AMD Phenom II X4 955 black or i5 750/760

[jsmessick]

I am trying to decide on a CPU for a New Desktop I would like to build. I will be playing some games (starcraft II, Guild Wars 2, MW2, Counterstrike) but I would also like to watch HD movies and do school work easily. Please include your choice and why this would be better in the comments. Thanks and if you have any other suggestions on hard drives feel free to include them but I have a budget of $1000.

AMD Phenom II X4 Black Edition:

http://www.newegg.com/Product/Product.aspx?Item=N82E16819103808

Intel i5 750

http://www.newegg.com/Product/Product.aspx?Item=N82E16819115215

Intel i5 760

http://www.newegg.com/Product/Product.aspx?Item=N82E16819115067

 

[enzo matrix]

The Phenom II is more than you need, unless you have crossfire 5870s or better for graphics. Even then it is arguable. Save yourself the money.

As for hard drives, the standard suggestion is to go with the Samsung spinpoint F3 or alternatively the WD Black. But to give you an exact model, I need to know how much space you need.

 

[jsmessick]

Thanks! My question about hard drives would be would I be better off Getting like an 80gb system drive and then a 500GB or 1TB storage drive. Or can i save myself the $30-$40 and use it on a video card or RAM and just buy a 1 TB drive?

 

[iqvl]

@Enzo Matrix:
Phenom II will even bottleneck 2xHD4870 in CF.

@jsmessick:
750/760 is the way to go if you want the best gaming performance.
Proof? Scroll down to the very bottom of the comparison benchmarks: http://www.anandtech.com/bench/Product/147?vs=109

 

[jsmessick]

Thanks which would be a better choice 750 or 760?

 

[iqvl]

Thanks which would be a better choice 750 or 760?

750 in my opinion as the extra 1 multiplier of 760 doesn't worth $10.

 

[jsmessick]

Alright thanks alot! Do you have any advice on my Hard drive Situation?

 

[iqvl]

Get a WD 1TB/64MB Black HDD and then add in a SSD later when it gets cheaper in Q4.

 

[jsmessick]

Best answer selected by jsmessick.

 

[AMW1011]

@Enzo Matrix:
Phenom II will even bottleneck 2xHD4870 in CF.

@jsmessick:
750/760 is the way to go if you want the best gaming performance.
Proof? Scroll down to the very bottom of the comparison benchmarks: http://www.anandtech.com/bench/Product/147?vs=109

Wow, no. A Phenom II won't bottleneck 2 4870s in crossfire. Also that chart is hardly much evidence as it isn't comparing everything equally. It is common knowledge that the Phenom II performs unnoticeably slower than the Intel i5/i7 in single and dual card setups. Unless the OP is going for more than 2 GPUs he has no reason to buy the i5/i7 over the Phenom II, as the performance is incredibly close and Intel's advantage isn't worth the extra cost.

Oh and before you go and start calling bias, I am currently using an i5 750 so its not like I'm an AMD fanboy, just a realist.

http://www.legionhardware.com/articles_pages/radeon_hd_5870_crossfire_cpu_scaling_performance_part_1,1.html

Something worth considering is that although the Core i7 clearly is the superior processor, can all that extra horse power be converted into better gaming performance? When we conducted our CPU scaling articles with the Radeon HD 5970 we found that the extra power of the Core i7 was of no benefit. This was due to the heavy use of AA/AF, which placed limitations of the GPU’s before the CPU.

Therefore it could be said that when enabling the kind of quality settings that owners of Radeon HD 5970 or Radeon HD 5870 Corssfire configuration would generally play with, there is almost no difference in performance between the Phenom II X4 and Core i7 processors.

Extra performance that is unnoticeable is performance wasted. The Phenom II X4 955 an be had for $160 now, and the 925 for $130. Why would he spend any more? Not to mention the $100+ saved on the AMD build can go into his GPU or such, which WILL make a noticeable difference.

 

[iqvl]

Check TOM's benchmark and you will find that 955+2x48'9'0 loses to i7+2x48'7'0. How can this happen if there is no bottleneck? Huge bottleneck!

 

[iqvl]

An example from http://www.tomshardware.com/reviews/phenom-versus-i7,2360.html :

 

[AMW1011]

An example from http://www.tomshardware.com/reviews/phenom-versus-i7,2360.html :

http://media.bestofmicro.com/T/W/216212/original/PhII vs. i7 World in Conflict.png

You can cherry pick all you want, bro. How much of a tangible difference will you notice with the i7 over the 955? How about we look at a more demanding game?

Also, look at the higher resolutions, unless buying two high end GPUs and a high end motherboard, CPU, and RAM while gaming at 1280x1024 makes sense to you.

Well isn't that interesting... The most graphically intensive game that is the standard still today for DX10 performance runs better on the Phenoms...

Why don't we try to get your benchmark under 60 FPS:

Can barely do it, though its funny how they are almost completely identical at 192x1200...

Wow, at higher resolutions the Phenom beats i7, again.

