AMD Phenom II X4 920 and 940 review [Guru3D]

[Guru3D News]

AMD today unleashes their Dragon and though that might sound like a pun it's actually an AMD marketing phrase for the infrastructure that carries Phenom II an AMD 790 chipset based motherboard Radeon...

More...

[kylzer]

That it was one of the most in-depth cpu reviews i'v seen in a long while overall i'm very impressed its not that far off the i7 but in general the prices are good i guess the guys with AM2+ boards are going to have grin on there faces

[kidswithguns]

http://www.techradar.com/news/comput...i7-down-480468

Reviewer might want to take a look at this.

The Core i7 performs almost the same as E8000 in gaming benchmarks.

Because, Vista thinks Hyper Threading = 8 true cores, which mean if your games support 2 threads, it will run on 1 core, and 1 virtual core.

[sbrehm72255]

Originally Posted by kylzer

That it was one of the most in-depth cpu reviews i'v seen in a long while overall i'm very impressed its not that far off the i7 but in general the prices are good i guess the guys with AM2+ boards are going to have grin on there faces

Performance between the i7 920 and the AMD 940 clock for clock isn't even close. ..............

[Aura89]

Originally Posted by sbrehm72255

Performance between the i7 920 and the AMD 940 clock for clock isn't even close. ..............

shh, go away intel fanboi, it depends on the situation, as if you read the review, it's true in some instances, and gets beaten by a long shot in many instances

[Denial]

If you look at the game benchmarks they are all virtually the same with a few fps difference here and there - and lets be honest, who cares if winrar takes 30 seconds longer to finish.. it's all about FPS. =d

[Aura89]

Originally Posted by Denial

If you like a the game benchmarks they are all virtually the same with a few fps difference here and there - and lets be honest, who cares if winrar takes 30 seconds longer to finish.. it's all about FPS. =d

true dat

[krayzieNRG]

come on now ,that's on a am2+ mobo with DDR2 and getting the performance that really great in my book..well done AMD,atleast it doesn't take a change of almost everything to get one unlike i7.

[RejZoR]

Originally Posted by kidswithguns

http://www.techradar.com/news/comput...i7-down-480468

Reviewer might want to take a look at this.

The Core i7 performs almost the same as E8000 in gaming benchmarks.

Because, Vista thinks Hyper Threading = 8 true cores, which mean if your games support 2 threads, it will run on 1 core, and 1 virtual core.

Erm, no. If game originally runs using 2 threads, i7 will run the game on 4 cores. 2 physical and 2 virtual, each attached to 1 physical core.

[Cartman372]

Sweet, reading the review right now. Maybe it gives me something to upgrade to ^^

Could upgrade my Phenom X4 9850 to a Phenom II X4 940

[Stukov]

It looks as though this is the best bang for your buck quad core at 1600+ resolution. Were there no temp tests? And was there a comparison of power consumption that I missed?

[K.I.T.T.]

powerhouse my a**. Its about as fast as my E6750 according the review at Tomshardware. Think AMD heading the way of the Dodo.

[Knox]

Originally Posted by K.I.T.T.

powerhouse my a**. Its about as fast as my E6750 according the review at Tomshardware. Think AMD heading the way of the Dodo.

LOL you said, tom's hardware.

IMO there just as reputable as Fudzilla or the Inquirer.

Reguardless for there price range, and decent overclocks they look like good bang for the buck.

[DES_MX]

Originally Posted by Knox

LOL you said, tom's hardware.

IMO there just as reputable as Fudzilla or the Inquirer.

Reguardless for there price range, and decent overclocks they look like good bang for the buck.

Well i have to say Hilbert is also a bit exaggerating the Hyperthreading impact (which on average gives -1 to 1% boost across multimedia, math, professional, and game benchmarks). So Intel's HTT isn't really that great. Turbo Boost is, since it gives at least 1 free multiplier bump. E.g. on i 920 from 20x to 21x.
Thats why no one notices the negative impacts of HTT, since it's usually enabled together with Turbo.

