Why is Core i7 920 better than Phenom 2 955

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
the diffrence between core i7 and phenom 955 isn't that big. altough there is a decent price diffrence and core i7 happends to scale a bit better in maxed out dual gpu configs. unless you are willing to pay alot extra and want to have everything the fastest possible phenom 955 dous fine. if you can get a phenom 955 around €170 yet it is a good buy if it is above €210 it is a rip off from that shop. both these cpus are ready for a couple of years to come and both there motherboards offer upgradability. for core i7 you can later plug in a gulf town. for phenom you just wait till a bit longer for the 32nm part but expact it to be cheaper then gulf town.
 
hmm... I like to be a little AMD biased but...

have been tempting myself lately and saw this little review here on Toms:

but AMD seem too far behind in a lot of benchies:

PhenomII @ 3.6GHz V I7 920 2.66GHz(stock) V I7920 @ 3.44GHz

PhenomII running 2 x 4890 and 4GB RAM
I7 System Running 2 x 260 and 6GB RAM

In most of those gaming Charts the higher clocked Phenom system gets left Far behind - and it has the stronger GFx setup.

I think in a system with a top of the line GFx setup or SLi / Xfire I7 is the only real option

A good Phenom system will be great for the more budget conscious setup

Based on these benchmarks at any rate...
 



What you posted is exactly the opposite in bias:

1. Tighter memory timings for Intel than for AMD.
2. Intel video cards at x16/x16. AMD video cards at x8/x8.
3. The NB should be tweaked as high as possible. (Any overclocker would do this automatically.)
4. Faster hard drive for Intel than for AMD.
5. We won't even go into the benchmark selection. (Just because the author claims the selection is not biased does not make it a fact.)


Oh WAIT... it's all okay be cause one system was made by a company and the other was hand built by him.
 
I was wondering about that comparison...

but my question is, does tighter timings or higher frequency matter more for these systems? (amd 2 x 2,048 MB,1,498 MHz, CAS 9-9-9-24 2T VS. intel 6gb DDR3-1378 at 1.56 V, CAS 8-8-8-19) i mean one can argue that intel is having a slower clocked ram and having 6gb of ram verses 4 gb, but the 2T deal for amd kinda makes me wondering about wtf is going on here....

where does it say the vid card at 16/16 and 8/8? i would love to find that out, although i honestly don't know if that limits the gfxes THAT much, but since they were top of the line gfxes, that may be worse than what i know/experienced.

edit: from the mobo's website, says that it has 1x 16 lane and 1x 8 lane with the 8 lane in 16x slot and it supports crossfire, so it would be x16/x8?

the manual tweaking vs production argument is what i was calling toms on, cuz when you mass produce something, you can't give it the attention and care of a home built unique machine.... would love to see either stock builds or both tweak builts

and on the hdd, good point there, the WD black are faster, and i wonder if that had some major impact in i/o intensive deals. i definitely missed that on the first run through of the review

and on bench marks, which ones are neutral and which ones are biased? i'm very much unaware and would love to know how they are bad (is it any game stamped with a twimtp a bad bench since the vid cards are different?)



 
ah, by my 1st line all I meant was that in a closeish matching I would go for AMD as a matter of choice rather than trying to qualify the validity of what was being said.

The content of that review to me seemed to weigh heavily on Intels side, meaning my ideal AMD choice would have been swayed to the much higher performing part - as apparent in the review.

and you comments on my post are most welcome and informative thanks :)

But it seems to me on point2 after a little reading:

"2. Intel video cards at x16/x16. AMD video cards at x8/x8. "
The AMD system is 2 PCI-E 2.0 so... so x8 pci-e 2.0 is like x16 pci-e 1.0.

They say that:
"PCI-E 2.0 x 8 is fine, you won't notice a performance decrease unless you have a GTX 295 or 4870 x2"

So on that front this review is fair.

So with the 4890 Xfire should not have caused a performance deficit to the Intel x16 slots it seems...
and if it was an issue I would have thought it could not match the "faster" x16 in Stalker Clear Sky benchmark, so I think that x8 and x16 here is a non issue:

Stalker Clear Sky

"4. Faster hard drive for Intel than for AMD. "
And was that even worth mentioning...
They would have run each benchmark multiple times to eliminate any such issues or lag spikes and used the avg... so the HDD speed would be irrelevant... and neither drive is "slow"

"5. We won't even go into the benchmark selection. (Just because the author claims the selection is not biased does not make it a fact.) "

Here it just seem you are being both a little inflammatory and misinformed as to your accusation of my post being misleading...


