To Anandtech and company

[SlickR12345]

Hi guys, I like the website and have been a reader for over 10 years, but I've always wanted to see some real life tests done, except for the best case scenario and no website has still done it.

What I mean is I'd love to see tests done with lower end parts and configurations that aren't optimal, you know like 90% of the people have.

Like I would really love to see a Intel I5 2300 with GTX 680 and 2GB ram tested, or AMD FX6300 with AMD 5570 and 8GB ram.

 

[Puppies04]

Like I would really love to see a Intel I5 2300 with GTX 680 and 2GB ram tested, or AMD FX6300 with AMD 5570 and 8GB ram.

Who in their right mind is going to blow a ton of money on a top of the line GPU whilst spending less than half that amount on a bottom end I5.

That isn't real world it is stupid.

 

[nemesismk2]

I would also like to see how a decent ati or nvidia video card performs with a cpu which isn't the best from amd or intel.

 

[AustinInDallas]

when you benchmark, you want to take out as many variable as possible.

ceteris paribus!

You dont want anything else being a bottleneck, and show the absolute best case scenario for each part

 

[biostud]

here you go
http://www.techspot.com/review/642-crysis-3-performance/page6.html

 

[jaydee]

when you benchmark, you want to take out as many variable as possible.

ceteris paribus!

You dont want anything else being a bottleneck, and show the absolute best case scenario for each part

Consumer-centric review != Hardware-centric review

 

[SlowSpyder]

I would also like to see how a decent ati or nvidia video card performs with a cpu which isn't the best from amd or intel.

I can tell you what its like... its not that bad! I'd always load up heavier on GPU power than CPU power. I'm sure some will disagree with that, but for that combo works fine. Obviously a fast CPU/GPU combo is best. But, at least for me, I feel like I get more life out of a system by upgrading the video card more so than the CPU, typically these days. Though I think I've pushed it about as far as I want and a new platform will be incoming sooner than later.

 

[AustinInDallas]

Consumer-centric review != Hardware-centric review

The reviews are not written for the general public. They are written for their readers. I would say the large majority of their reads have an i5, i7 or at the least a quad core amd.

 

[Vesku]

Agreed to some extent, surprised there are no real comprehensive reviews of what are probably the most purchased CPUs. Not much out there on Pentiums and Core i3s.

 

[Maximilian]

Seems like a waste of time and something you can already figure out anyways.

- Look at CPU benchmarks for gaming, note the FPS the CPU in question gets

- Look at GPU benchmarks for gaming, note the FPS the GPU in question gets

The smallest number is the FPS you will likely be seeing assuming settings/res are equal.

 

[Exophase]

A review site can't handle every possible test combination.. each additional one costs time and money.

What most sites do is try to test CPUs under conditions where they're not GPU limited and GPUs under conditions where they aren't CPU limited. You can then pretty much find where the two intersect to see if a given combination fits for you.

 

[Arkadrel]

here you go
http://www.techspot.com/review/642-crysis-3-performance/page6.html

FX 4170 = ~120$ = 56 fps
i5 3470 = ~199$ = 60 fps
FX 8350 = ~199$ = 61 fps

i7 3770k = ~330$ = 64 fps
i7 3960x = ~1070$ = 67 fps

AMD is good value for Crysis 3.



same for Battlefield 3:

 

[lakedude]

FX 4170 = ~120$ = 56 fps
i5 3470 = ~199$ = 60 fps
FX 8350 = ~199$ = 61 fps

i7 3770k = ~330$ = 64 fps
i7 3960x = ~1070$ = 67 fps

AMD is good value for Crysis 3...

AMD is an excellent value for a lot of things against the i7-3770k, if you don't care about power usage.

AMD's value does not hold up nearly as well against the i5-3570K @ $219.99...

 

[Exophase]

It always bugs me when people look at heavily GPU limited scenarios and show that CPU A is ever so slightly faster than CPU B, when in CPU limited scenarios (much lower resolution for instance) the situation is drastically reversed. Rather than looking at who gets 61fps vs 60fps averaged over a long period of time, most of which is CPU limited, you should look at frame time histograms and CPU utilization (if you care about power consumption).

 

[exar333]

AMD is an excellent value for a lot of things against the i7-3770k, if you don't care about power usage.

AMD's value does not hold up nearly as well against the i5-3570K @ $219.99...

I think the 3570k was there and he erased it, but forgot to get rid of the space. :biggrin:

 

[mikk]

same for Battlefield 3:

GPU bottleneck is new to you.

 

[Abwx]

Theses times Anand seems more interested by DT cases reviews....

