MSI Radeon HD 7970 Lightning 3 GB

[W1zzard]

MSI's Radeon HD 7970 Lightning is a heavily customized implementation of the HD 7970 with focus on performance and overclocking. The card uses a beefed up voltage regulation circuitry and massive dual-fan heatsink to ensure stable operation at MSI's clocks of 1070 MHz core and 1400 MHz memory.

Show full review

 

[swirl09]

All going well a new build will be happening next week. Over the past month Ive gone back and forth from the red team to the green team a couple of times, well the wheel is about to stop rotating and we're currently sitting on red ^.^

Its not going to be this card however, the lack of dual link DVI is not acceptable for me. Altho on a positive note for this card I would like to point out that while the price is listed as a con here, its actually readily available for just under the price of most reference 680's (UK&IE), which just a month ago was a cornerstone of the praise nVidia were getting.

 

[FreedomEclipse]

 

[hhumas]

in comparison where is evga gtx 680 sc signature

 

[dj-electric]

W1zz, where is overvolting?

 

[Completely Bonkers]

single link DVI! :shadedshu

W1zz... could you please add TDP and technology nm as a standard in all your spec tables? (http://www.techpowerup.com/reviews/MSI/HD_7970_Lightning/) Thanks!

 

[the54thvoid]

Nice. The 7970 at 1070 performs somewhere between the 680's operating range (1006-1110).

I don't like the perf summary graphs as they (by definition) have to iron out the quite massive diffs between cards. I took the liberty to take W1zz's review and plot a spreadsheet using the 1200 res results on all games tested.

It shows MASSIVE swings between the two cards.
Red is an AMD win, Green is an Nvidia win. Anything under 5% isn't coloured - it's too low to be significant. Note the whopping 20% higher average power draw too. But then again, in four titles it beats the 680 by 20%

This chart is nothing but a mathematical model of W1zz's review. Hope it helps. (The minus simply means it went Nvidia's way).

One caveat - I'm sure the fan profile could be altered to make it quieter but also, Hardware Canucks tested the 7970 Lightning and got this result (they confirmed W1zz's power draw results).

 

[W1zzard]

I'm sure the fan profile could be altered to make it quieter

of course. but then temps would be higher.

does hwcanucks test card only noise or whole system? what meter do they use? 12 inch seems a bit close

 

[the54thvoid]

From HC

What you see below are the baseline idle dB(A) results attained for a relatively quiet open-case system (specs are in the Methodology section) sans GPU along with the attained results for each individual card in idle and load scenarios. The meter we use has been calibrated and is placed at seated ear-level exactly 12” away from the GPU’s fan. For the load scenarios, a loop of Unigine Heave 2.5 is used in order to generate a constant load on the GPU(s) over the course of 20 minutes.

Methodology section.
http://www.hardwarecanucks.com/foru...53263-msi-hd-7970-lightning-3gb-review-4.html

Guru3D gets this: http://www.guru3d.com/article/msi-radeon-hd-7970-lightning-review/12

 

[W1zzard]

"seated ear level".. maybe they use a highly directional microphone (like most sound level meters) and don't point it towards the graphics card.

even subjectively the 7970 lightning is extremely noisy

guru claims the card does not get much noisier going from idle to load (36 vs 37). you want me to make a recording how that sounds here?

 

[the54thvoid]

I never find recordings useful. I never know what volume to use

It's academic to me, I'm under water now with my card - which is a point, EKWB make blocks for Asus DCII cards. Someone should make blocks for Lightnings.

Tweaktown tested the Lightning, albeit at 1225 core and it also noisy as hell. I guess there's many factors concerned and the luck of the draw how it will work out in each persons rig.

My GTX 580 Lightning hit the sound barrier and ran at about 50-60 degrees load. I tweaked it on a fan profile to hit closer to 70 degrees but it kept it very quiet.

 

[BigMack70]

Nice review to have.

I do think it might have been more helpful if it included the Lightning's overclocked results in all the tests. The entire purpose of the card is overclocking (as is even said in the review), so I think it's a disservice to not include those tests.

 

[Lionheart]

Hey Wizz, will you be showing more games in the Overclocked page other then COD4 in the near future?

