I can't beat the 4ghz on phenom, but I've used 1.45 volts on this chip can gotten 44C. at idle :D 3.2ghz. also note I'm in 64 bit so I'll still have higher temps then most user in XP.
also id still depends on what your using to cool it.
Printable View
The operating voltage of Phenom II is higher than the original Phenoms, I see that as a bad thing. In terms of power consumption it is not going to be pretty, there is a reason both the 920 & 940 have 125W TDP. Now it is obvious that AMD's 45nm process is a big improvement over 65nm but the power results we have seen have been with chips clocked lower and with much lower voltages. For example the old reviews of RB-C0 Phenoms clocked at 2.2GHz and with 1.224 volts, the power consumption was impressive. But crack it up to 3.0GHz with 1.35V..... power consumption will soar.
I'll reserve final judgment until more thorough reviews appear with ocing/power consumption and more tests, but so far Phenom II does not live up to the hype we have seen recently. Once again AMD built up insane hype for a chip and once again that hype is turning against them. Deneb is what we expected it to be, it is Agena with additional cache and on 45nm, allowing for higher clocks, ~5-10% better performance/clk, and lower power consumption..... it is nothing more and nothing less. It is good competition for the Q6600 / Q9400 and will perhaps shake up some of Intel's lower end lineup if pricing is good. But it is still not competition for Q9550 and higher, pricing on those chips remains at Intel's discretion (which is horrible for consumers, if you haven't noticed we have seen very few price drops this past year or so as AMD has become more and more uncompetitive). Certainly there is no pressure to lower prices on Core i7 CPUs or motherboards, as they remain in a league of their own right now in most applications.
Yes I have seen that, but I don't get your point.
Intel's 45nm process used on Core2 starts using A LOT of power when you reach over 3.8+Ghz
http://images.anandtech.com/graphs/p...1228/16135.png
but you cannot say this will happen to phenom2 too, because of the differences in manufacturing technology
And Phenom II isn't going to be consuming a lot of power when it is overclocked to 4.0GHz with 1.6V, which seems to be what is required? When 3.0GHz/1.35V is already ~125W TDP.......... any CPU is going to consume a lot of power when you are pushing it to the limit with insane voltages/clocks.
amp x volts = watts. just because the volts are higher doesn't mean the power consumption will be higher. my fan uses 12V but does that mean that it consumes more power than my cpu? no so stop using volts as a comparison. the volts have nothing to do with anything. phenom II uses more volts than i7 yet somehow from what we have seen so far it overclocks better. phenom II uses more volts than i7. yet somehow from what we have seen it has a lower power consumption. i don't believe you can use volts to compare different cpus. not even cpus of the same brand.
If it clocks like 3.6+ ghz on average and its priced fairly than it will be an nice alternative up to mid/high end :)
i am sorry for ati seems they are gonna die with amd :( c'mmon amd you can do better than that at least game benchs at high res aren't bad and thats only positive thing about this new cpu
Intel's TDP and AMD's TDP are two different things in general. Both technically are the maximum power usage, but AMD's CPUs tend to actually consume close to TDP under maximum load. Intel, on the other hand, bundles up together CPUs into "classes" and gives them all the same TDP, regardless of actual power consumption. For example, Intel rates all 45nm duals at 65W despite them consuming 30-40W at most, meanwhile AMD rates certain duals at 45W that consume more than chips that Intel rates at 65W. Same thing with quads....... Intel's 130W QX9650 consumes less than AMD's 95W launch Phenoms. In general, AMD's TDP is much more aggressive than Intel's TDP.
See above, and performance/W comes into factor there..... only time i7 gets close to consuming that much power is when all 8 threads are under use, and in an app that takes advantage of HT performance is going to be significantly higher for the Core i7 CPU.Quote:
i7 TDP is 130W that's my answer
Not sure what everyone is so shocked about. Phenom + more L3 + higher frequencies = faster than Q6600, mostly keeping up with Yorkfield but a little behind still. Anyone who claimed it'd be faster than Yorkfields and 'almost as good as Nehalem' clearly decided to not let facts get in the way of their search for fantasy and drivel.
