Phenom II 955 Versus Core i7 920: Gaming Value Compared

[JimmiG]

One thing to keep in mind is that you can also overclock the GPU/video RAM. If the 4890's are running closer to the limit, the i7 system would catch up in terms of GPU power as well.

 

[JAYDEEJOHN]

Id point out, using AA, the P2 wins a few, and if you really want a setup like this, youll be using the AA. So, throw out the first set of numbers where theres no AA or eyecandt, and use the second, youll see P2 is much closer than you think

 

[dattimr]

Great review, but it's greater knowing that Tom has been paying quite some attention to external feedback coming directly from its readers. We usually don't get useful responses from reviewers when we comment about something that goes against their own opinions (not even an explanation as to why they should be considered better than what we suggested). It's even more rare for a whole new article to be solely based on this same feedback.

 

[kartu]

[citation][nom]pulasky[/nom]This crap site is worse than expected why http://www.tomshardware.com/review [...] 60-12.htmlvshttp://www.tomshardware.com/review [...] 350-7.htmlwhat the hell, one single 4870 is faster than 2 4870 or 2 4890 (it seems that crossX do not work in any game tested) some explication is more than welcome.[/citation]
I'd like to hear official explanation on this.

 

[apache_lives]

show me benchmarks with my usual apps running - voice chat, torrents, messenger apps, antivirus, and a slightly cluttered system - paint us a true image of that things are like, not some fresh rig with fresh perfect install.

 

[sohei]

[citation][nom]kartu[/nom]I'd like to hear official explanation on this.[/citation]


you know the official explanation ..."sorry guys bla bla"

 

[dirtmountain]

Thanks Cleeve for going back and doing the work on this system. I had thought that the Cyberpowers numbers were off and with a more level playing field an AMD build performed well.

 

[JimmiG]

[citation][nom]apache_lives[/nom]show me benchmarks with my usual apps running - voice chat, torrents, messenger apps, antivirus, and a slightly cluttered system - paint us a true image of that things are like, not some fresh rig with fresh perfect install.[/citation]

No game uses four cores fully, or even three. That doesn't mean that quad-core is useless for gamers, precisely because you can leave background processes running without them affecting performance. So the games would not be any more CPU-bound on a cluttered system. Loading times might increase because of fragmentation etc. and background file accesses, but that wouldn't show in benchmarks.

 

[kevin1212]

Excellent article... covered every corner.

 

[cimtaurus]

Very good article.
I recently built an i7 system and I can say that the performance is incredible.

Two interesting points mentioned above:
Gamer systems are always going to run with all of the eye candy maxed out, that is the point of building a gaming rig. Dump all of the other test numbers and focus on full AA and AF.
Gaming machines typically also run things like anti-virus, voice chat, a mail client, a non-IE browser, etc. Show us some benchmarks for a system in more of a "real world" situation as well as these clean install set ups. Having both would give us a better comparison.

My final decision to go with the i7 was based on more personal preference than anything else. The fact that I can overclock my i7 so much and so easily is what tipped the scales in Intel's direction for me.

Again, good article and thanks for doing the research.

 

[radnor]

[citation][nom]pulasky[/nom]This crap site is worse than expected why http://www.tomshardware.com/review [...] 60-12.htmlvshttp://www.tomshardware.com/review [...] 350-7.htmlwhat the hell, one single 4870 is faster than 2 4870 or 2 4890 (it seems that crossX do not work in any game tested) some explication is more than welcome.[/citation]

I would aswell. I'm in the middle of a upgrade,and rocking a 4850 CF with a 4800 x2 (Oceed to 3.0Ghz). There was great diferent in putting the second 4850. Basicily the frame rate is the same in the same situations, just only ALL bells and whistles are on. Although i reckon is a higly cpu dependant game, CF makes a diference.

 

[skora]

I wish everyone would stop screaming AMD INTEL AMD INTEL. We each have our own set of priorities. Some its maximizing performance/price value, some its sheer performance, and others are based on price. This article only examines how much computer you can get for a specific price point. The trade offs at this price point is More CPU/ less GPU, or less GPU/more CPU. Just because the intel system did better in the benchmarks, doesn't mean they have the most value, just that its the most system you can get at this price point. AMD will have its victories on lower price points and that's where their market strategy seems to be. The next SBM could be set at a price point just below where an i7 could be afforded and I'm sure AMD will be sitting in the case. But please stop with the THG hates AMD and learn how to interpet perspective. Also realize the rest of the world doesn't always share your priorities and accept that articles are written for other people too.

