AT's Phenom review up (11/19)

[BitByBit]

Phenom's performance is certainly something of a mixed bag. Intel's own benchmarks put Barcelona well ahead of Xeon in HPC and Floating point performance, per clock, which is in direct contrast to performance on the desktop.

I'm inclined to agree with other posters here that Anand's Phenom review was severely lacking in depth. I for one would love to know why there is such a huge disparity between server and desktop performance.

 

[Accord99]

Originally posted by: BitByBit
Phenom's performance is certainly something of a mixed bag. Intel's own benchmarks put Barcelona well ahead of Xeon in HPC and Floating point performance, per clock, which is in direct contrast to performance on the desktop.

Those are the niche areas where the superior memory bandwidth of the Opteron platform comes into play.

I'm inclined to agree with other posters here that Anand's Phenom review was severely lacking in depth. I for one would love to know why there is such a huge disparity between server and desktop performance.

The main advantages of Barcelona in Opteron form is more memory bandwidth and lower latency, which allows it to win a few benchmarks. On the desktop, this memory bandwidth and latency advantage is far smaller. Plus, few desktop applications actually are memory bandwidth limited and most are more dependent on integer performance.

 

[gOJDO]

@Gary Key

What happened to the 2.4GHz Phenom which was supposed to have higher IPC than the lower clocked Phenoms?

 

[harpoon84]

Originally posted by: gOJDO
@Gary Key

What happened to the 2.4GHz Phenom which was supposed to have higher IPC than the lower clocked Phenoms?

Huh? Thats news to me!

 

[AlabamaCajun]

The 2.4G 9700 that was supposed to be released is being held back until a BIOS patch is available. Some of the steering logic can lock up with 4 cores running full saturation in the 2.4G and higher bins. The change will see to it that the process management steers clear of the situation. No doubt this will delay newer steppings.

 

[AlabamaCajun]

Originally posted by: myocardia

Originally posted by: AlabamaCajun
[sarcasm]Where was the Phenom article that Lal Shrimpi wrote it was all conroe and penryn.[/sarcasm]
Besides the fact that Intel pays him more or has better wine and cheese parties or beer and pizza, couldn't we have just the info on Phenom and leave all the effing Intel crap for the benchmarks and closing. [blunt]That review sucked[/blunt].

Yeah, you're right, you didn't actually call him a name, you merely implied that he's a liar. That more than makes up for your ignorant rant, huh?

Barcelona though less glorious in speed has a better power factor than it's competition with the exception of the low power xeons that might use less total RMS power.

So you not only can't read articles, you also can't read graphs?

Icall this reply a personal user attack, don't escalate it or one or more of us may see posting vacations.

Liar was never mentioned nor implied. The closest I came was calling it bad journalism. Someone can give a bad review without lying. I had a lot of respect for your posts in the past but that like AMD reign has passed.

 

[Sylvanas]

Originally posted by: AlabamaCajun

Originally posted by: myocardia

Originally posted by: AlabamaCajun
[sarcasm]Where was the Phenom article that Lal Shrimpi wrote it was all conroe and penryn.[/sarcasm]
Besides the fact that Intel pays him more or has better wine and cheese parties or beer and pizza, couldn't we have just the info on Phenom and leave all the effing Intel crap for the benchmarks and closing. [blunt]That review sucked[/blunt].

Yeah, you're right, you didn't actually call him a name, you merely implied that he's a liar. That more than makes up for your ignorant rant, huh?

Barcelona though less glorious in speed has a better power factor than it's competition with the exception of the low power xeons that might use less total RMS power.

So you not only can't read articles, you also can't read graphs?

Icall this reply a personal user attack, don't escalate it or one or more of us may see posting vacations.

Liar was never mentioned nor implied. The closest I came was calling it bad journalism. Someone can give a bad review without lying. I had a lot of respect for your posts in the past but that like AMD reign has passed.

No, you certainly came off as if to imply hes a liar no need to change your attitude now to play innocent. Judging by the 3 AMD rigs in your sig I'd say you are disappointed with Phenom and took it out on the reviewer as was hinted at earlier in the thread.

 

[SickBeast]

People complaining about an intel-bias here are silly. AT was reccomending the A64 when it was in its prime.

The Phenom simply makes no sense in comparison with a Q6600. It's slower per clock, more expensive, uses more power, *and* can't reach the same clockspeeds. Hopefully things will get better once the platform matures a little.

Also, to those that say that there has been a lack of AT articles about AMD products: What were they supposed to write about? Prior to this past week, they have released *nothing* of substance for months. Do you expect them to waste their time writing articles about obsolete A64 boards that went into production 6 months ago?

 

[rchiu]

Originally posted by: AlabamaCajun

Originally posted by: myocardia

Originally posted by: AlabamaCajun
[sarcasm]Where was the Phenom article that Lal Shrimpi wrote it was all conroe and penryn.[/sarcasm]
Besides the fact that Intel pays him more or has better wine and cheese parties or beer and pizza, couldn't we have just the info on Phenom and leave all the effing Intel crap for the benchmarks and closing. [blunt]That review sucked[/blunt].

Yeah, you're right, you didn't actually call him a name, you merely implied that he's a liar. That more than makes up for your ignorant rant, huh?

Barcelona though less glorious in speed has a better power factor than it's competition with the exception of the low power xeons that might use less total RMS power.

So you not only can't read articles, you also can't read graphs?

Icall this reply a personal user attack, don't escalate it or one or more of us may see posting vacations.

Liar was never mentioned nor implied. The closest I came was calling it bad journalism. Someone can give a bad review without lying. I had a lot of respect for your posts in the past but that like AMD reign has passed.