 

[sm625]

I definitely agree with the OP. Say you have $400 to spend on a cpu, GPU, motherboard, RAM, case, and power supply. (Leave drives and OS out of it for now.) That's it. Now build a machine that yields the fastest framerates in a selection of games. There is almost zero information on what these components might be. Sure I could make a few guesses (pentium, A10, FX-6XXX or i3?), but who really knows?

Is an i5 paired with a 7770 faster than an i3 paired with a 7850? That seems to be the biggest most relevant question. The rather snobby answer seems to be "buy a 7850... and a 3770k!"

 

[aigomorla]

when you benchmark, you want to take out as many variable as possible.

ceteris paribus!

You dont want anything else being a bottleneck, and show the absolute best case scenario for each part

+1

this basically answers everything.

Why would u want to set yourself for other bottlenecks when your running testing?
How would that help you or anyone else with any tests when your bottlenecked on something else outside the video card.


WOW.. it amazes me at how much this hobby has been stupified.
No.. you cant really test low end eq.. with low end hardware because of said bottle necks.
Also no one would want to invest in the low end hardware with low end crap.. because if you collected all that... it would come out to be more then the high end stuff with less then 1/2 the performance numbers.

Also most of these tests are for games.
You dont game on a low end crap gpu unless its an indie / flash game.
And you most definitely dont Distribute Computing on a crap end GPU.

So the whole point of the low end gpu was to get office production done... while drawing a tiny bit of power.. while having enough prowess to run flash / movies...

NOT PLAY GAMES or subject it to torcher test runs doing F@H.

 

[ElFenix]

GPU bottleneck is new to you.

if a top of the line GPU is bottlenecking at a very common resolution, then CPUs aren't the thing to worry about. that's exactly the point.

 

[SlowSpyder]

if a top of the line GPU is bottlenecking at a very common resolution, then CPUs aren't the thing to worry about. that's exactly the point.

I'm more like this, too. I get why low res benches are important, they show what the CPU is capable of, which is handy when benching different CPU's. But I also like, and put more value on as far as what to expect in my real world use, is common higher resolution benches, like 1080P, since I game at 19x12.

For what I do, I'll take my Thuban and 7970 over a 5GHz Ivy with a 7750 level card.

 

[frozentundra123456]

GPU limited benchmarks tell you only one thing.

They tell you in that particular game, at the settings tested, all the cpus that give the same framerate are "good enough" to max out that particular gpu. So from that point of view, they tell you something. Just like a prius and a corvette with 100 mph tires would have the same top speed. It tells you something besides the engine (cpu) is limiting performance. But it does not tell you which has the fastest engine. Give them better tires and the corvette would obviously win. (edit: even in the graph you show, the athlon X2 dual core is as fast as the bulldozer and ivy i5. Are you saying it is as good a gaming cpu as either of those?)

That is the problem with GPU limited benchmarks. They dont really tell you which cpu is faster in a wide variety of games, or even in that game if you used a stronger video card. To get a true indication of a cpus value in gaming, you have to test a wide variety of games at a variety of settings. To try to evaluate a cpu in one game only, especially in a gpu limited benchmark, is misleading at best.

 

[frozentundra123456]

I'm more like this, too. I get why low res benches are important, they show what the CPU is capable of, which is handy when benching different CPU's. But I also like, and put more value on as far as what to expect in my real world use, is common higher resolution benches, like 1080P, since I game at 19x12.

For what I do, I'll take my Thuban and 7970 over a 5GHz Ivy with a 7750 level card.

That is like saying you would take our vikings quarterback Christian Ponder at the helm of an NFL team vs Tom Brady quarterbacking a high school team. Which team do you think would win??

Does that make Ponder a better quarterback?? we wish, but that hardly makes it the case.

It is a totally unfair comparison.

 

[ElFenix]

That is like saying you would take our vikings quarterback Christian Ponder at the helm of an NFL team vs Tom Brady quarterbacking a high school team. Which team do you think would win??

Does that make Ponder a better quarterback?? we wish, but that hardly makes it the case.

It is a totally unfair comparison.

not if you already have thuban (or a core2, for that matter) and, say, a 4870, and want to know which way to go with an upgrade (though you'd likely keep your 4870).

 

[aigomorla]

not if you already have thuban (or a core2, for that matter) and, say, a 4870, and want to know which way to go with an upgrade (though you'd likely keep your 4870).

but people arent going to be running high ghz overclocked cpu's, or monster cpu's with budget cards...

that just doesnt make sense...

the op is asking for tests with non optimal parts.
So basically he wants to strap an atom processor to a 640GT class and bench it.

We're all going :hmm:
Can the CPU even handle the benching software to begin with?