 

[mrwoody]

I really enjoy reading these articles, and am very impressed to see an author go back and actually respond to criticism. It's very nice to see, when generally responses are ignored .

 

[hungryjackal]

I have to say, though I have never complained about an article before, this goes a long way in making me feel like we the readers are listened to. I appreciate that Tom's took the time to do this for us.

 

[brendano257]

I would like to see how i7 wins with the 4890 pairs, not just price/performance, but raw performance, what most people buy Intel for, not concerned with the price.

 

[Miharu]

I don't think we can compare an AMD 4890 card in a AMD Phenom PC and an Geforce 260s with a Intel i7, telling AMD win in gaming ?!?

If you can't use the same video card and just change mb/CPU... it's not an benchmark for compare "Phenom II 955 Versus Core i7 920: Gaming Value Compared".

i7 with 4890
i7 with 260
VS
Phenon II with 4890
Phenon II with 260

and you compare the cost at then end (in %).

I deserve a good measure of blame for rationalizing...

The conclusion is... this is not a rational benchmark, someone need to redo it with rational benchmark.

 

[heltoupee]

[citation][nom]lashton[/nom]its annoying to see that toms hardware seems to be very pro intel[/citation]
Tom's is doing exactly what it should be doing -- exactly what my uncle tries to do at the racetrack -- pick the fastest horse. It's not Tom's fault that AMD doesn't compete well at the high end anymore. Actually, it seems to be AMD's strategy right now to take over the budget market, and let Intel and their i7 take the high end crown.

 

[wh3resmycar]

intel i will buy p2x3 and tom articles has no value for me because they are 5 people and they don't use a cpu like an end user ... they use a cpu just for rating and i spent my own money not their money ......open your eyes

i opened my eyes and saw a pile of crap on top of your post. seriously why are you here?

 

[theholylancer]

[citation][nom]wisdom_learner[/nom]Wow. You take things way too seriously.[/citation]

tell that to the people whom bashed Pentium 4s when A64 was the king, tell that to the people whom bashed nvidia when 3dfx was the king, tell that to....

this is the internet, when you get burned enough times, you too demand swift retribution when you get the chance (at times, the only chance before what ever you liked goes under)

 

[Onus]

I thought this article was very good. Not an end-all, be-all, which no one in his right mind should expect, but another good impression to file away in the "index."

 

[scook9]

Good article, this is an excellent comparison of the different platforms. I would also have liked to seen the same thing done but with SLI this time as well. And then see the 2 articles compared.

My only gripe with this article, why the little overclock on the i7 920? You pushed the P2 955 to its limits, do the same to the i7 (that would be 4 GHz for the average overclocker). This would tip the scales heavily in Intel's favor, despite what all the readers are saying in these comments.

Everyone who reads this article needs to be aware that the Intel cpu would easily win if it had a REAL OC on it. AMD pulled a close win with a larger OC than the Intel. Also note that the AMD needed a lot more voltage than the Intel = less headroom for further OC.

 

[sandmanwn]

[citation][nom]pulasky[/nom]This crap site is worse than expected why http://www.tomshardware.com/review [...] 60-12.htmlvshttp://www.tomshardware.com/review [...] 350-7.htmlwhat the hell, one single 4870 is faster than 2 4870 or 2 4890 (it seems that crossX do not work in any game tested) some explication is more than welcome.[/citation]
Seems like an odd number. Maybe its one of those games that does terrible at utilizing CF or SLI.

 

[spoondigity]

I agree with scook9. They previously limited the overclock on the i7 920 because of the mini case they were putting it into. At this point, they should have overclocked it further without the limits of a small case and heat issues.

Maybe they figured this would be a good base point comparison and decided to leave it up to the reader to understand that this is a 'best case' scenario for that particular AMD chip in the matchup (since it is going up against a 'gimped' core i7 overclock).

 

[thearm]

For all the AMD fanboys: Suck it.

 

[glawk]

Why you guys keep including games like Stalker and HAWX is beyond me. A quick peek at the xfire and steam statistics will tell you that no one is playing these games!

Steam only lists the top 75 played games and neither game makes that list. Among xfire users, neither game is in the top 200 in hours played! What gives? Please start including more relevant games in your benchmarks.