It's not like it's a new architecture. Barcelona was the CPU division's R600, this is their RV670.. Now let's just hope they have an RV770 to show as well. Given their current roadmap where Bulldozer is mothballed and this arch will be milked until 2010 and beyond I am pretty damn pessimistic. :(
well i tried working out the math but then i found out it makes no sense at all. how many watts can i draw at that wattage? do you mean how many amps it will have to draw to reach that tdp? i did the math for that and it is no more than 92.6 amps no more than hmmmmm. so are we looking at how much power the cpu will consume under the absolute worst conditions or are we looking at what its average power consumption is? the average is all that matters. as we don't look for maxiumum fps on a video card review we don't look for maximum power consumption either.
I should have mentioned that I was comparing it on a clock for clock bases. Even with a 600MHz advantage, it still can't completely dominate 2 year old quad core. Looks like I know what my next upgrade will be.
And Abe, it has been shown multiple times in various reviews Intel's TDP is very relaxed. The 45nm Penryns does not teach their max TDP @ stock clocks, unlike their AMD counterparts but I guess you will have a hard time believing that. :rolleyes:
to all the ungrateful idiots saying: oh i'm never going to look at a review from this site again because they used superpi :zombie: etc etc... give them a break. they spent however many countless days i'm sure coming up with the first real results to compare the i7, q9450, q9550, q6600 and the phenom II. not to mention breaking nda just to give the community what they want.. they didn't have to publish these. they're the first to give us something real to talk about, other than bull :banana2: :bananal: :rehab: :banana: speculations and rumours.
anyways, on to the results: from the limited gaming perspective, it seems like this chip is a winner. I'm definitely interested. Results are on par with my expectations for everything else as well. I'm getting one of these chips.
it can't? :| :confused:
you obviously didn't read the whole review.. even with a 400Mhz clock advantage (920) it completely dominates Q6600, whilst almost certainly using less power.
A 9950 for reference would have really helped here, but for a die shrink, like barc-shanghai, looks like they've done a great job.
This isnt a FULL review...Where are the power consumption and power dissipation tests?
I have nothing angst the board I actually plan on buying a destroyer. The Asus seem to be a favored board for overclocking, at least in the amd section. I don't see much about the destroyer here though, not many people will to put phenom in SLI/TriSLI. I would like to see some mutli-card difference with phenom 9950/9850 vs phenom II
you are right.
P2 @ default
http://img224.imageshack.us/img224/6...ebench1si1.jpg
qx9650 @ 9550 clocks
http://img68.imageshack.us/img68/2566/cinebench2fo4.jpg
Lets say that its not AMD optimized(which i dont know), how many "real" x64 progs do you use on your every day use?
I use Cinema 4D, Realflow, Maxwell Render, Maya and Houdini.
All are x64. :) I'm not your average guy, but there you go.
hey, thanks for the review, i found a lot of good information in there.
do you think you could run the games tests with HT off?... just for comparison sacks.
@Tiridum
sure AMD is much much better.
You can "feel" instantly the difference, browsing over windows. :)
Well in most tests it was that much better then a Q6600 @ 2.4GHz vs a 3.0GHz PII. I'd like 2 see how they all did at the same GHz.
Anyway very good review.
in our next review(which is on the way) for clock to clock comparison there will be the option for HT on and HT off.
Otherwise you can see here @ our older review for Core i7( http://www.hwbox.gr/showthread.php?t=2915 ) what you should expect
I'd like to see max and reasonable air overclock comparison's between them all.
Considering for core i7 3.6 is the sweet spot before heat gets to bad, phenom II seems like the sweet spot could be from 3.8 to 4.2, q6600 has 3.4 to 3.6 and q9450/q9550 has 3.8 to 4.0.
the phenom that i got in my hands, tops out @ 3.8 with ~1.5v.(stable)
@4g im not able even to boot windows with any voltage.
Huh, guess you got a bad one because most everyone that I've heard testing them has gotten to 4.0 pretty easily.