Anyways, comparing highly OC'd Phenom @ 3.8ghz vs intel at i7 @ 2.66, which is more than 1ghz slower is just UNFAIR. e.g mixing apples and pears. Pro's don't do that.

OC vs OC
non-OC vs non-OC

thats the way to go

otherwise it's punching below the waist. Hilbert.

[noteleet]

The real steal is the Phenom II 920 imo, at 1600 res it was on par or ridiculously close with both the Phenom II 940 and the i7 920.

[wud03]

True, but this is a review of the Phenom II's. If you want to compare it to OC'ed i7s, write down the numbers and compare it to OC'ed i7's in earlier reviews. Just read the graphs; it tells the truth. Hilbert's conclusion makes sense. He goes out of the way to tell the readers i7 is more powerful, but points out that Phenom IIs are a great value. I don't see the problem.

Originally Posted by DES_MX

Anyways, comparing highly OC'd Phenom @ 3.8ghz vs intel at i7 @ 2.66, which is more than 1ghz slower is just UNFAIR. e.g mixing apples and pears. Pro's don't do that.

OC vs OC
non-OC vs non-OC

thats the way to go

otherwise it's punching below the waste. Hilbert.

[Tat3]

What I was interested was gaming at high res and there Phenom2 X4 940 @ 3.8GHz did very good job.

I really dont have very much extra money atm, but I hope to upgrade in next month or two and this looks good

[DES_MX]

Originally Posted by wud03

True, but this is a review of the Phenom II's. If you want to compare it to OC'ed i7s, write down the numbers and compare it to OC'ed i7's in earlier reviews. Just read the graphs; it tells the truth. Hilbert's conclusion makes sense. He goes out of the way to tell the readers i7 is more powerful, but points out that Phenom IIs are a great value. I don't see the problem.

I don't see a problem either, except for the fact that it is all too well used by reviewers to bias the results in either way. So in effect I have more of structured critique than a "problem". And besides, some things are just unethical. So to boil it down:

1) HTT on i7 platform DOES NOT benefit on average. (taking different benchmark suites and averaging out the performance gains. In Mathematica and MatLab for example i7 with HTT can deliver up to -50% performance DECREASE)
2) Turbo does. 1x free multiplier, nuff said.
3) Give us a clock-for-clock performance comparisson.

OR

4) Give the "nominal" comparisson complemented by max overclock comparisson. e.g. same price range cpu's (indeed i7 920 vs Q9550 vs Phenom 940) and then overclock each platform to it's max stable clocks and compare again. Simple, yet very, very revealing to each group of users.

[Hilbert Hagedoorn]

Originally Posted by DES_MX

I don't see a problem either, except for the fact that it is all too well used by reviewers to bias the results in either way. So in effect I have more of structured critique than a "problem". And besides, some things are just unethical. So to boil it down:

Are you done bitching like a little kid in every thread I run into ? Here's my answer to your HT remarks.

http://www.guru3d.com/article/intel-...-965-review/10

If our reviews do not suit your needs ... then read them elsewhere.

Sjeesh

[HeavyHemi]

Originally Posted by DES_MX

Well i have to say Hilbert is also a bit exaggerating the Hyperthreading impact (which on average gives -1 to 1% boost across multimedia, math, professional, and game benchmarks). So Intel's HTT isn't really that great. Turbo Boost is, since it gives at least 1 free multiplier bump. E.g. on i 920 from 20x to 21x.
Thats why no one notices the negative impacts of HTT, since it's usually enabled together with Turbo.

Anyways, comparing highly OC'd Phenom @ 3.8ghz vs intel at i7 @ 2.66, which is more than 1ghz slower is just UNFAIR. e.g mixing apples and pears. Pro's don't do that.

OC vs OC
non-OC vs non-OC

thats the way to go

otherwise it's punching below the waist. Hilbert.

Of course if you had actually read the entire article you would have noticed that every single one of the benchmarks included the stock 940 and OC'd 940 versus the others for comparison. So there was non OC vs non OC. You're implying you're a a pro huh? Schwing and a miss dude.