So all in all with a little more checking myself, I think my post was an interesting addition, though quite possibly there are a bunch of other benchmarks that may favour the Phenom more...

I just happen to be interested in some of those games and was a little surprised by the size of lead the I7 made



 
A bit late here. But i7 has internal Core Virtualization which allows processing power from other cores to be utilized in other processes....if that makes sense at all. I am personally drift towards AMD because of their Price/Performance Ratio. If you are looking for the TOP of the TOP of the line, then of course, i7 is going to be the better choice. If money is a factor, a Phenom may suit you better.

 


Good question on whether they ran multiple times... I tried to see briefly in the article but saw no mention about how they conducted that part of the tests

As my understanding goes, anyone running a serious benchmark set of tests do this to iron out any irregularities
I would be very surprised if they didn't... but really don't know for certain.

If they did, they the HDD issue does not count

I will have a bit more of a check about on that later :)
 


2. Using x16/x16 versus x8/x8 creates between a 3% and 5% difference in benchmark results. You are correct that it is not noticeable IN NORMAL USE. However a 3% to 5% speed difference in a comparative system benchmark is not acceptable if you even want to pretend you are professional and/or unbiased.

4. It is worth mentioning because is shows that this "review" was not subjective.

5. Let me summarize these benchmarks: "Hey the benchmarks that benefit from more threads actually do better with the processor that services more threads." How shocking.
 
Which is why a more thourough and encompassing re review needs to be done. If we saw this much threading in everyday use, if we only played games that favor nVidia arch, then Id agree, wed see this as the norm.
The 2T for the ram, choice of games, and the freedom of any user to tweak his/hes system accordingly to best case scenarioas was obviously done with the SBM, all should have been expounded upon more thouroughly as well, and not laid at the feet of AMD/ATI.
If I had that boutique setup, Id get the ram right, Id set the NB settings to max for that particular cpu (as ymmv), and Id play any games, not ones leaning in a particular direstion for HW. Then make the assessment. Even the reviewer said he was surprised by the disparity seen, and wants to do a followup, as Im sure hes not satisfied with what he showed.
He was hamstrung with the setup, not being able to choose particular ram etc, and choice of HW (HDD for example) I admire and respect the reviewer, but he still shouldnt have left it at AMDs feet like that without alot of caveats
 
here we go:

"2. Using x16/x16 versus x8/x8 creates between a 3% and 5% difference in benchmark results. You are correct that it is not noticeable IN NORMAL USE. However a 3% to 5% speed difference in a comparative system benchmark is not acceptable if you even want to pretend you are professional and/or unbiased."

Well in these benchmarks the I7 system is getting not too far off Double the performance ins some cases... so this 3 - 5% really does not change things much - but an interesting point I was not aware of nonetheless


"4. Faster hard drive for Intel than for AMD. "
"4. It is worth mentioning because is shows that this "review" was not subjective. "

According to Toms Benching processes:

Linky

Make sure not to use the results of the first cycle, as the increased I/O activity will produce distorted results.

and

In most cases, it doesn't even matter what kind of hard drive you use. However, audio or video encoding programs (Auto Gordian Knot, Lame, MP3 Maker) always need a lot of computing power, and in some cases a fast main memory also makes a difference.

Seeing as we are talking gaming benchmarks and they Do Not take 1st set results... the HDD plays No part in the result a data is cached to RAM by then!

So as I said point 4 is not worth mentioning...

"5. We won't even go into the benchmark selection. (Just because the author claims the selection is not biased does not make it a fact.) "
"5. Let me summarize these benchmarks: "Hey the benchmarks that benefit from more threads actually do better with the processor that services more threads." How shocking."

Hmm
Surprisingly we are looking to see which processor is better and we Do want to see this even if the I7 has an advantage with HT, after all that is the point --- we want to find out which is faster... Surprise Surprise...
and also seeing as we are looking at games ( often critisied for not handleing support for more than 2 core in reality ) what you say here is just plain wrong lol

I never claimed to be all knowing and even pointed out that maybe this set of benchmarks might overly favor the I7... but were interesting (and hence inviting more positive posts on the subject)

I only did a tiny bit of research to find most of what you said was basically rubbish or irrelevant - upon which you felt justified to put someone down

We are meant to be posting to share and gain knowledge and have fun, not try to shoot each other down
especially by someone like you who clearly knows very little about the subect we are talking
 
Id point out, if its cached to memory, and the memory is running 2T....

