Actual deneb review/comparison to Intel

Page 9 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.


Seems to match what I've seen so far - haven't been able to check AT yet because the server was down or swamped at lunchtime. I suspect these 4+ GHz overclocks on air may have been cherries that got picked by AMD :)...

That said, I think we all should acknowledge that AMD is heading in the right direction, and the P2 is what they should have done with P1. While I don't personally see any reason to go with a P2 on an AM2 for my next build in a couple months, I also would like to see what the P2 can do with an AM3 mobo and DDR3, esp. when the 40nm GPUs come out in 5 months or so. Maybe AMD will make me regret my i7 920 by then :).
 
Seems to match what I've seen so far - haven't been able to check AT yet because the server was down or swamped at lunchtime. I suspect these 4+ GHz overclocks on air may have been cherries that got picked by AMD :)...

That said, I think we all should acknowledge that AMD is heading in the right direction, and the P2 is what they should have done with P1. While I don't personally see any reason to go with a P2 on an AM2 for my next build in a couple months, I also would like to see what the P2 can do with an AM3 mobo and DDR3, esp. when the 40nm GPUs come out in 5 months or so. Maybe AMD will make me regret my i7 920 by then :).

I think we will be safe with the i7 920's, when AMD release the AM3
they should compete with mid to low range Q9000 thats about it.
 
^ OK I managed to get to AT's review of P2, and in the gaming benches which JDJ brought up frequently as being "disappointing" for i7, here is what they had to say:

The take away point is that compared to Penryn, Phenom II is slower clock-for-clock. The gap grows with Nehalem; Phenom II only gets close in older game engines, while the rest of the time Nehalem is 30-60% faster at the same clock speed.

 
My biggest issue is that AMD has nothing planned after P2 945 on AM3 till Bulldozer in 2011. Obviously there will be the odd stepping changes but no new arch or shrink is planned. So essentially when AM3 comes out we will find out what AMD has to compete will Intel for the next 2 years which currently barely competes with Intels 2 year old cpu's. So by the time Bulldozer comes out AMD will be 4 years behind Intel.

Its such a shame because this is a move the right direction for AMD but at a minimum at had compete with Core clock for clock which doesn't. Seeing as Core is now becoming obsolete with i7 and i5 to take over it doesn't look that good for AMD.
 
^ OK I managed to get to AT's review of P2, and in the gaming benches which JDJ brought up frequently as being "disappointing" for i7, here is what they had to say:



Quote :


The take away point is that compared to Penryn, Phenom II is slower clock-for-clock. The gap grows with Nehalem; Phenom II only gets close in older game engines, while the rest of the time Nehalem is 30-60% faster at the same clock speed.

This is why JDJ shouldn't become a cheerleader before the real result
are posted, again cherry picked cpu's just like first phenoms.
Now when AM3 comes out lets hope he learned his lesson wait
before being a AMD sock puppet.
 
ah jdj didn't make any personal guarantees. he was excited because of the preliminary info that was sprouting up. the price point for a 3 Ghz quad core and the fact the boards are very affordable is significant. it doesn't have to beat the i7. you can build a heck of a nice p2 system for not alot of money.
 


I think AMD plans a 6-core version of Deneb/Shanghai, codenamed Istanbul IIRC, to debut late this year. And they might just come out with HKMG, which JDJ poo-poohed as unnecessary but may actually do AMD a lot of good if added to their current 45nm process which is already a huge step up from their 65nm process. I wouldn't be surprised if HKMG got them up at least 2 speed bins and with better power consumption. So all is not lost with AMD, although admittedly they are pretty much at the mercy of the economy as well as Abu Dhabi, IBM, etc.

BTW, I saw a post on **** from our old friend, JumpingJack, where he estimates the hyperthreading power draw on i7 to be around 20-25 watts. Clearly it needs to be turned off for single or low-threaded apps or games. However some of the newer games will make use of a lot of threads, it seems.
 


Well some of us think he did a bit of spinning, and taking many preliminary positives for Deneb as being for granted. So yeah that seems to be cheerleading :). However I will say that he was correct in that 45nm is a really good improvement for AMD - not as great as he seemed to think it would be, as I haven't seen any reviews out today with 4GHz stable on air, but still when you compare it to the P1, it looks golden.
 
it is exactly what AMD needs. they needed something that could trade barbs with the yorkies. i am building one and if it doesn't meet my expectations, no big deal. my investment will be low enough to sell it or give it out as a Christmas gift lol
 
I forgot about the six core version. HKMG will help get power consumption down and increase the clock speeds abit. However I don't think AMD really wants to push the clockspeeds above about 3.2ghz. Becuase if they did Intel will probably just raise there clockspeeds to compensate so neither side gains anything.

The good thing about P2 is that there is clockspeed to spare so they can increase the clocks if they want. It also means that as the process matures if AMD keeps the stock clock speed to 3-3.2ghz room for overclocking will continue to grow.
 
