ASUS Radeon HD 7970 CrossFire

[W1zzard]

AMD's Radeon HD 7970 has quickly established itself as the fastest single GPU graphics card in the world. In our ASUS HD 7970 CrossFire review we take two of these beasts for a spin to see what benefit they offer. One of the highlights is reduced power consumption thanks to AMD's new ZeroCore power technology.

Show full review

 

[alexsubri]

Thanks as always! The max power consumption is amazing! Compared to the 6970 Crossfire! (525watt vs. almost 600 watt)

BTW, first

 

[v12dock]

Need drivers otherwise is very fast

 

[entropy13]

W1zz teasing readers with a multi-monitor setup in the future.

 

[alexsubri]

This is obviously too expensive for me, however I'd be good with 2 7950's (Coming in Feb) , hoping that they will be less than $400 or at least around the same price of the 5850s when it came out.

The 7990 (coming in march--will be two 7970s in xfire) is going to be wicked sick. I guarantee you that it will be $800-$1200 price range ($1200 being like special editions or limited editions)

 

[dj-electric]

1 - drivers, derivers drivers
2 - worse scaling then HD6000 at 1920\2560 = 7970 needs to stretch it's legs, its not really worse

 

[Lionheart]

Oh how I want, improved drivers will fully unleash these beasts ^_^

 

[alexsubri]

1 - drivers, derivers drivers
2 - worse scaling then HD6000 at 1920\2560 = 7970 needs to stretch it's legs, its not really worse

All ATI drivers are shit once a new product is released. It will take time, just like any other GPU to work its kinks out. The scaling is due to the drivers. But, you can't complaing playing Crysis 2 maxed out Ultra Package @ 2560 x 1600 with 70 FPS

 

[buggalugs]

Yikes, still using bad drivers.

I'm pretty sure AMD will release WHQL in the next few hours.

 

[theJesus]

1 - drivers, derivers drivers
2 - worse scaling then HD6000 at 1920\2560 = 7970 needs to stretch it's legs, its not really worse

I don't think the scaling is worse; it's wholly dependent upon the drivers. If you notice, where it does work, scales almost 100% in most situations.

 

[hardcore_gamer]

"Remember, this money can buy you an Xbox 360, plus Kinect, plus 46" TV......"


Who the 4^cK buys Xbox 360 in 2012 ?

 

[Trackr]

No offense, W1z, but I can't trust a review with a 920 anymore.

I went from 65 FPS to 75 FPS in Crysis @ 1080p going from a 920 to a 2500k.

I suppose it wasn't until I updated myself that I saw the difference.

 

[Jack Doph]

With this generation of cards AMD introduced their new ZeroCore power saving technology which really can shine in CrossFire. In a CrossFire setup only a single card will be active when not gaming; the other cards will be essentially switched off, consuming only 1W of power and producing no heat or noise. Once a gaming session starts, the additional card(s) will instantly wake up and be ready to provide their full power without any manual switching or software configuration required. It seems AMD has found the holy grail for multi-GPU power consumption reduction.

I'm not sure I fully understand this, as I'm running 5850s in CrossFire and they behave in the exact same way, leading me to think it's driver-related, rather than 'technology' related as such.
Am I missing something?

 

[W1zzard]

I'm not sure I fully understand this, as I'm running 5850s in CrossFire and they behave in the exact same way, leading me to think it's driver-related, rather than 'technology' related as such.
Am I missing something?

your cards go into a lower power state where they are all off including the fan ?

No offense, W1z, but I can't trust a review with a 920 anymore.

I went from 65 FPS to 75 FPS in Crysis @ 1080p going from a 920 to a 2500k.

go overclock your 920 and it'll be faster than the stock 2500k in many benches (happens here)

my 920 does 2x pcie x16 for this review, yours can't

 

[Jack Doph]

your cards go into a lower power state where they are all off including the fan ?

Yes, exactly as you describe in your review is how mine behave as well.

 

[pantherx12]

Mmmm I see 100% scaling in places

 

[buggalugs]

Whats the noise like running 2 of those bad boys?

 

[W1zzard]

Yes, exactly as you describe in your review is how mine behave as well.

hmm amd claims this is a new feature.

maybe your card behaves the same but consumes normal idle power and not 1W like in zerocore

 

[W1zzard]

Whats the noise like running 2 of those bad boys?

in 2D, like a single card due to ZeroCore power.

in 3D, twice that of a single card. almost, since the second card might be loaded less. maybe higher because of worse airflow to one card. single card is already quite noisy in 3d

 

[Jack Doph]

hmm amd claims this is a new feature.

maybe your card behaves the same but consumes normal idle power and not 1W like in zerocore

Well, the only difference I find is that the fan on card 2 does not turn off (ever), but this may be due to my ambient temps never going below 25C (and is currently at 35C - we had 40C+ today btw).
Everything else is at 0%
Power usage measured is less than 1w (which I have to presume as being the fan drawing this power).
:/

Any insights?

 

[W1zzard]

how did you measure 1W ?

 

[Jack Doph]

how did you measure 1W ?

CCC & GPU-Z XD
However, if I launch more than one tab in my browser, or a second GPU-Z, this goes up a tad.
The annoying thing is that whenever I try do do a screen-grab, card #2 kicks-in :/
Without any useful independent hardware tools, how else can I measure this accurately?
CCC only goes so far and GPU-Z seems to add more to GFX stress in that it forces the second card to kick-in (even if only at the lowest MHz - 157core/300MHz mem).

Is there something more reliable I can rely on?

 

[blibba]

in 2D, like a single card due to ZeroCore power.

in 3D, twice that of a single card. almost, since the second card might be loaded less. maybe higher because of worse airflow to one card. single card is already quite noisy in 3d

Two sources of an amount of noise do not make twice that amount of noise.

 

[W1zzard]

Two sources of an amount of noise do not make twice that amount of noise.

how would you explain it to him then without giving any specific numbers and confusing him even more?

CCC & GPU-Z XD
However, if I launch more than one tab in my browser, or a second GPU-Z, this goes up a tad.
The annoying thing is that whenever I try do do a screen-grab, card #2 kicks-in :/
Without any useful independent hardware tools, how else can I measure this accurately?
CCC only goes so far and GPU-Z seems to add more to GFX stress in that it forces the second card to kick-in (even if only at the lowest MHz - 157core/300MHz mem).

Is there something more reliable I can rely on?

software only shows measurements for the gpu (excluding everything else on the card, like memory).

you can't measure card only power consumption without complicated hardware tools. thats why only few reviewers test it

i'm quite sure that your second card will consume its idle power in 2D, which is probably around 25 W for 5850

 

[Frizz]

I just sold my rig to enable myself to buy a new car.. now I wanna sell my car to enable myself to buy three of these babies . Zero core feature definitely looks like a winner being a Crossfire user for over 2 years this looks to be very beneficial in terms of $$$ and noise.