The boutique vendor was held somewhat blameless in this article, which is truly a shame. Also, the under emphasizing of homebuilt vs boutique (with the somewhat poor optimizations as well) was again, left at the feet of AMD, not the vendor.

In all fairness, one would need to build a 1300$ AMD rig, with all the tweaks, and better HW (due to pricing disparity of P2 vs i7) to acually then show the difference. To me, this is nothing more than a HP with tweaks vs a home built, but blaming AMD, not HP
 
Oh and a little more info froma little more research...

According to this Toms review on the difference in performance between x8 x16 etc...

Linky

In almost every test there was no where even close to a 3 - 5% difference

At least on that test setup.

Seeing as the bandwidth available is beyond most cards and in SLi or Xfire setups the board uses the same hardware there is no reason there should be a 3 - 5% difference.

if x8 was so bad I don't think the new I5 motherboards coming soon would "limit" themselves to this lowly standard

But there we go :)



Stop trying to shoot people down and post something a little more useful... 😉
 
Read the uses planned for the enthusiasts here for i5. They DONT recommend i5 for xf/sli. And the next coming gen of gfx card is sure to only challenge the BW or a 8x slot even more so, besides what we already see.
You arent singing to the choir by saying i5 is fully capable in those scenarios, not here in Toms forums anyways
 
I was only considering th I5 platform for a single card solution in this context, just that it must be considered that X8 was considered enough for current and next gen cards - for which rumours say may be as much a twice the power of current ones...

though I imagine from what I have read it will maybe start to become a limiting factor with the top of the line cards maybe ?? But then it is not aimed at the enthusiast market so much.

Ah well, am learning lot on areas I have not researched before :)
 
Currently, like you said, x2 solutions do show loss at 8x, and depending on game, even some single core solutions as well. Games use such a variety of things, requiring such a wide usage of gpu/cpu/bw etc, and shouldnt be the determining factor for cpu benches anyways, unless theres a wide amount of types of games used. So, this in itself doesnt truly give a decent picture of either cpus, since the games was just a small sample.
It is telling, but to what direction? AMD? Intel? Games? Lack of data? or amount of games? Particular games used? Favoring which HW? Build? HW choices? Tweaking?
Thats my point, to have it all laid at the feet of AMD is somewhat misleading
 
Im looking forwards to his followup.
Take a look at FC2, and how it favors Intel cpus. Look at Crysis, and how it favors nVidia cards, as well as many other games. Look at GRID, and see how it favors ATI's arch. All clues to either actual HW design, or, in some instances, pure marketing/dev creations, ala Assassins Creed DX10.1 removal
 


Wow. You picked an article that shows 3dMark06 and a bunch of game that are already bottlenecked at both x8 and x16 for your "proof".

What you need to do is focus on something that moves a LOT of textures. Oh wait... your article shows one: The MS FlightSim. If I look at that then I discover that I was WRONG. That shows a 17%-21% difference between x8/x8 and x16/x16. So my 3%-5% number was a too low.

That showed that on bottlenecked games it won't matter much between x8 and x16. But on some things it could matter by as much as a 20%. (My personal experience moving from x16/x8 to x8/x8 was a only 6% difference in 3dMark Vantage GPU scores. Other scores varied in results. Ergo my conservative 3%-5% figures.)

Thanks for showing me that my numbers were WAY too conservative.

 
hmm...

maybe the Flight Sim low performance is also linked to having not enough memory on the card, 1GB for a dual core card - so effective 512MB cards

so having to utilise main system memory as they mentioned... something Well known for lowering your FPS

The extra stress put on the x8 bandwidth of additional access to system memory due to too low a card memory to hold the texture distort the result...

something I considered when I pointed out the results.

As such it Does not show that the X8 would hold back a strong card, with enough memory to not have to resort to using main system memory to store textures.

This is something Everyone knows in chooseing a GFx card - you want to run High res or extra large textures, you get a card with more than 512MB memory as appropriate.

So what I say still stands...


Think about what you write before you get all excited again please.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom