Quad Core Round Up: AMD Vs Intel Clock to Clock(deneb vs agena + deneb vs intel)

[Zucker2k]

yes its a completely horrible comparison since how i have seen many more. it seems hes just picking the best yorkfield he can find and then find the worst phenom II benchmark out with the absolute worst settings. and the phenom has 800 mhz ram not 400.

You of all people are just uninformed, and I've chosen to ignore your posts, but let me say this: just scroll up and jou'll notice somebody used this same screenshot as proof that PII benches at 3.7Ghz should have been faster.

Why don't you guys do this: get your PIIs and beat that!!

[roofsniper]

You of all people are just uninformed, and I've chosen to ignore your posts, but let me say this: just scroll up and jou'll notice somebody used this same screenshot as proof that PII benches at 3.7Ghz should have been faster.

Why don't you guys do this: get your PIIs and beat that!!

i really don't understand how you can call any of us uninformed. i know more than you do about the topic and personally you don't even seem to care as long as you can find certain benchmarks and try to prove that phenom sucks.

[sundancerx]

As i had promised here about the clock to clock comparison....

finally here it is...

http://www.hwbox.gr/showthread.php?t=3253

Its in greek(again) but full of charts so its easy to understand what we are saying and whats going on.
Here is a link withgoogle translator

Deneb vs Bloomfield vs Yorkfield vs Kentsfield
225*16.5 __169*22____411*9______411*9

and

Deneb vs agena @ 3.0Ghz deneb with 1800nb clock as it is by default and agena(9950BE) 200*15.

Best wishes for the new year from HwBox team

hi man, great follow up. sorry if im wrong, but i dont see power consumption between agena and ph2, whats the excuse now?
btw, do you have review when nehalem launched?

[Zucker2k]

i really don't understand how you can call any of us uninformed. i know more than you do about the topic and personally you don't even seem to care as long as you can find certain benchmarks and try to prove that phenom sucks.

Maybe you're dense too; so let me repeat, I said YOU are uninformed! It's glaring from your posts. The best way to make your point is to post a screenshot, preferably, one that you yourself created; and one that is duplicable. With all your babbling and seemingly AMD affiliations its quite surprising we haven't seen a single screenshot from you.

Edit: I don't have anything to prove; the results are on overclockergr's website. You have a tall order trying to defend the indefensible.

[roofsniper]

hi man, great follow up. sorry if im wrong, but i dont see power consumption between agena and ph2, whats the excuse now?
btw, do you have review when nehalem launched?

he doesn't have the agena anymore to do the power consumption with.

Maybe you're dense too; so let me repeat, I said YOU are uninformed! It's glaring from your posts. The best way to make your point is to post a screenshot, preferably, one that you yourself created; and one that is duplicable. With all your babbling and seemingly AMD affiliations its quite surprising we haven't seen a single screenshot from you.

hmmm i thought providing links and saying obvious things were fine. what do you want screens of? if you have no clue what you are talking about maybe you need to find out before you start posting.

[OverClocker_gr]

hi man, great follow up. sorry if im wrong, but i dont see power consumption between agena and ph2, whats the excuse now?
btw, do you have review when nehalem launched?

the excuse that agena is not in the power consumption test,is that i gave him back the day before i got my watt meter.

Yes i did a Core i7 few days after their launch but i cannot understand what you want to say.
What is your point?

[HuffPCair]

Reconfirms my purchase of my q6600 being epic.

Its just mind blowing that even now, AMD has not caught up to Kentsfield. Good step in the right direction though.

Was thinking the samething. Intel still killin AMD. Makes me a happy man.

[AbelJemka]

Clock for clock is good to compare architecture but just for that!
At stock clock PII 920 will beat Q6600, Q9300 and probably Q9400 and that will be sufficient to make him competitive.
That's the point.

[Pillo-kun]

Am I doing something wrong, I can not see the update of the article at all.



Edit: Wow, now i see tehe... I thought that phenom2 would fall back even more at same clock but i think it does rather well, but still to damn expensive. q6600 is a much better buy right now but if we will see phenom2 over 4ghz i guess even they can get a bit ahead, no? seems like phenom2 is what pentium 4 was, a cpu in need of speed(Mhz).

[roofsniper]

Am I doing something wrong, I can not see the update of the article at all.

its a whole new thread. everything is the clock to clock unlike the other one which is stock.