Hmm both approaching 100 FPS at high resolution... Looks pretty similar to me.

Hmm, both well over 60 FPS again...

Oh no! Phenom beating or tying the i7 again!

Wow, your right! If the OP only wants to play Prototype then the choice is clear. I mean sure its a horribly optimized console port and really doesn't show anything probative. But what do I know?

Notice the largest difference, except for Prototype, at 1920x1200 is 10%. That is less than 2 FPS more than the Phenom if we are comparing to a theoretical 29 FPS (considered almost unplayable). It is also less than 5 FPS when comparing a theoretical 59 FPS (Considered under maxed out).

Show me one example from that article that shows a noticeable performance difference in a game that isn't Prototype.

 

[iqvl]

Cherry picking?!

From the benchs, 955+2x48"9"0 loses to 920+2x48"7"0 in most cases , is it?!

The best case for 955 setup is only a little FPS more than 920 one which is far worse than it should perform, because 2x4890 is significantly better than 2x4870.

The performance difference is not enough?! Are you serious? You get LESS fps on 955 even paired with two BETTER GPUs! WOW, 955 is actually beating i7 up
to the pulpe if i7 were paired with 8400GS instead.


Imagine what would happen if both 955 and 920 are paired with two SAME gfx.

You don't see huge difference ONLY IF YOU SPEND MORE MONEY ON GFX FOR 955 SETUP!

 

[AMW1011]

Cherry picking?!

From the benchs, 955+2x48"9"0 loses to 920+2x48"7"0 in most cases , is it?!

The best case for 955 setup is only a little FPS more than 920 one which is far worse than it should perform, because 2x4890 is significantly better than 2x4870.

The performance difference is not enough?! Are you serious? You get LESS fps on 955 even paired with two better GPUs!

Imagine what would happen if both 955 and 920 are paired with two SAME gfx.

You don't see huge difference ONLY IF YOU SPEND MORE MONEY ON GFX FOR 955 SETUP!

Your missing the point. The differences in percents are so low that it doesn't amount to much at all. The phenom II does about as well as it should, when the FPS aren't quite as high and the GPUs become the heavy bottleneck, we see the 4890s pull ahead, when the FPS are really high and the GPUs aren't a heavy bottleneck we see things run together or the i7 pull slightly ahead.

The point is that in a circumstance when a discernible difference can be seen, the difference will be very small because the bottleneck will move on to the GPU instead of the CPU.

I apologize for the flaming earlier, that isn't my intention. I just want to explain that the Phenom II setup won't be noticeably slower than the i7 setup in high end gaming with one or two cards.

 

[iqvl]

Two questions.

1. 2x4890 is significantly better than 2x4870, is it?
2. 955+2x4890 loses to 920+2x4870 in most cases, is it?

If you still think there is no bottleneck on 955, then so be it.

 

[AMW1011]

Two questions.

1. 2x4890 is significantly better than 2x4870, is it?
2. 955+2x4890 loses to 920+2x4870 in most cases, is it?

If you still think there is no bottleneck on 955, then so be it.

1. A 4890 is 10% faster than a 4870.
2. In 3 out of the 7 tests, the 955+4890 win, of which are the most demanding games. In every other instance, extremely high FPS, the bottleneck has shifted away from the GPU where the CPU is the factor, where the 955 loses. Where the GPU is the bottleneck the 955+4890s win.

I'm not sure whats so confusing.

For the 955 to bottleneck these GPUs, the GPUs would have to be under-performing where they are the bottleneck, not where the CPU is the bottleneck. We do not see this happening, therefor the 955 is not bottlenecking these GPUs.

 

[iqvl]

1. A 4890 is 10% faster than a 4870.
2. In 3 out of the 7 tests, the 955+4890 win, of which are the most demanding games. In every other instance, extremely high FPS, the bottleneck has shifted away from the GPU where the CPU is the factor, where the 955 loses. Where the GPU is the bottleneck the 955+4890s win.

I'm not sure whats so confusing.

For the 955 to bottleneck these GPUs, the GPUs would have to be under-performing where they are the bottleneck, not where the CPU is the bottleneck. We do not see this happening, therefor the 955 is not bottlenecking these GPUs.

Does it make sense for under-performed 4890 to lose to under-performed 4870?

No matter there is a GPU bottleneck or not, 4890 should always performs better than 4870, agree?

 

[iqvl]

Another example:

Both HD5850 and HD5750 are the bottleneck of system for Crysis under 2560x1600. Do you think HD5850 would get less FPS than HD5750 given that there is no CPU bottleneck?

 

[AMW1011]

Does it make sense for under-performed 4890 to lose to under-performed 4870?

No matter there is a GPU bottleneck or not, 4890 should always performs better than 4870, agree?

And they always do, if you count only when the extra performance CAN be noticed.

I hate to break this to you, but it doesn't matter whether you have a 4870 or a GTX 480, Half Life is going to play the same. Does that mean the difference is not there?