By that 4.0 I meant 940. I've seen 3.8 and 4.0 for 920 as well from high htt overclock, so I'm assuming they have pretty close air overclockability.
How high are you getting the htt/nb with your overclock?
Or are you just multiplier overclocking?
I just hope the PhII's get released pretty soon, so some of the XS guys with experience tweaking Phenoms can post some results...
Thanks for the review OverClocker_gr, it's nice to see some solid stock numbers but I can beat alot of the PhII scores with my OC'd 9950BE (@3.2Ghz)...
Phenoms have always been temperamental beasts. ;)
Before anyone makes snap decisions about their potential performance, why not wait until they are in the hands of people who have been taming it for the last year...
I'm not sure about ACC or AOD or w/e for higher then 3.6 I think you must have read 3.8, since from what we've been shown from a lot of people playing with them 3.6 can be achieved at stock voltage.
Yes I'm an addict keeping up with any info about this I can :rofl:
Great review, guys. Answered pretty much all the questions except for how retail samples will clock. That could make all the difference.
Actually retail samples may have slightly difference performance overall as well...such as fixing areas that they didn't do so well in with the ES's.
Perhaps not but as a price/performance based consumer I have to hope so.
Destroyer,M3A79-T,DFI 790FX
@3.6 i am with default voltage too.
The cpu that i have its not an ES.
http://www.hwbox.gr/images/imagehost...3f0a87b72c.jpg
In addition to the Intel favored software, as suggested above, maybe you could add some encoding and 7zip. But thanks for the numbers, as they are perfect for what matters to me! :) Definitely looking forward to this purchase. :)
I've only tested with laptops; a 1.73GHz Merom 2M Dell 640m with WI-FI and display set to minimum uses a total of 29W from the battery while running dual P95 at stock voltages, and 19W when undervolted to 0.95v. A 2.53GHz T9400 Penryn Thinkpad W500 uses 41W running P95 at stock and 33W running at 2.4GHz/1.15v. I haven't bothered with desktops but I've read enough of Silentpcreview forums to see that the results are accurate.
Knowing my luck, my 1st "Deuce" will be an 0843BPMW... :yepp:
Well, let's hope at this point that PII does overclock as well as initially rumored. And not just 1 in 10,000 chips.
It'll be a pretty big disappointment if 3.6 is around all that people can get, but we'll see. I'm not going to be buying
a PII in any case as I already have a board I love and a Q9650 to upgrade to, but I've always liked AMD and would
love to see them return to XS prime time.
*sigh* With all these performance results starting to surface it's painting a rather grim picture of the Phenom II. Rather than even hoping for clock-for-clock parity with Yorkfield, we're just wanting it to be a close match with its price-equivalent competitor at this point. It looks as though for the most part, not even that is going to happen.
As someone who wanted to keep the AMD family "tradition" alive (never owned an Intel system), I was really hoping for a return to form with this series. Sadly what does even comparable overclocking matter when 3.6GHz on Yorkfield gets you so much more than 3.6GHz on a Deneb? Forgive me if this seems "ranty", but I'm just really disappointed right now. Obviously I'll refrain from final judgement until the big-name reviews appear, but I'm not holding onto false hope.
people seem to forget this a great drop in upgrade for anyone with a AM2 mobo. As long as the price is right can't really say AMD failed. Now if this was a totally new chip for AM3 it be another story.
P2 is about what I was expecting.
Totally buying it. I better start shining my boot to kick Q6600's @ss good bye.
too expensive, but i hope the price will go down.
OverClocker_gr,
First, Thanks for your great review.
Well, Would it be possible you to test with SLI of GTX 260-216 and test in 1920x1080 ?
Thank You. :up:
I sure have heard of overclocking, Thats why my q6600 is at 3.6ghz, What I am asking you is that since you find i-7 a complete waste of money why are you so excited to change a whole platform for = performance?
And how is AMDS platform better? Any links or is this just your opinion?
That sums up what we have been hearing for weeks , PII still slower clock per clock even when compared to two years old kentsfield and that shows that Core 2 wasn't just small jump in performance but still waiting for that OC review thought .