You come into this forum and insult the founder by implying that he was paid off by Intel to write something bad about AMD, and you complain about people's reply as personal attack? What's wrong with Anand talking about conroe and penryn? You may want to close your eyes and ignore the fact that Phenom is not up to par compared to Conroe and Penryn, but for the rest of us, how Phenom perform relative to Intel's offering is the key to our decision whether to buy it or not.

I realize the Internet and free speech go hand in hand, but at least show some respect for people who put a lot of work to get this forum to where it is today, and spend lots of time and effort to write reviews for us to read without charging a cent.

 

[myocardia]

Originally posted by: AlabamaCajun
Icall this reply a personal user attack, don't escalate it or one or more of us may see posting vacations.

Liar was never mentioned nor implied. The closest I came was calling it bad journalism. Someone can give a bad review without lying. I had a lot of respect for your posts in the past but that like AMD reign has passed.

I honestly could give a shit, whether or not you respect my posts. What I do care about is someone implying that one the most respected computer hardware testers/journalists around is being paid to write bad reviews, just because said complainee doesn't like hearing the truth. Now, whether you're intelligent enough to be able to comprehend that is another story. And please, report this to an administrator, preferably Anand himself. I'm willing to bet that if one of us gets a vacation, that someone won't be me.

 

[AlabamaCajun]

Nothing in that initial post says he's a liar. As for the Intel Paying more, right after I read the article I went to log in and saw the Intel Resource Center on the front page of this site. FTW, AMD, Intel and others do buy ads on all these boards. My implication was more about who provides the most funding, could this persuade him to interject Intel articles inside and AMD review, at 430am it seemed that way to me. If I were to scratch any part of the original post, it might be that line but that was part of "my" rant. It's amazing how many words have been turned around as this thread has progressed. It doesn't call him a liar and I still don't imply he is a liar but some of the attacks Ive received in this thread do. I've been back insulted by a handful of people here. People are trying to sell me Intel hardware, I won't take it personal and attack anyone over it.

I think if I had said the same about an Intel review someone did that included an AMD blurb right in the middle, I wonder is I would still be getting beat up over it?

BTW, what happened to the foul language filters here, we have minor reading these forums that Anand has put his name on.

Once again, the defacto statement that I'm just mad about AMD loosing is false. I was simply annoyed about the position of page 2 and so little coverage of Spider. But I'm human and did not get my intentions across.

 

[myocardia]

Originally posted by: AlabamaCajun
Nothing in that initial post says he's a liar.

You're right, you never used the word liar, but it was very obvious what you were implying. Yet when someone implies something about you, you say it's a personal attack, even though when you do it, it isn't? That isn't how it works in the real world.

As for the Intel Paying more, right after I read the article I went to log in and saw the Intel Resource Center on the front page of this site.

I realize that you're pretty new here, but back in the "old days", Intel was the only CPU manufacturer who bought ads on anandtech, just like today. Guess which company's CPU's were recommended by the anandtech staff? I'll give you a hint, it wasn't Intel. Way back then, they recommended the top performing processors, just like they do today. If you can't handle that, maybe you should only post here: http://forums.amd.com/

I think I had said the same about an Intel review someone did that included an AMD blurb right in the middle, I wonder is I would still be getting beat up over it?

If you said the same things you've said in this thread, yes. BTW, since you still haven't read the article, not one person from anandtech attended the AMD Lake Tahoe event. That's revealed in the article, which you've continually claimed to have read.

Needless to say, I wasn't happy. I refused to go to Tahoe.

Don't get me wrong, a free trip to Tahoe is a wonderful thing, but Phenom deserved better. It deserved dedicated testing, it deserved a thorough review, not a quick glance over a couple of days.

http://www.anandtech.com/cpuch...howdoc.aspx?i=3153&p=3

edit: Oh yeah, and if you had been around, back when the first Athlon 64's came out, you would have seen 10-20 times as many people saying exactly the same things about the Pentium 4, as you've said about the Phenom. So don't feel bad, it's really just human nature, obviously.

 

[JumpingJack]

Originally posted by: BitByBit
Phenom's performance is certainly something of a mixed bag. Intel's own benchmarks put Barcelona well ahead of Xeon in HPC and Floating point performance, per clock, which is in direct contrast to performance on the desktop.

I'm inclined to agree with other posters here that Anand's Phenom review was severely lacking in depth. I for one would love to know why there is such a huge disparity between server and desktop performance.

Alas, the Barcelona core does indeed have a market in applications where there is concern for BW limitations, these 'IPC' beating benchs you site are more of Intel's core being hindered by lower BW comparitive to Barcelona, which has done a good job on the BW side of the equation.

The reason you see the DT benches lopsided the other way is because DT applications have working sets so much smaller that only a fraction of the BW is consumed, these type of applications are called CPU bound or core-limited. The overall performance comparing a high octane HPC application to Doom 3, Lame MP3, or WinRAR is not appropriate, they are completely different classes.

If your interest is calculating the fuild field and Bernoulli coefficients for a turbin pumped water based tooth pick, the Barcelona will suit just fine....

So while clock for clock, instruction for instruction without any limitations -- the Xeon will out perform the Barcelona Opty 2350 or such, there are situations when the amount of data needed by the core exceeds the amount that can be supplied by the pipeline (i.e. the data/instruction bus), in which case the Barcey will appear faster because it does indeed deliver more effective BW to the cores than the Xeon.

This is no different when Conroe/Woodcrest launched. While woodcrest overtook K8/Opty's in several workloads, the K8/Opty was still much more competive in the BW applications comparitively where as Conroe just dominated on the desktop.

Jack

 

[JumpingJack]

Originally posted by: gOJDO
@Gary Key

What happened to the 2.4GHz Phenom which was supposed to have higher IPC than the lower clocked Phenoms?

gODJO -- you know, it was delayed.

http://www.digitimes.com/systems/a20071121PD215.html