 

[W1zzard]

Hey Wizz, will you be showing more games in the Overclocked page other then COD4 in the near future?

no plans for more than one game.

this review adds eyefinity testing, temperature and oc comparison tables. last review added a new ivy bridge test system and alan wake. nobody had anything to say about that. not very motivated right now to add extra work

 

[Delta6326]

Awesome review! Love the new setup W1zz just when I thought your reviews couldn't get any more pro you come out with this. I really like the Temp page.

I agree with you W1zz that the GTX 680 is the better buy.

 

[douglatins]

Weird, the lightning 580 is perfect, low noise and low temps, this have neither WEIRD

 

[the54thvoid]

no plans for more than one game.

this review adds eyefinity testing, temperature and oc comparison tables. last review added a new ivy bridge test system and alan wake. nobody had anything to say about that. not very motivated right now to add extra work

The work you do is second to none for in depth analysis. Your games spread is top notch and the review is clinical. I always place your reviews at the top of the quality pile. I think many take your work a little for granted but please understand your work is easily THE reference point for other reviews.

 

[dj-electric]

I concur w1zzard's opinion, the sharp shape of TF II-IV's fans does make A LOT of noise. A lot of the times an unnecessary noise even at low temperatures. MSI really needs to fix the fan\temp ratio operation of their TwinFrozrs in a lot of graphics cards such as the lightning

 

[Darkleoco]

Very nice looking card with very nice performance in BF3 O_O.

 

[cadaveca]

Damn you MSI for using black PCB and gold chokes. Can haz three, please?

 

[Delta6326]

no plans for more than one game.

this review adds eyefinity testing, temperature and oc comparison tables. last review added a new ivy bridge test system and alan wake. nobody had anything to say about that. not very motivated right now to add extra work

If it makes you fill any better I ONLY use TPU for my tech info and I do notice the big and small changes you make with your reviews.

I'm good with hardware, but suck with software I only use the internet for TPU, Newegg, Gmail.

 

[BigMack70]

As to fan noise... I own one of these cards and it is pretty quiet at 1.22v 1200/1600 OC. It's significantly quieter than my XFX DD Black Edition 7970 was at that clock, though it is not as quiet as my previous card (8800 GTS) was.

Granted, I have very good case ventillation (2x120mm front + side intake, 3x140mm top/rear exhaust), but still... I dunno about the noise results here, especially since noise isn't a big complaint in other reviews of this card.

I'm getting good temps (sub-75C) with only marginal fan noise, at a pretty high overclock. Now, I don't use OCCT/Furmark, so if that's the test here for noise then I could see things being different. However, for hours of gaming, I have no complaints with temps/noise.

Wish that the one OC test for this card was something where the 680/7970 perform more similarly in... the 7970 does really poorly in some old DX9 games like COD. It would be like testing a 680 OC on AVP... doesn't give a good indication of the average performance boost compared to the other card. I think 54thvoid's point above is very important to note when comparing the 7970 and the 680.

 

[W1zzard]

i'm open to suggestions for a test that checks oc performance

 

[Casecutter]

Sure we knew this would promote a huge TDP once OC'd and max'd that not a true bellwether, when a GTX680 no longer permit you the traditional OC gains like this. In all fairness, today it’s really moot. Looking at “Peak” (Crysis 2 at 1920x1200), we get a better representation of a typical gaming power draw and there’s no real difference. (ref to ref)

I'd now like to see the power figures for each game during real play. I think looking at Nvidia GTX680 (or AMD) those Max power number (Furmark) matters little anymore. I'd like to see how [H] has been reporting it, they pull power on all B-M runs (they do 5 titles). There some clear advantages for the GTX680 (BF3), but once averaged the GTX680 is 4-5% more efficient for what not any big performance difference in what you see coming through the monitor for those five.

Nvidia got it right this time a small fast Chip, but with complexity and cost, while negating the long-established enthusiast OC’n. They made their card "plug-n-play"... while idot proof. If you do bump it up, it still ends up maintianing OC'n limits based on dynamically adjusted clock/voltage against the rendering load, temperature, and other factors.

 

[FreedomEclipse]

I would be nice if you included tests with the GPU reactor on and off as I think the GPU reactor will be sold seperately.

there are some sites that are saying the reactor is more of a gimmick and doesnt really do much for overclocking at all but just makes the card look more sexy