[YukonTrooper]

Am I doing something wrong, I can not see the update of the article at all.

Like Roof said, every graph with Kentsfield, Yorkfield, Nehalem, Deneb reflects all chips at 3.7GHz. Then, at the bottom of every page, there is a 3.0GHz Deneb/Agena comparison.

[informal]

Clock for clock is good to compare architecture but just for that!
At stock clock PII 920 will beat Q6600, Q9300 and probably Q9400 and that will be sufficient to make him competitive.
That's the point.

Some intel users forget that even 9950 beats or equals Q6600 when both are at default clocks .Phenom II 920 will surely beat both Q6600 and Q9300.As for Q9400 I'm not sure,probably equal or win some/lose some. If superpi or similar "apps" are in question,intel quads will be faster .On the other hand if one uses mainconcept or does divx compression or games at high settings(the settings actually used in games by enthusiasts), I'd expect Deneb to be faster.

[Hornet331]

Some intel users forget that even 9950 beats or equals Q6600 when both are at default clocks .Phenom II 920 will surely beat both Q6600 and Q9300.As for Q9400 I'm not sure,probably equal or win some/lose some. If superpi or similar "apps" are in question,intel quads will be faster .On the other hand if one uses mainconcept or does divx compression or games at high settings(the settings actually used in games by enthusiasts), I'd expect Deneb to be faster.

In game settings at high i dont expect any processor to be faster especial on 1920x1200 + AA and higher, cause they run into gpu limitiation.

[informal]

In game settings at high i dont expect any processor to be faster especial on 1920x1200 + AA and higher, cause they run into gpu limitiation.

Being faster could mean having higher min fps .Even if it means running at 27fps in that resolutions versus a 23-24fps,the difference could be spotted.
Also notice that I mentioned the gaming scenario last(as it is the least important one) .

[Zucker2k]

Some intel users forget that even 9950 beats or equals Q6600 when both are at default clocks .Phenom II 920 will surely beat both Q6600 and Q9300.As for Q9400 I'm not sure,probably equal or win some/lose some. If superpi or similar "apps" are in question,intel quads will be faster .On the other hand if one uses mainconcept or does divx compression or games at high settings(the settings actually used in games by enthusiasts), I'd expect Deneb to be faster.

More bull. We've seen the game benchmarks, we've seen x264 benchmarks, we haven't seen much SSE benchmarks, but nobody is complaining. Only AMD fanbwois like you are. Imho a true clock for clock is an app that loads the processor 100%. It's funny how AMD fanbwois are calling for gpu-bound benchmarks as if that is what is being reviewed. Come on, you've got to be smarter than that.

[demonkevy666]

Sigh! Please point us to the evidence that more ram helps cinebench. If anything, my NB is being tasked with such high fsb and 8Gbs. This is my everyday system and this is currently the hardware config. that's all. Of course, you guys can consider this a target when you get your PIIs. All you hear is excuses, DDR3, Expensive, Hot, etc.

This is a Q9550 with 8GB of cheap DDR2 ram at DDR2 990Mhz with loose timings.

phenom is weird with memory and Nb settings some chips will run high on both other run high one one and low on the other.

as for the memory being loose it's really that lose it's like set for a rate 1000mhz-1100mhz speed (with various volts). tight timings doesn't benefit phenom at all from what I've seen). The IMC on the chip can be picky about how much volts you use too.

[informal]

Zucker,I can't see your (probably useless as always) post due to magical button . Sorry!

But to repeat what i have already said:

I do appreciate the tests they(hwbox) done,but I'd rather wait for more reputable sites to post their scores .It may support the numbers from Greek site,but it may as well show some other aspects. Deneb needs to be just 6% faster than Agena,per clock on average in order to beat Kentsfield. From the initial tests we have seen the delta between the Deneb and Agena is north of 6%.

We have been discussing the hwbox.gr review a lot.It has a thread of its own... And yes,some of their numbers are a bit low(er) than expected or previously reported for Deneb. And yes,Kenstfield is beating Yorkfield and Nehalem in number of those tests from hwbox.gr so you can draw you own conclusions from that.

