Intel Core i7-3770K Ivy Bridge GPU Performance

[OneCool]

Originally Posted by Aleksander Dishnica

It is not my fault though.
Both benches say Dragon Age II 1280x800 2xAA.
Why don't you state them on the bench?
writting more makes it harder to understand

Not to mention this was a IvyBridge setup too.

W1z 7970 review was done with this

CPU: Intel Core i7 920 @ 3.8 GHz
(Bloomfield, 8192 KB Cache)
Motherboard: Gigabyte X58 Extreme
Intel X58 & ICH10R
Memory: 3x 2048 MB Mushkin Redline XP3-12800 DDR3
@ 1520 MHz 8-7-7-16
Harddisk: WD Caviar Blue WD5000AAKS 500 GB
Power Supply: Antec HCP-1200 1200W
Software: Windows 7 64-bit Service Pack 1
Drivers: NVIDIA: 285.62
ATI: Catalyst 11.12
HD 7970: Dec 16 Reviewer Driver
Display: LG Flatron W3000H 30" 2560x1600

[TheLaughingMan]

To the guy asking about why the 6670 is coming out ahead of a much more powerful card. If the whole "its at low" does help. It means the 6670 has enough power to max out the game at those settings as well as the 7970 lightening. The variance in performance can be attributed to:

miscellaneous variations in the test run
better driver optimization at settings for lower end cards
Windows 7 background tasks running that you cannot control
Pudding

Any of these could have created the very pointless and very minor anomaly. It is the nature of testing systems with hundreds of contributing variables. This is why W1zz and everyone does multiple tests across different games and why W1zz normalizes the results for the overall score.

[eidairaman1]

Originally Posted by hardcore_gamer

I'm not impressed. I expected much higher performance since the GPU takes 1/3 of the total die area.

this shouldnt come to you at any surprise.

Your machine whoops IVB in Graphics capability despite the CPU not being as fast as a SB/IVB...

Originally Posted by Dent1

Looks like the old school AMD APU is spanking Ivy Bridge. Trinity should be exciting.

No Kidding, now if AMD could get the performance numbers up to better spec than SB/IVB/E Models with the GP portion being superior it be a winner

Originally Posted by NdMk2o1o

Meh, no ones buying IB 3770K for igp, when do we get CPU review?

and with a test with and without the IHS in temperature testing

With IHS (not uncapped)
With IHS reinstalled using aftermarket Thermal compound/TIM
With IHS removed and cpu shimmed

see what thermals are and if that is infact true as to why the IVB runs hotter...

[Sah7d]

Nice review.

Many of the reviews do not have test about temperatures why ?
It is not important ?

Some kind of reference would be nice to compare Sandy Vs Ivy in that area
becouse if you want to do OC this is an important fact.

Also there are reports about that the Ivy is hotter becouse of some thermal material
inside the die wich has been replaced to low the cost had an impact in the temperatures.

Anyway if you are already have Sandy Bridge there is no point to upgrade becouse
of the "tic toc" known cpu process with Intel.

[evulmunk33]

nice review!!!

i really would have liked to see a highend card in that comparison... just for fun
just how slow is IVB compared to a 680 or 7970?

btw, did you observe any glitches? my laptops intel igp drivers are terrible, even in 2d... colors look washed out, no option to adjust it, and they seem to dynamically adjust the contrast based on what the display is showing, and i cant turn it off...

so whenever i can, i switch to the nvidia optimus graphics... even if all i do is browse the web and watch videos or read text...

[Frick]

Originally Posted by Aleksander Dishnica

It is not my fault though.
Both benches say Dragon Age II 1280x800 2xAA.
Why don't you state them on the bench?
writting more makes it harder to understand

I actually agree with this. It does makes it easier even if you're reading the review anally.

[Absolution]

Nice and balanced review, unlike the one at THG, where Intel cpu's must be at the top of each chart no matter the configuration.

[phanbuey]

Originally Posted by Aleksander Dishnica

It is not my fault though.
Both benches say Dragon Age II 1280x800 2xAA.
Why don't you state them on the bench?
writting more makes it harder to understand

I must agree with him... "set to low quality" by itself (out of context) means nothing... his post brings up legitimate questions a new reader would have: what was the other bench set at? why are these set to low and the others to something else? The initial, and natural, assumption is that all benches are normalized.

This is very good constructive criticism, best course of action is to improve chart by putting settings of game clearly at the top - or to not, and use same settings for all benches. This would make everything apples to apples (because bananas are pure, misleading evil).

I know everyone will disagree with myself and the two other posterss for the sake of TPU, but as an unbiased 3rd party - those charts are misleading and are easily taken out of context. Not to mention, they are very easily fixed. There is no reason not to improve them.

[entropy13]

Originally Posted by phanbuey

This is very good constructive criticism, best course of action is to improve chart by putting settings of game clearly at the top

None of the major review websites that feature at least 15 benchmarks puts settings at the top of each benchmark's page...because TPU is the only review website that uses that many benchmarks. In the "battle" of making each benchmark's pages even longer v. making just one page that would already specify the settings for everything, the latter won.

Originally Posted by phanbuey

or to not, and use same settings for all benches. This would make everything apples to apples (because bananas are pure, misleading evil).

I agree...if I want to see 0 fps graphs.

Originally Posted by phanbuey

I know everyone will disagree with myself and the two other posterss for the sake of TPU, but as an unbiased 3rd party - those charts are misleading and are easily taken out of context. Not to mention, they are very easily fixed. There is no reason not to improve them.

There's nothing misleading in those charts, it only becomes misleading if you don't read properly, and/or skip pages. Similar to how Tech Report's other graphs (frame times, frame latencies, time spent beyond 50ms) would become misleading if you don't read a certain page ("Our Testing Methods") that precedes those graphs.

[Steevo]

Great review, and we apologize for the village idiot, he must have snuck in here, and he can't read.

Originally Posted by Sah7d

Nice review.

Many of the reviews do not have test about temperatures why ?
It is not important ?

Some kind of reference would be nice to compare Sandy Vs Ivy in that area
becouse if you want to do OC this is an important fact.

Also there are reports about that the Ivy is hotter becouse of some thermal material
inside the die wich has been replaced to low the cost had an impact in the temperatures.

Anyway if you are already have Sandy Bridge there is no point to upgrade becouse
of the "tic toc" known cpu process with Intel.

If you can afford this chip and are pushing it to the max then a water, or high end air cooler isn't an issue for you. Or you also plan on phase or other silly tests using LN or the like and will probably remove the IHS anyway.