Read Anands article, where it says the cpus given reviewers arent the newest rev, and higher ocs are expected. Unfortunately, I havnt seen any NB ocing, only straight multi, and thisll eventually become important in a couple of ways, as we will eventually see. Alot of these sites dont really know how to oc a AMD cpu as well, and they certainly wont take the time to learn, before deadlines.
Perf is where I expected, but Im thinking the ocing will rear its head, and will ultimately show to be greater than what weve seen
 
Heres a quote:
"AMD has proven in early demonstrations that the Phenom II x4 will offer overclocking headroom similar to the Penryn series. Early production sample processors have clocked anywhere from 3.9GHz on air to 4.4GHz on water and all the way up to 6.3GHz on LN2. We have matched their results on air-cooling and been impressed with the potential headroom offered by the new 45nm manufacturing process on extreme cooling setups. In fact, the latest retail steppings that AMD displayed this past week showed significant improvements in overclocking headroom compared to the press samples we utilized. Our 940 topped out at 3.9GHz, which is not bad, but after reviewing AMD’s results and seeing some early retail numbers on the forums, the expectation level for air-cooling is now set to the 4.1GHz range with the 920 hitting 3.8~3.9GHz on the right motherboard"
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3492&p=10
Just so no one thinks Im "spinning"
 

So, it comes down to me and Anand spinning eh? Now, whos doing the spinning? What has happened to the ocing, as the only claim in perf Ive made was 2-3% better than Kentsfield. So, explain your spin on the non ocing? Better yet, ask Anand himself, why he lied about this?
 

Ok ive been reading phenom 2 benchmarks seemingly the whole day, and what it has taught me is that AT's gaming benchmarks are flawed. Testing with low or medium details at a low res produces duff results. Turn details up to high/max and you stress the cpu in different ways, as those extra graphical features themselves use more GPUI and CPU power. Especially games that use streaming, where you cannot simply extrapolate/guesstimate what high streaming system load-performance by processor will be like by looking at the low load scenario results. Game engines and the inner workings of cpus are far too complex to have the sum of their interactions properly represented by analysing a lowest-load highest-raw-throughput set of numbers. Turning up the details at an appropriate resolution to keep the frame rate high/very high makes for different hierarchies on performance tables than if low settings were applied.
Cases in point, note the position of the dual core (e8600):

http://www.firingsquad.com/hardware/amd_phenom_2_940_performance/page7.asp

For good measure another example:

http://www.techspot.com/review/137-amd-phenom2-x4-940-920/page12.html

and again note the farcry 2 results, P2 cant push out lots of low quality frames but just going to medium detail evens things out while still maintaining a high average frame rate for all the cpus tested:

http://www.extremetech.com/article2/0,2845,2338347,00.asp

ah screw it heres some other good reviews testing games at something like realistic settings:
http://www.elitebastards.com/cms/index.php?option=com_content&task=view&id=667&Itemid=27
http://www.bit-tech.net/hardware/2009/01/08/amd-phenom-ii-x4-940-and-920-review/1
http://www.techreport.com/articles.x/16147
http://www.guru3d.com/article/amd-phenom-ii-x4-920-and-940-review-test/1
http://www.legitreviews.com/article/860/1/

look at more reviews than just whatever AT says, at least for the gaming, overclocking and power consumption figures, as they tend to fluctuate alot between review sites.

Low resolution: good cpu test
Low details: poor representation of real-world performance (playing with details pushed as far as an individual can reasonably push them).
 


Jed you really are a hopeless fanboy. P2 has the q6600 licked. Not too mention more power efficient. Where the q6600 edges out p2, at other times p2 urinates on the q6600. Do some reading, pay attention to things, get out more.

Peace
 
theirs about 12 sites that couldn't get these chips
to 4Ghz on air thats fact. the clock for clock you talk
about in games the Q6600 is ahead, between the
phenoms or right behind while running 400 to 600
Mhz decrease in clock speed .
now crank up the speed of the Q6600 to 3.0Ghz
equal to the top of the line Phenom II it will beat
them both.
 


i would rather have a P2 than a Q6600 any day of the week. nice try though.
 
Its been known for awhile, and as Anand has said, seen as well, the new rev will allow for higher ocing than what the reviwers were given. http://www.xtremesystems.org/forums/showthread.php?t=209491&page=49
So, its pointless to defend your comments as these chips are being sold and used now. If being happy for AMD users having these chips is hyping/spinning, then Im guilty. As for the ocing, where have I spun/lied/hyped? As per my guesstimate of 2-3% better than Kentsfield, well, its not over till its over, and either way, I wont be far off, and again, if thats spinning or lying, then again Im guilty, but not by much heheh. You just dont like hearing about AMD, and thats fine, but others do, and if I can bring some much needed good news for those users, then so be it.
 


Hopefully. My tone may be off, but im not bashing on the processor, im bashing on the price. $275 just cannot be justified in my eyes, at least to a person in my situation (aging, hot, greedy with voltage Q6600, and with new motherboard price aside) or anybody on comparable ground. Ive supported these chips in these threads up till now baring price, and I just think these prices are disasterous.
 
Possibly youre right. Im sure AMD will correct anything thats needed in pricing. This is a release price, as weve already seen the i920 in bargains, besides slight price reductions. Itll be interesting tio see who drops their prices first