For Q6600 being able to outperform it,the same must work at 3.7+Ghz on air cooling,and this is assuming parity in IPC between the two(which in real apps is not the case-Deneb will be somewhat faster,be it a few percent so Q6600 would need to run even above 3.7Ghz on air for 24/7).

[sundancerx]

the excuse that agena is not in the power consumption test,is that i gave him back the day before i got my watt meter.

Yes i did a Core i7 few days after their launch but i cannot understand what you want to say.
What is your point?

ok. nice excuse.
well i want to see if core i7 results are consistent with what you got before, just to be sure. anyway, the site giving me headache, so slow to load on my crap net. maybe, ill just take the data as they are, and wait for other reviews to compare.
i would say you did the best you can in the review, props to you for that. really good effort you put up there.
its the first deneb review, so i hope you understand people will be asking for most minute details they want to see.

[TL1000S]

Relax guys.

There will soon be many reviews from all the familiar sites.

Regarding power consumption, most people wants to see this at stock.
Here I think PHII will do well and much better than PHI.
We'll see.
BTW:
Even for a given CPU model the power consumption can vary.
This was the case for my two Q6600 G0, which puzzled me.
The first (early G0) OC'ed "like hell" but used just as much power as my older B3. The later G0 OC'ed "like crap" but used significantly less power at same clock/volts.
Strange..
This may also be the case with Deneb, and later revision of Agena (9950) surely improved in this area.

[leoy]

@ zucker2k
All i see is that i am right again.

Score of yorkfield @3.7 @ hwbox : 14602 (with 411 fsb )
http://www.hwbox.gr/showthread.php?t=3253&garpg=12

Wanna put the nb little higher too???
Unless yorkfield also wins 2000 points when going 64bit

[accord99]

@ zucker2k
Wanna put the nb little higher too???
Unless yorkfield also wins 2000 points when going 64bit

Yorkfield gains about 1600 points:
http://www.anandtech.com/cpuchipsets...px?i=3184&p=12

[Daveburt714]

@ zucker2k
All i see is that i am right again.

Score of yorkfield @3.7 @ hwbox : 14602 (with 411 fsb )
http://www.hwbox.gr/showthread.php?t=3253&garpg=12

Wanna put the nb little higher too???
Unless yorkfield also wins 2000 points when going 64bit

leoy: I couldn't agree with you more...
Ever since the 1st word of PhII's potential has come out Zucker2k has done everything in his power to discredit the improvements make in K10.5.... Call anyone who speaks well of the improvements blind fanboi's etc...

When he posted his CB10 results from his Yorkie, I was not that impressed.

Thanks for the PhII results... Here is a ss of my tweaked 9950 and even it gives his beloved Intel a run for it's money.



BTW: His Q9950 run was on a 64bit OS.....

[roofsniper]

In game settings at high i dont expect any processor to be faster especial on 1920x1200 + AA and higher, cause they run into gpu limitiation.

exactly thats why where im at 1680x1050 with at least 4xAA i don't think any cpu would make any difference. with the way phenom II caught up on the 1680x1050 benchmarks im curious if at 1920x1200 if it would pass up i7 or just be equal.

and guys theres no need to even argue with zucker anymore. every single thing he posts that is "proof" why phenom II sucks has nothing to do with the topic or it is the worst possible setup for it. even tho we all have seen better scores he will use the lowest one to show the performance. he is obviously a troll and thats why i pressed the magic button since he doesn't contribute anything to any of the threads he posts in. you might as well ignore him too so we can actually find out information about deneb instead of arguing over senseless stuff that we all already have the answers to.

[accord99]

exactly thats why where im at 1680x1050 with at least 4xAA i don't think any cpu would make any difference. with the way phenom II caught up on the 1680x1050 benchmarks im curious if at 1920x1200 if it would pass up i7 or just be equal.

It'll be equal, plus or minus a noise value. Just like the i7 will regain the lead it had at lower resolutions when paired with more powerful GPUs.

[viccan]

i call some bull because there should have been ddr2 on yorkfield or kentsfeild

Benches have shown no real improvement of ddr3 on kentsfield and yorkies... I was going to go with the DDR3 version of my MB and opted out after careful examination of multiple benches, so #'s wise there is no major difference.

Great work Awesome charts... I've been toying with the idea of moving to the i7 this deffinatly has made me pause and say wait... Truley apprecaite the hard work keep it up !!!