Marest's Sample Builds

[Ron Burgundy]

Okay.

http://www.anandtech.com/bench/Product/618?vs=555

Why?

He has the GTX 600s dominating in almost all the other builds. Frankly, however, the 7970 GHz Lightning is pretty damn good. Not to mention is has more VRAM natively and a bigger membus, takes less power both at load and idle... Also it supports 6 monitors.

I think what Marest really wants from you, is to provide cold hard evidence and reasoning to switch them.

ur comparing two reference cards. the cards im comparing are not reference.

[BicycleMafioso]

it's actually rather simple, if you plan on doing a multi card setup, get the 680, if not, get the 7970

I agree with this. Actually, it's my entire stance on nVidia vs AMD in general.
AMD gives you a lot of performance, and for a great price. They overclock well, but consume a lot of energy. They "scale well" in crossfire, but the microstutter offsets that worse.
nVidia works really great in multi-card setups, and well in single-card as well (of course). But. Price for money, they are except at the GTX 660 at pretty poor spots, really compared to the opposition.
I wouldn't recommend a GTX 680 over an HD7970 for a single card config unless you're after something specific that nVidia offers.

And the strength of the HD7970 is not the memory amount but the bandwidth. While it's wrong to say it "gets better" the more textures and resolutions you throw at it, it suffers less.

[Deleted]

it's actually rather simple, if you plan on doing a multi card setup, get the 680, if not, get the 7970

This.

I won't replace the 7970 in that build. I might swap out the MSI iteration of the card, but it will stay a 7970.

[Ron Burgundy]

thats the first time ive heard ppl say single go amd, multi go nvidia.

why?

[Drunkenvalley]

I really ought to do something stupid one of these days. Like building a 2011 testbench and comparing the two families now, more than 6 months after the initial release. (This is quite important because the reviews may be outright outdated, even if they were correct at the time.)

[DeltrusDisc]

ur comparing two reference cards. the cards im comparing are not reference.

Oh, sorry. If you can find me a website that has direct number comparisons between the MSI Lightning GTX 680 and MSI Lightning HD 7970, be my guest. I used a site that I know, love, and trust, however, that happens to have a plethora of tests that have been run, as well, comparing many things.

---------- Post added 2013-01-10 at 11:43 PM ----------

I really ought to do something stupid one of these days. Like building a 2011 testbench and comparing the two families now, more than 6 months after the initial release. (This is quite important because the reviews may be outright outdated, even if they were correct at the time.)

I would share the crap out of your reviews, honestly. That is what bugs me about a lot of the reviewers, they get engineering samples (90%+- of the time) and only really ever test them right as they launch... not later on with new drivers or anything.

[Cyanotical]

can u explain why? i dont think the extra 1gb on the 7970 really does much

SLI works much better than crossfire, this is a proven fact, not taking a fanboy stance on this, but crossfire has loads of issues, and often new games dont see fixes for months, crossfire does really well on benchmarks though, 4 7970s will produce higher results than 4 680s or 2 690s, but almost every reviewer will tell you that its not something you want to use every day, while quad SLI has almost no issues running games

going back to a single card, the higher vram and memory bandwitdh on a 7970 give it some great attributes, but the lower overall price means you can improve your computer elsewhere, such as a larger SSD, faster ram, or even a good soundcard

[Drunkenvalley]

So here's a lil' thing I did. I looked at the techpowerup reviews. The relative performance ones. For these values I looked at the 2560x1600. I took the % of the benchmarked item (100%) and divided by the unidentified GTX 680 in the list. This should produce a number describing the relative performance increase over the GTX 680 on the benchmark.

TL;DR - 100% / GTX 680 * 100 = x%

@ 2560x1600
HD 7970: 105.26%
GTX 680: 106.38%

I'll be the first to admit that I'm not fully sure if I'm correct in my method, so any corrections would be just dandy.

[Ron Burgundy]

and at 1920x1080 ?

[DeltrusDisc]

Did you mean 2560x1440 or 2560x1600? O.o

---------- Post added 2013-01-11 at 12:00 AM ----------

and at 1920x1080 ?

You're the one who linked the reviews in the first place, can't you figure it out yourself?

[Ron Burgundy]

Did you mean 2560x1440 or 2560x1600? O.o

---------- Post added 2013-01-11 at 12:00 AM ----------



You're the one who linked the reviews in the first place, can't you figure it out yourself?

im not very good at math

[Drunkenvalley]

and at 1920x1080 ?

What, do you need me to hold your hand? I'm pretty sure I just described what I did in pretty much perfect detail. I am making your arguments for you here, so stop being lazy.

Did you mean 2560x1440 or 2560x1600? O.o

Err, 2560x1600. I'm not sure how I derped that bad. :P

[DeltrusDisc]

im not very good at math

Then how can you argue about the 7970 being worse if you aren't good at math? It takes some math skills to figure it out.

[Drunkenvalley]

@ 1680x1050
HD 7970: 98.04%
GTX 680: 105.25%

@ 1920x1200
HD 7970: 100%
GTX 680: 105.25%

@ 2560x1600
HD 7970: 105.26%
GTX 680: 106.38%

Conclusion: We have learnt that at the time of the review, the HD 7970 did not perform as well as the GTX 680. And I'm sorry, but that's really all we've learnt. It's a completely meaningless basis imo for actually comparing the cards. Especially since this doesn't describe where and how the card was better, only that the rounded together number indicated the GTX 680 performed better.

[Ron Burgundy]

Then how can you argue about the 7970 being worse if you aren't good at math? It takes some math skills to figure it out.

i looked at the graphs and the gtx 680 was better ESPECIALLY at the 1920x1080 resolution which MOST of us play at.


u guys r trying to gang up on me...

[Drunkenvalley]

u guys r trying to gang up on me...

Yes, mostly because your entire argument is based on extremely lacking information. Hell, I had to provide an argument for you.

[Ron Burgundy]

@ 1680x1050
HD 7970: 98.04%
GTX 680: 105.25%

@ 1920x1200
HD 7970: 100%
GTX 680: 105.25%

@ 2560x1600
HD 7970: 105.26%
GTX 680: 106.38%

Conclusion: We have learnt that at the time of the review, the HD 7970 did not perform as well as the GTX 680. And I'm sorry, but that's really all we've learnt. It's a completely meaningless basis imo for actually comparing the cards. Especially since this doesn't describe where and how the card was better, only that the rounded together number indicated the GTX 680 performed better.

thnks for making my case.

long story short folks msi gtx 680 > msi hd 7970

[Butler to Baby Sloths]

thnks for making my case.

long story short folks msi gtx 680 > msi hd 7970

At the time of writing.

[DeltrusDisc]

thnks for making my case.

long story short folks msi gtx 680 > msi hd 7970

Read his conclusion again... good lord.

[Ron Burgundy]

Yes, mostly because your entire argument is based on extremely lacking information. Hell, I had to provide an argument for you.

the numbers dont lie