Nehalem?
Nehalem would be a complete platform overhaul (CPU, board, memory [of which 2 latter are way too expensive]). That means i7 is an option I don't even consider. Besides, what would be the gain? Nothing in performance in applications I care about.
Yorkfield?
LGA775 mobos and bios' suck (and are boring), so Yorkfield is not an option. All i965Ps and P35s I've used since the launch of Core 2 have let me down.
What else?
That leaves Deneb - which uses a platform I've used before (pre-K10) and was pleased with. I loved tweaking (RAM especially) back in the days of S939s but when I moved to Core 2 tweaking became tedious due to quirky LGA775 mobos and bios'. Sure, the CPUs are quick, but in the end, I have no use whatsoever for CPU power. I don't have the need to experience the illusion: "Wow. Is this CPU is blazing fast..." That is, while I browse the internet, listen to music and the system uses maybe 1-2% of CPU resources. Even in gaming (casual) - there is no difference.
But before anything; I want the fun back in OC'ing. And a better conscience.
Then again, the most effective way to gain overall system performance would be to get an SSD. No CPU is fast enough to match that kind of gain.
Why are so many of you saying this? Am I looking at a different set of graphs or something?
From what I saw it was quite comfortably duking it out with the Q9550 90% of the time, it was only the tests that have always favored Intel that it was at the same level as the Q6600....
Jesus christ even now some of you seem to be trying to highlight the few bad points over the masses of good ones.:rolleyes:
I think people mean clock for clock, Dont forget that the q6600 is 2.4ghz, If they clocked it up to 2.8-3.0ghz ph2 would be having a hard time with it, Even if you clock them both to the max I doubt that ph2 would be much better. Thats why I am disappointed, Ph2 looks to be 2 years late.
Pretty much what I expected for stock clocks, thanks for the preview. Will be looking out for reviews showing real world results that simulate how most users will be using their quad cores.
Something like, running music, watching video through web browser, multitabs open, virus scanning in background and maybe some file browsing, win raring in the mix.
Also am very interested to see reviews showing performance scaling with increase of nb speeds. If AMD had been able to release the Phenom II at default HT/NB speeds of 2600mhz, me thinks there would be a lot less negative posting here.
So at stock it is a little disappointing, however I expected that, now its just upto some peeps who have the chips to start increasing HT/NB frequencies and showing us comparitive performace gains........
Oh, yeah, thanks for the numbers.
But what's the idea behind the flash charts? To simulate slow internet?
¦)
I dont understand how people can say that amd has a better platform? bought a am2+ for my brother for a phenom 2 sys and although the cpu right now is a 5200+x2 i feel that the am2+ sys feels cheap and not as well thought out as the 775 sys. lets hope that the phenom 2 falls in price cus if not then i have made a bad platfor choice for my brother.
arent anybody that wondering why tahts so different core corei7 in 1280*1025 to 1680*1050. yes in that res that is more gpu limited but this doesnt explain for example in far cry 2 corei7 give %50 more then phenom2 @ 1280*1024 but in 1680*1050 it gives %10 less. you cant explain this by gpu limitation because they all use same gpu.
So what i am thinking is the problme is all about x58. It seems like it has no problem with cpu but may x58 is limiting gpu more the previous chipsets?
Thanks for the review, I think the community needed it.
PhII looks quite decent, coming in somewhere between the Kentsfields and Penryns. Not bad, that is quite a lot of performance for any use.
The prices are a bit high though. I hope availability brings some price drops for amd and price cuts from intel... :yepp:
Sure if you have a thousand Plus Dollars to upgrade to Nehalem .....GO FOR IT!
I will spend a couple of Hundred and and be thrilled compared to what we have witnessed over the Past years disappointment. Taking what I just said into consideration it is indeed a huge leap for AMD Compared to initial release of Phenom.
But the Fact is that Nelhalem is still a RIP from AMD whatever which way you slice it. I guess with unlimited capital of Intel & RD Team compared to AMD they just better make it sing.
