They was close...only 50% wrong...
They was close...only 50% wrong...
Crunching for Comrades and the Common good of the People.
Well, he said "many workloads". And for multiple socket systems, Barcelona's do offer higher performance versus an equivalent intel xeon system, at least that's what the benchmarks showed. Or maybe he was only referring to specific SPEC benchmarks.Originally Posted by Periander6
The exact quote, from Randy Allen was:
"Barcelona is going to be able to deliver performance levels for many workloads in excess of 40% beyond what Clovertown can."
http://www.youtube.com/watch?v=G_n3wvsfq4Y
A pretty comical video in hindsight.
SpecFp_rates and stream bench...real world was a whole other story where the marketing bluff cracked.
Crunching for Comrades and the Common good of the People.
CnQ is buged, disable it for bench people.
You lost a lot of power with enabled, the cpu stay at 1.1ghz on single thread.
Agena has 10-15% faster IPC from a kentfield
Proof? Just like your claims of CF having 90% scaling, which you failed to show any proof whatsoever, I call you out again on BS.
yea, pretty annoying. It's their job to spin it positively, not much of a defense, a lame excuse.
The only one you are fooling is yourself...
Crunching for Comrades and the Common good of the People.
AT numbers for 2P comparison:
Considering that AMD prices this Opteron 2350 under the Xeon 5345, AMD has an attractive price/performance offering for most applications. The only exception is a chess engine and highly optimized Intel binaries. Although our testing is not finished yet, there is very little doubt that AMD's newest chip is very energy efficient. Add to that the fact that the AMD platform is not burdened with the extra power consumption of FB-DIMMs, and it is clear that the third generation of Opterons will lead in the performance/watt area for a few months. When you are looking for the highest performance however, Intel has still a solid advantage with it's 3 GHz Xeon x5365
The future looks very interesting with the 45nm Xeon Harpertown and a 2.5GHz AMD quad-core in the next quarter. AMD hasn't clearly hit a homerun this time, but at least they're playing in the same ballpark.
Why would CPU-Z not show that though?
Just a shame they disabled hardware prefetching and forgot to list the disk systems used. Specially when most of them are heavy disk I/O based.
Nice try, no saviour.
Also doing scaling calculation is not the right way. Specially since it looks like they asume 100% scaling..and thats never true.
Crunching for Comrades and the Common good of the People.
If someone tells you something, swears it's the truth, then later is proven wrong, then what's that saying about them? One of two things; they just didn't know any better or they were trying to willfully deceive people.
IMHO I believe AMD was being deceitful to keep their share prices up and avoid panic.
Ah lol,now even Anand is no good
First ,you should read up,the hardware prefetching is disabled since it is recommended by intel as it was found that it negatively affects the performance (hence they disabled it ,duh..)
Second,yeah,now disks are the problem.Right.It is logical that Intel must have provided Anandtech with the worst possible hard drive setup in their systems,since they knew they would only be compared to the latest AMD's quad core based server
As for scaling,they have numbers for 2Ghz AMD 2P system and 2.33Ghz 2P intel system.No brainer.It could only get worse for intel since Clowertowns doesn't scale linearly with freq.(fsb bottleneck).And the lowest number for K10 in the table is 1.8Ghz,so it was pretty much constant(they didn't have to presume scaling up for it since it was faster clock/clock in 2P most of the time).
Last edited by informal; 11-20-2007 at 04:33 PM.
I think you are dreaming again. Also if you ever bought an OEM server you would know its on per default for the same reason.Ah lol,now even Anand is no good
First ,you should read up,the hardware prefetching is disabled since it is recommended by intel as it was found that it negatively affects the performance (hence they disabled it ,duh..)
Second,yeah,now disks are the problem.Right.It is logical that Intel must have provided Anandtech with the worst possible hard drive setup in their systems,since they knew they would only be compared to the latest AMD's quad core based server
As for scaling,they have numbers for 2Ghz AMD 2P system and 2.33Ghz 2P intel system.No brainer.It could only get worse for intel since Clowertowns doesn't scale linearly with freq.(fsb bottleneck).And the lowest number for K10 in the table is 1.8Ghz,so it was pretty much constant(they didn't have to presume scaling up for it since it was faster clock/clock in 2P most of the time).
You do know much big difference there can be and is on servers diskssystem, right? No..I guess so.
Again you asume scaling is linear. Yet it aint. They could have used E5335, yet they didnt.
So you are even counterargueing youself? Since the result is based on a 2.33Ghz and not a 2Ghz? Bravo...bravo!It could only get worse for intel since Clowertowns doesn't scale linearly with freq.(fsb bottleneck).
You have been wrong since day one on K10, and you still are.
Crunching for Comrades and the Common good of the People.
Well, that particular review was junk given the choice of benchmarks which was unrepresentative of server applications.Ah lol,now even Anand is no good
It's also pretty amazing that two months after the launch and AMD still hasn't been able to release a single official score for important enterprise benchmarks like TPC-C, SPECjbb2005 or SAP-SD. AMD's silence says everything you need to know about Barcelona's lack of performance.
