMD's offerings LEGALLY (!) Intel will be around.
I still can't see how Intel will fail in a year or two or five. As long as they keep plugging away they will always be around.
True, they are often humbled by AMD offerings, but so is AMD sometimes.
"Software designed for current processors will linger on for eons, long after multiple processors are commonplace in the home PC. Therefore Intel's old stock will remain useful long into the future, regardless"
This I have to disagree.
An Intel CPU is only useful if you have a matching motherboard, good RAM, good video cards and even power supply to work with. Those things change. The old power supply and RAM from my old Athlon and P-II are different, and those corresponding chips, no matter how capable they are, are essentially useless to all but the few determined people or special market.
Anyone who owns an old machine knows to pray that none of the component breaks; because once it does, the whole thing becomes junk (unless he's so lucky to bump into a replacement of that component).
Software designed for current processors will linger on for eons, long after multiple processors are commonplace in the home PC. Therefore Intel's old stock will remain useful long into the future, regardless
While true, in order to stay alive, any company needs to seel products. Just having old hardware lingering around isn't enough to stop bankrupcy. Since hardware (and even software) gets outpaced so quickly, you really need to stay on top of your game the whole time.
Secondly, all of that software for intel CPUs will run on AMD (or any other upstart x86 CPU maker) processors as well. If the alternative is obvious, there really isn't necessarily a need to buy intel products, if things went that way.
Sharikou, why should anyone trust your twisted 4P/2P comparisons for SPECfp_rate?
The best 2P SPECfp_rate2000 publications are 85 for Opteron and 85.9 for Woodcrest. Woodcrest wins in 2P. I dare you to produce different 2P numbers.
BTW, Opteron's 85 score benefits heavily from Sun Studio11. Without it, Opteron gets only 77.6.
"The best 2P SPECfp_rate2000 publications are 85 for Opteron and 85.9 for Woodcrest. Woodcrest wins in 2P. I dare you to produce different 2P numbers."
The fastest Opteron boxes aren't 2P.
BTW, Opteron's 85 score benefits heavily from Sun Studio11
Unfortunately nobody has put up a Woodcrest score without the Intel compiler, but most likely its the same deal.
In a perfect world everyone would post benchmarks with the latest version of gcc (4.1.1 right now), but for now we have to settle with some compilers being better than others.
price_4p = 10 x price_2p
unit_hp = unit_dell
revenue_hp = 3x unit_dell
Isn't that 3 equations and 5 unknowns??
Hint: Revenue = Price x Unit
I am not sure about this, but ...
CFP2000 rate shows Itanium is 34% faster than Opteron for the following case:
SPECfp_rate2000 = 182
SPECfp_rate_base2000 = 167
(Sun Blade X8400 Opteron 885 2.6GHz 8 cores, 2 cores/chip)
link here
SPECfp_rate_base2000 = 244
(SGI Altix 4700 Itanium2 9050 1.6GHz/24M 8 cores, 2 cores/chip)
link here
SPECfp_rate_base2000 = 244
(SGI Altix 4700 Itanium2 9050 1.6GHz/24M 8 cores, 2 cores/chip)
Well. I don't trust any data from a bankrupted company. If a company can't even balance its checkbook and end up BKed, how can we trust its SPEC scores?
I found that SGI's stock is trading on OTC market at $0.03 per share (3 cents). I guess you can own 100% of AGI if you want.
So the question is: why hasn't Intel bought SGI?
price_4p = 10 x price_2p
unit_hp = unit_dell
revenue_hp = 3x unit_dell
Isn't that 3 equations and 5 unknowns??
Hint: Revenue = Price x Unit
It is still 4 equations and 5 unknowns
Hint: To solve uniquely for n unknowns you need n equations.
Let us get the objective facts right.
"In terms of floating point performance, Opteron ruins Itanium 2. the SUN x4600 got a SpecFP_2000 score of 3538. A Itanium 2 9050 got 3017. A 2.2GHZ Power 5+ got 3513."
The x4600 was running the parallelized version of the benchmark, on 4 Opteron 856.
Parallel: Yes
Shell Environments:
Stack size set to unlimited via "ulimit -s unlimited"
export PARALLEL=4
The Bull and IBM were running only a single thread. So we can only compare between the Bull and IBM, and the POWER5+ is faster than the Montecito.
We should always check whether Parallel: Yes or No in the hardware software information.
I should this clears the confusion.
Chin Gim Leong
Yap! Few of the people are right! Why you compare the data with 1.6 GHz Itanium 2 and a 2.6 Opteron? And the result is the one directly from AMDs site, plus the Itanium is not design to use under Bull advanced server, me personally as Dell system engineer, we got the real chance to test those system out, and yes, we do a private testing using Intel and AMD processor, not disclosed fact, Dell do those testing to determine if we are going to ALSO use the AMD chip for our high end server market. Different OS, different clock speed, different OS can make allot of difference! You stated you are Ph.D, if that IS true, you should read the benchmark content more carefully next time when you post, because that makes you look very bad to the people who actually know the system inside and out.
Post a Comment
<< Home