Where it counts, we see the 4890 winning. When there is no need for the extra GPU performance and where it is not the focus, we see the CPU become the focus and we can both agree that in that instance the i7 is faster. However this instance only happens at very low resolutions or when there are a ridiculous amount of FPS as it is and the GPUs are not working anywhere near 100%.

 

[iqvl]

And they always do, if you count only when the extra performance CAN be noticed.

I hate to break this to you, but it doesn't matter whether you have a 4870 or a GTX 480, Half Life is going to play the same. Does that mean the difference is not there?

Where it counts, we see the 4890 winning. When there is no need for the extra GPU performance and where it is not the focus, we see the CPU become the focus and we can both agree that in that instance the i7 is faster. However this instance only happens at very low resolutions or when there are a ridiculous amount of FPS as it is and the GPUs are not working anywhere near 100%.

Let say, both HD5850 and HD5750 are the bottleneck of system for Crysis under 2560x1600. Do you think HD5850 would get less FPS than HD5750 given that there is no CPU bottleneck? Hmmmm, maybe peoples who can afford 2560x1600 monitors should get 6600GT instead. Worse GPU FTW for high resolution monitors!

 

[towjm]

Why are the two of you going on and on over the difference between an i7 920 (a $300 part) and the AMD part (about half as much), when the OP is looking at i5 parts vs. the AMD part?
According to Tom's CPU charts, the i5's may very slightly outdo the 955 but cost 25-30% more.

 

[mindless728]

@ iqvl, I tested the PII bottleneck with 2 x 4870 1GB's myself, going from stock PII X4 810 (2.6GHz) to 3.198 GHz netted maybe 1 fps in games with AA/AF enabled @ 1080p which is within margin of error (games tested: Battlefield: Bad Company 2, Crysis, Metro 2033)

EDIT:

Why are the two of you going on and on over the difference between an i7 920 (a $300 part) and the AMD part (about half as much), when the OP is looking at i5 parts vs. the AMD part?
According to Tom's CPU charts, the i5's may very slightly outdo the 955 but cost 25-30% more.

because we are trying to show worst case scenario against AMD, between the i5 and PII is even less (takes either extremely high settings or fps way over 90 to see a difference)

 

[AMW1011]

Let say, both HD5850 and HD5750 are the bottleneck of system for Crysis under 2560x1600. Do you think HD5850 would get less FPS than HD5750 given that there is no CPU bottleneck? Hmmmm, maybe peoples who can afford 2560x1600 monitors should get 6600GT instead. Worse GPU FTW for high resolution monitors!

That makes no sense.

If we went by your criteria for bottlenecking, then the i7 would too because eventually it will as you relieve the focus on the GPU.

Where these modern CPUs bottleneck the GPUs is at extremely low resolutions and at extremely high FPS. This means absolutely nothing. At higher resolutions, with less than 60 FPS, and in future games, where the performance is needed and noticeable, the GPUs will be taxed a lot more, and this is where no CPU bottleneck will occur.

Now add in that you can OVERCLOCK these CPUs still and it becomes really obvious that there will be no bottlenecking.

You can point at results with low resolutions at over 100 FPS, but you will be missing the point completely. If you look at Tom's article again, you will see that it supports this as I've stated above.

 

[iqvl]

That makes no sense.

If we went by your criteria for bottlenecking, then the i7 would too because eventually it will as you relieve the focus on the GPU.

Where these modern CPUs bottleneck the GPUs is at extremely low resolutions and at extremely high FPS. This means absolutely nothing. At higher resolutions, with less than 60 FPS, and in future games, where the performance is needed and noticeable, the GPUs will be taxed a lot more, and this is where no CPU bottleneck will occur.

Now add in that you can OVERCLOCK these CPUs still and it becomes really obvious that there will be no bottlenecking.

You can point at results with low resolutions at over 100 FPS, but you will be missing the point completely. If you look at Tom's article again, you will see that it supports this as I've stated above.

Do you have any kind of common sense? Don't think in such a complex way.

Just think if it is possible for better GPU to get less fps worse one given that there is no CPU bottleneck.

I will state the common sense one more time. Given that there is NO CPU bottleneck, better GPU will always get higher fps.
In other words, better GPU will only get less fps when there is CPU bottleneck on the better GPU and not on worse CPU.

I really don't understand why you don't get such an easy logic. You really beats me up with your own logic. I quit now.

Why are the two of you going on and on over the difference between an i7 920 (a $300 part) and the AMD part (about half as much), when the OP is looking at i5 parts vs. the AMD part?
According to Tom's CPU charts, the i5's may very slightly outdo the 955 but cost 25-30% more.

Because i5 750 is as good as i7 in gaming. 750 offer you much better performance on almost all situations costing only a few dollars more.
In addition, the 750 has much larger room in OCing and so it will beat 955 up to the pulp any day especially when OCed, so it would last much
longer than 955 and won't make you have to spend more money on upgrade very soon.

955@3.2GHz < 750@2.66GHz!
Imagine what would happen between 955@4GHz and 750@4GHz.