AMD Phenom II Review Thread

[Miss Banana]

Did you even read my post pan? Or are you just going to ignore it? It seems very obvious that it is completely unclear what results in the review can be attributed to the chipset, the motherboard and the ram, and what results say something about Phenom.
Even kyle himself says something weird is going on with that motherboard.
Click my signature to read the post again.

[Zucker2k]

Could you explain why, after reading my post about the chipset he used, you are not questioning the numbers?
Just curious here.

Ok, this is ridiculous! You said the numbers don't matter! Now, you're saying they do?

[marten_larsson]

Could you explain why, after reading my post about the chipset he used, you are not questioning the numbers?
Just curious here.

I'm not questioning that he is posting the actual numbers he got. There's no point in doing that (if you want to do that, why not just make up all the results). But I'm not considering his results "true" or "valid" in any way. As I said, there are good reviews and then we have this not so good one...

panfist: If you look strictly on performance yes. But if you add price in the equation it's more like i7 ~ C2Q ~ PII. And it's not 10% faster per clock (maybe in the given benchmark but not overall. Look at table in previous page.

[panfist]

Did you even read my post pan? Or are you just going to ignore it? It seems very obvious that it is completely unclear what results in the review can be attributed to the chipset, the motherboard and the ram, and what results say something about Phenom.
Even kyle himself says something weird is going on with that motherboard.
Click my signature to read the post again.

You are attacking Kyle's methodology. I agree, it was very flawed.

My position is that regardless of Kyle's methodology, I agree with his conclusion. Why? Because you can cut and paste benchmark charts from ANY website, and the conclusion is still the same: i7 > Core 2 Quad > Phenom II.

[Miss Banana]

Ok, this is ridiculous! You said the numbers don't matter! Now, you're saying they do?

Yes Zucker, it IS getting rather complex isn't it? Don't worry I will explain it to you again.
After the mess kyle made of the hardware setup, the numbers become irrelevant as it's not clear whether a benchmark result is corrupted by motherboard issues, pcie bandwidth issues, memory size issues, or memory speed issues.

Now look at the word irrelevant and think about its meaning for a second.

Some results could be right, some results are likely to be wrong. Noone knows what is the case for what results.
Trusting these results(as in thinking they say something valid about the Phenom II) is rather naive, as kyle himself pointed out how strangely the motherboard for the AMD system performed in an earlier review.

Anything else you want cleared up?

[gosh]

...because this is a CPU test so they tried to take the GPU out of the equation, by using old games at low resolutions. The GPU can crank out as many frames as the CPU can feed it.

Ok, but that has little to do with how good the processor is for the game. It's like super pi then. Just used to see how good number you can get but doesn't have any real value to see how good the processor in real use.

Gosh, then you go on to attack the actual benchmarks saying, "well what if you're just standing around in game that's not a real situation." Look at all of HardOCP's other reviews: real world game situations with FPS over time graphs. What makes you think that they would do something stupid like load up level 1 of a game and just stand there?

Why don't you please prove the legitimacy of the other benchmarks that you say are so much better...oh wait...[B]it's irrelevant, because every single website has the same ranking: i7 > Core 2 Quad > Phenom II.

I don't know about HardOCP's other reviews. I was informing about this and if they was using a configuration to crank out as many FPS as possible and used that to conclude how good the processor if for the game then that is really stupid.
Reading in the Hardforum and if that represent how HardOCP does review hardware then it makes sense because there are some intel fantatics that seems to spend all time to write about how bad AMD is. That forum is just one big joke

[gosh]

Why? Because you can cut and paste benchmark charts from ANY website, and the conclusion is still the same: i7 > Core 2 Quad > Phenom II.

Well, if you like me do a lot multitasking and vmware then it is more like: Phenom II > i7 > Core 2

[ryboto]

You are attacking Kyle's methodology. I agree, it was very flawed.

My position is that regardless of Kyle's methodology, I agree with his conclusion. Why? Because you can cut and paste benchmark charts from ANY website, and the conclusion is still the same: i7 > Core 2 Quad > Phenom II.

The conclusions I read implied Phenom II was competitive with Core 2 Quad. Though not mhz for mhz competitive, it is competitive. If you're considering overall performance, then yes, there are faster more expensive Core 2's not competing directly with Phenom 2. The conclusion can be correct in a certain light, though it's still a vague one.

[panfist]

Well, if you like me do a lot multitasking and vmware then it is more like: Phenom II > i7 > Core 2

I would like to see some reviews that say this, because actually I am very interested in multitasking and vmware. I didn't find any reviews that discussed virtualization.

The conclusions I read implied Phenom II was competitive with Core 2 Quad. Though not mhz for mhz competitive, it is competitive. If you're considering overall performance, then yes, there are faster more expensive Core 2's not competing directly with Phenom 2. The conclusion is correct, though vague.

Clock for clock is the only comparison that is really relevant to me, because I'm not going to buy the top of the line, I'm going to buy a much cheaper model and overclock past the top of the line clock speed, anyway.

[Zucker2k]

It's really not about his numbers, it's about his attitude and subjective test setup.

Stop dilly-dallying and be serious. You're actually competing with Informal for contradictory stands in the same thread.

[Donnie27]

http://www.extremetech.com/article2/...2338339,00.asp

AMD will also be at a memory bandwidth disadvantage with its DDR2 controller. While some DDR2 modules support 1066MHz operation, we've had difficulty overclocking the CPU while maintaining 1066MHz memory speeds; 800MHz seems generally more stable. Contrast that with DDR3 speeds starting at 1066MHz and going up. AMD will have an update to Phenom II sometime in the first half of the year, which will support DDR3 and DDR2 memory.

Kyle wasn't the only one to have problems with the RAM, maybe just early shake down jitters.

[ryboto]

Stop dilly-dallying and be serious. You're actually competing with Informal for contradictory stands in the same thread.

That was Miss Banana's opinion, then, after some great research, found more information and the opinion was re-evaluated. So, you can either berate them for becoming MORE informed and ignore the information that's been dug up, or realize you've got nothing to add and not click the "reply" button.

[Chad Boga]

What Kyle did in his test was to focus on areas that isn't important and used that to conclude how good the processor was.

Why don't you post on his forums and explain to him the error of his ways?

[gosh]

I would like to see some reviews that say this, because actually I am very interested in multitasking and vmware. I didn't find any reviews that discussed virtualization.

You won't find any.
Have tried vmware on three different intel computes and there vmware workstation just sucks on intel. I can't use it because it is too slow.
Multitasking you need to read between the lines in tests and also read about how the processor works. As I understand it AMD is built to handle worst case scenarios good. It isn't built to be a dragster when all data is coming perfectly. Running a lot of applications then these worst case scenarios is showing up a lot more.
Intel has more focus on beeing fast when applications behave good. Worst case scenarios they will have bigger performance hit compared to Phenom.

I don't think a computer user could say which computer he if it was a i7 or a Phenom II though.

[Zucker2k]

That was Miss Banana's opinion, then, after some great research, found more information and the opinion was re-evaluated. So, you can either berate them for becoming MORE informed and ignore the information that's been dug up, or realize you've got nothing to add and not click the "reply" button.

Since when did you become Miss Banana's spokesperson? I'm calling her (assumption) out. I'm sure she can defend herself.

[Zucker2k]

You won't find any.
Have tried vmware on three different intel computes and there vmware workstation just sucks on intel. I can't use it because it is too slow.
Multitasking you need to read between the lines in tests and also read about how the processor works. As I understand it AMD is built to handle worst case scenarios good. It isn't built to be a dragster when all data is coming perfectly. Running a lot of applications then these worst case scenarios is showing up a lot more.
Intel has more focus on beeing fast when applications behave good. Worst case scenarios they will have bigger performance hit compared to Phenom.

I don't think a computer user could say which computer he if it was a i7 or a Phenom II though.

Gosh stop being silly, Nehalem has eight threads and has proven capable of using them... well. I think you're an intelligent person, but the continuous use of this example to prove your point about nothing other than multi-tasking undermines that assumption on my part. You can't compare a dualcore to a quadcore in a multitasking environment - that is the sole purpose and strength of quads. Please save that comparison, I'm not interested. Have you tested Nehalem in a vmware environment?

[ryboto]

Since when did you become Miss Banana's spokesperson? I'm calling her (assumption) out. I'm sure she can defend herself.

Because, it's obvious what happened. She(assumption continued) developed an opinion, then presents us with NEW(technically old) info, and the opinion changes. It doesn't need defending because it's blatantly obvious. Though it apparently needs to be explained for some people.

[panfist]

Gosh stop being silly, Nehalem has eight threads and has proven capable of using them... well. I think you're an intelligent person, but the continuous use of this example to prove your point about nothing other than multi-tasking undermines that assumption on my part. You can't compare a dualcore to a quadcore in a multitasking environment - that is the sole purpose and strength of quads. Please save that comparison, I'm not interested. Have you tested Nehalem in a vmware environment?

I wasn't going to respond to Gosh because there is no data either way, that I can find. My gut instinct was that i7 would be better due to hyperthreading. I have an E8400 machine running Gentoo that hosts three virtual machines with VirtualBox and "slow" is not a word that I would use to describe it. I access the VMs solely through a remote desktop connection and they feel just like remoting into bare metal.