Perhaps he had a bad experience with Intel platforms, but most people would agree that Intel does have the better platform at the moment. Platform quality goes hand-in-hand with what socket chips are doing well in the market. S939 was prime back in the day, as AMD was selling chips like hot cakes. There were tons of boards to choose from; therefore, the competition was good and you could pick up a solid mobo for dirt cheap. Now, we have a large Intel influence in the market and the same thing is happening with LGA775. Companies are simply putting more resources and effort into the LGA775 based mobos, as the chips that correspond to those platforms are selling better. Seems pretty straightforward to me.
I'm probably gonna get flamed for this, but personally I'd go PII just to be rid of the :banana::banana::banana::banana:ty ICH10 controller. If PII can offer similar performance to yorkfield at a lower power consumption, that coupled with a better mobo chipset imo then it would be a viable alternative. Especially seeing that i wont have to swap out my ram. I primarily use my machine as a development rig and right now i barely strain my wolfdale.
This is all dependent on pricing tho, a lot of people might find it cheaper to move to yorkfield if they already have intel systems. CPU wise I'm running a wolfdale because yorkfield was prohibitively expensive when i bought my CPU. then again by feb the prices of the yorkfield chips may have dropped.
well i did hoped for more. Also for AMDs' sake.
But i am still considering buying a PII. Curious about spi and pif times. also those wprime charts didn't looked so bad! if the PII rly clocks that high it may do some nice scores!
And for rl it is totaly uninteresting what those benchmark apps tell.. they are so much plattform depending. Mostly it won't show the real thing.
btw i am only gaming with 1680x1050 right now and even thinking about going higher.
and if u play with lower res. does it rly matter if u have 170Fps or 270Fps?? but than it may matter if u have 40 or 50Fps in higher res.
But the Ppc is amazing for i7, intel did a good job while integrating the mc.
I think most important now is how prices are developing.. and both platforms have their pros.
You should really try GTA4 on the AMD platform...
I think that power consumption is important for OC. New c2q E0 do not consum to much power. This is my latest Q9550 11h prime – 4GHz@1.216@air and you do rest of mat.
I think that AMD must release new CPU stepping with lower consum, in future.
Not much go in Phenom 2. And the IPC like 65nm Core 2 was true.
I don't think we have enough info to be claiming performance-per-watt as a strength. I'll believe it when I see it. Performance-per-watt over Yorkfield isn't looking too promising...
i never stated that.... *sigh*
both Intel's and AMD's cpus have their own advantages/disadvantages
but lets consider the facts, AMD finally has produced something that is competitive somehow
I dont feel anything about being smoother with a K8 platform, my E8400 @ 3.8Ghz is doing a fine job at making things smooth...
@Zucker2K
quite funny how you misinterprete my post
let me clarify it for you
- phenom2 <<<<<< i7 (so it isn't a competitor)
- phenom2's prices is starting at around 200$, that is midrange
- people can finally start choosing (between yorkfield and deneb) both platforms have similar performance and similar price, so people have choice now, a few months back Intel dominated the whole market, it's always good to have competition for us, so that prices go down
I don't care what people are posting. I want official reviews from respectable sources :)
Either way I'm getting a 940 BE :D
yes,but nothing different about the performance of the cpu's
@OP: Thanks for the results, nothing unexpected here.
What did anandtech use to generate load? techreport used cb10. Use CoreDamage or Core2MaxPerf to get close to the worst case, should be at least 20W AC more.
Can you please run one of the above mentioned apps and post consumption in idel and load, so we can compare once PII is available for all of us?.
I'd vote for 3.0/3.6GHz and 3.8-4GHz. 3GHz at stock all others at lowest possible voltages. I assume no chip will have problems with 3.6GHz at reasonable voltages.
For power consumption it would be nice to have two alternatives generation load. One can be CB10 or povray to simulate an average load and the other should be an worst case tool like coredamage or core2maxperf.
I wonder if the scaling on PII's diminished because the single core runsnow makes better use of the bigger L3 cache but it's not enough to improve the multi core test as well?