[HeavyHemi]

Originally Posted by Hilbert Hagedoorn

Are you done bitching like a little kid in every thread I run into ? Here's my answer to your HT remarks.

http://www.guru3d.com/article/intel-...-965-review/10

If our reviews do not suit your needs ... then read them elsewhere.

Sjeesh

He is so much of a 'PRO' he missed your non OC to non OC comparison in every benchmark. Perhaps he can link us to his professional reviews eh? LOL What a nack.

[DES_MX]

Originally Posted by Hilbert Hagedoorn

Are you done bitching like a little kid in every thread I run into ? Here's my answer to your HT remarks.

http://www.guru3d.com/article/intel-...-965-review/10

If our reviews do not suit your needs ... then read them elsewhere.

Sjeesh

First of all, Hilbert, I am not a little kid. Second - I am not bitching by any account, but pointing out how and why the article seems biased. Third, before anyone will start using the "fanboy" labels, I have been switchinb back and forth from Intel to AMD and back to Intel for several years now.

As to your rather *rude* response, I would like to give a few counter-arguments:

1) You probably just got angry and didn't pay attention to my post at all, since I was saying that ON AVERAGE the HTT does not give any benefit to performance. Now you just gave me a link to your own review in KribiBench, where indeed IT DOES. Ironically that's a SINGLE bench result, and HTT at ITS BEST.

To oppose this, I would like to repeat again that if you take an average of different purpose test suites (Matlab, Mathematica, 7zip, WinRar, FLAC, Lame, XVid, Divx, PHP Calculator, PHPSpeed, Synthetic MySQL, Call of Duty 4, Company of Heroes, Call of Juarez, Crysis, S.T.A.L.K.E.R., Unreal Tournament 3, World in Conflict) you gain around 0% increase in performance.

2) To back my words up, I include a link on an in-depth performance testing on iXBT.com, a very respectable Russian hardware portal. I do not think it needs any further introduction to you.

http://www.ixbt.com/cpu/ci7-turbo-ht.shtml

Now the review is initially in Russian, but the tables are all English-based, and include all the info needed to comprehend.

To make things easier, here is the same page translated into English by google:

http://translate.google.com/translat...hl=EN&ie=UTF-8

I hope everyone can see these results and judge for themselves. As for you Hilbert, I am very disappointed by your rather rude and categorical response, with unstructured insults. By the way I do not think that being 23 and having one BSc and one MSc title in general economics, business studies at the University of Amsterdam no less makes me a lil bitching kid.

Originally Posted by HeavyHemi

Of course if you had actually read the entire article you would have noticed that every single one of the benchmarks included the stock 940 and OC'd 940 versus the others for comparison. So there was non OC vs non OC. You're implying you're a a pro huh? Schwing and a miss dude.

To answer you as well, no I am not implying I am, by no account. I do imply that after reading hardware reviews for many, many years one does get a feeling of whats done properly, and what is not.

In fact, Hilbert does include non-OC and non-OC results of i7 vs 940. On the other hand he does NOT include the results of OC'd i7 vs OC'd 940. The view I take on this is not only CPU wise, but PLATFORM wise (e.g X58 vs 790FX/GX) as well. I hope I am being clear now.

[hallryu]

Nice to see AMD on the up. This new CPU will be serious contender for all the folks that don't want a complete overhaul and all the expense and heart ache that comes with it! Justifying the cost to the Mrs, buyers remorse, the disappointment when thing run just slightly quicker than before.

Good review and nice one AMD.

[kylzer]

@ DES MX

Hilberts probably pissed cause that review looked like it had taken a while to do and then gets ridiculed WTF man thats not nice.

[Hilbert Hagedoorn]

Originally Posted by DES_MX

As to your rather *rude* response, I would like to give a few counter-arguments.

Last response: This is a fair and objective written review. The fact that you look at it from another POV doesn't make your opinion right. Quoting one source is a little too limited really.

Anyway, so be it. But now please move on before you get into an arguement with me.