I don't remember where I saw it, but I did see some previews benching Phenom II in a server situation. It was indeed much improved over Intel's Xeons, but then when they benched i7 it blew Phenom II out of the water in multi-threaded situations.

Oh, and the sole purpose and strength of quads is not multitasking. There is also parallelized computation, like video encoding, servers, etc.

[Tonucci]

You won't find any.

kk

Stock clocks, imagination, emotional attachment or anything ?

[marten_larsson]

I would like to see some reviews that say this, because actually I am very interested in multitasking and vmware. I didn't find any reviews that discussed virtualization.



Clock for clock is the only comparison that is really relevant to me, because I'm not going to buy the top of the line, I'm going to buy a much cheaper model and overclock past the top of the line clock speed, anyway.

Yeah, and if you look at it, if PII clocks 10% better than similarly priced Intel chip, the difference is gone.

[Zucker2k]

Because, it's obvious what happened. She(assumption continued) developed an opinion, then presents us with NEW(technically old) info, and the opinion changes. It doesn't need defending because it's blatantly obvious. Though it apparently needs to be explained for some people.

In case you forget, I was the one who raised the probability of a chipset or board limitation. I also, said maybe before folks rush into judgment, it might be wise to be sure the hardware wasn't the issue since Kyle had indicated he run into problems. That is showing sound judgment, don't you think? Miss Banana is the one with links in her sig and iswaging an all out foul cry, mixed with name callings and other outrages which can all be found in this thread and others in the AMD section. Where's the fire? None has done what Kyle did, and even though M B admits no one knows whether more ram would impact the reviews positively for PII, while at the same time admitting there was nothing wrong with the numbers, attacked Kyle's "attitude" and "subjective setup." The one constant for most in this thread is that regardless of ram timings or size the fact still remains that Ci7>C2Q>PII; that is the conclusion Kyle arrived at, even though his expression of that is very problematic and in bad taste; still takes nothing from the result, ultimately. I do reserve the right to give M B a hardd time for contradicting herself, it is a very bad thing indeed in a debate which I believe this thing is.

[Zucker2k]

I wasn't going to respond to Gosh because there is no data either way, that I can find. My gut instinct was that i7 would be better due to hyperthreading. I have an E8400 machine running Gentoo that hosts three virtual machines with VirtualBox and "slow" is not a word that I would use to describe it. I access the VMs solely through a remote desktop connection and they feel just like remoting into bare metal.

I don't remember where I saw it, but I did see some previews benching Phenom II in a server situation. It was indeed much improved over Intel's Xeons, but then when they benched i7 it blew Phenom II out of the water in multi-threaded situations.

Oh, and the sole purpose and strength of quads is not multitasking. There is also parallelized computation, like video encoding, servers, etc.

Yeah, I should have said multi-threaded, which is also a plus for multitasking.

[LiquidReactor]

Sorry Abel, but this is ridiculous! My God, Jesus, Joseph, and Mary!!! You made a mistake and you get straightened and then you start talking about some other nonsense!

Let's see the excuses here from the AMD crowd:

Board is defective/8x PCIE Lanes/Not enough PCIE Lanes Immature bios
Ram is timings are loose/only 2GB used
The Games tested are too old
i7 is too expensive so the comparison is not fair
QX 9770 is on DDR3 and costs an arm and a leg
Why test with SLI/Why not test on Ati card, AMD will do better
The reviewer is biased/paid/drunk/incompetent/amateur
Etc.

Pathetic!

Hey Genius, if you think your so $%@ing smart and know better then every high paid engineer at Amd why don't go over there and help them out with their next gen cpu design. Oh wait, you can't because your another idiotic fanboy troll. Keep trying to make yourself feel very important just because you have Intel in your system.

[panfist]

Yeah, and if you look at it, if PII clocks 10% better than similarly priced Intel chip, the difference is gone.

Where in the world did you get the idea that PII clocks higher than intel chips?

Most Intel 45nm Quads can get to 4GHz, or very close...how high does the Phenom II go?

[VoodooProphetII]

Yes Zucker, it IS getting rather complex isn't it? Don't worry I will explain it to you again.
After the mess kyle made of the hardware setup, the numbers become irrelevant as it's not clear whether a benchmark result is corrupted by motherboard issues, pcie bandwidth issues, memory size issues, or memory speed issues.

Now look at the word irrelevant and think about its meaning for a second.

Some results could be right, some results are likely to be wrong. Noone knows what is the case for what results.
Trusting these results(as in thinking they say something valid about the Phenom II) is rather naive, as kyle himself pointed out how strangely the motherboard for the AMD system performed in an earlier review.

Anything else you want cleared up?

Very well said, and pretty objective when looking at the facts.