No they didn't they said in; "across a wide variety of workloads". What and where I ask? I saw NOTHING of the sort LOL!amd never said that it would be 40% faster overall................
http://www.news.com/AMD-Go-to-Barcel...3-6152645.html
That was parroted here by many on the Green team and asked loudly, "surely AMD wouldn't lie that much?" Ask Informal? The whole epilogue was to insert FUD and have shoppers wait or delay buying from Intel when AMD knew they didn't have a damned thing.When it comes to quad-core chips, good things come to those who wait, Advanced Micro Devices believes.
After years breathing AMD's dust, Intel beat its rival to the punch by releasing its quad-core Xeon 5300 "Clovertown" processor for servers in November. But AMD's "Barcelona" quad-core chip, due to arrive midway through 2007, will be a significant notch faster than the Clovertown chips expected to be on the market at that time, said Randy Allen, AMD's corporate vice president for server and workstation products.
"We expect across a wide variety of workloads for Barcelona to outperform Clovertown by 40 percent," Allen said. The quad-core chip also will outperform AMD's current dual-core Opterons on "floating point" mathematical calculations by a factor of 3.6 at the same clock rate, he said
Posted by duploxxx
I am sure JF is relaxed and smiling these days with there intended launch schedule. SNB Xeon servers on the other hand....
Posted by gallag
there yo go bringing intel into a amd thread again lol, if that was someone droping a dig at amd you would be crying like a girl.qft!
So what did you think got Enron in trouble? You can't do that in the Business world and get away with it. But that wasn't the point. Many folks saw through the FUD, lies and etc... then were flame just for pointing out just what we all KNOW now.
Posted by duploxxx
I am sure JF is relaxed and smiling these days with there intended launch schedule. SNB Xeon servers on the other hand....
Posted by gallag
there yo go bringing intel into a amd thread again lol, if that was someone droping a dig at amd you would be crying like a girl.
Don't you mean Front Side Buses?Ah lol,now even Anand is no good
First ,you should read up,the hardware prefetching is disabled since it is recommended by intel as it was found that it negatively affects the performance (hence they disabled it ,duh..)
Second,yeah,now disks are the problem.Right.It is logical that Intel must have provided Anandtech with the worst possible hard drive setup in their systems,since they knew they would only be compared to the latest AMD's quad core based server
As for scaling,they have numbers for 2Ghz AMD 2P system and 2.33Ghz 2P intel system.No brainer.It could only get worse for intel since Clowertowns doesn't scale linearly with freq.(fsb bottleneck).And the lowest number for K10 in the table is 1.8Ghz,so it was pretty much constant(they didn't have to presume scaling up for it since it was faster clock/clock in 2P most of the time).
Posted by duploxxx
I am sure JF is relaxed and smiling these days with there intended launch schedule. SNB Xeon servers on the other hand....
Posted by gallag
there yo go bringing intel into a amd thread again lol, if that was someone droping a dig at amd you would be crying like a girl.
QFT!So what did you think got Enron in trouble? You can't do that in the Business world and get away with it. But that wasn't the point. Many folks saw through the FUD, lies and etc... then were flame just for pointing out just what we all KNOW now.
You mean performance. Because cnq enable is suppose to consume less power.
DO you realize the shrug emoticon you use very frequently at the end of your statements implies you don't know for sure what you're talking about? It confuses me so much as to if you're asking something, guessing something based on anecdotes or trying to affirm a fact. What is it supposed to mean for you?You mean performance. Because cnq enable is suppose to consume less power.
A couple of sites made note that AMD would need to be ahead of Intel 200 MHz or so to achieve parity .. this is a bit off as the relative % compared to IPC is lower at higher clocks. For AMD to be even with a 3.0 GHz Yorksfield, they would need about 3.3 or slightly higher bin. This is not likely going to happen on their 65 nm process.
Well, AMD worded their 40% statement very carefully, at the time they made their claim, the mean 40% faster using SPEC2006FP_rates. They did not clarify if they meant clock for clock and such... they also stated in a wide range of applications, and the SPEC2006FP bench has 17 different code bases in it's suite, so one could argue that is a wide range.
Where they missed was their clock speed estimate, 2.0 GHz launch speed is significantly slower than what they planned and they had to revise their expectation downward as a result, recall Hector Ruiz downplaying the product.
Jack
Anyone else think AMD just got this into the market so that its out there so they can start working on the 45nm process instead of trying to tweak this thing?
Let's hope 45nm goes better than 65nm has! I totally agree with your statement. This chip looks more like it was meant to please shareholders than anything else.
slightly offtopic
@informal did you consider that harpertown is (almost) available and penryn's server variant is much better than core2 in server workloads because of the the 1600fsb (up to 20% per clock improvement in memory intensive apps). thus barcelona doesn't look so good anymore even for servers...
edit: I don't want to go further offtopic on Barc vs. Harpertown. So I just want to add; I don't know much about the availability of serverparts here in Europe as I'm not purchasing any. Maybe they are widely available or not..
Last edited by Jacky; 11-21-2007 at 11:46 AM.
Posting Permissions
Bookmarks