My benches so far. Adding GTX480 SLI benches.

[Janooo]

Some possible reasons have already been discussed in this thread, but unfortunately Keys isn't in a position to go through all the permutations to test out some of the theories. That would require things he doesn't have like an SSD and an entire new mobo+CPU combination.

AT shows 10% and Keys almost 50% in some cases and you are saying it could be SSD? You are kidding, right?

 

[Dark4ng3l]

Unbiased Keysplayr wrote:
Actually just reading AT,HardOCP and Bit Tech gave me that info.
I don't think I need to see the "revised" Focus group version.
Imagine that...

Touché. All I would like is some O/C numbers (and O/C temps) to compare a MAXED OUT 480 to a MAXED OUT 5870 but that is too much to ask for apparently.

 

[SirPauly]

I don't. When a focus member discusses ATI -- they're not going to be believed by some at face value and will be challenged. No matter how nice, respectful they are, if affiliated with nVidia -- it's a way of voicing what nVidia may do wrong in their eyes at them.

A way of getting away from the ATI vs nVidia paradox syndrome that effects many forums is simply comparing the GTX-480 to the GTX-285 and discuss the advantages of the new architecture to the old one from nVidia. It may not be sexy or as emotionally charged but the flow of data can be presented easier.

Just a thought.

 

[tviceman]

Touché. All I would like is some O/C numbers (and O/C temps) to compare a MAXED OUT 480 to a MAXED OUT 5870 but that is too much to ask for apparently.

Even if he did it, you'd just dispute and argue those results too so it's a lose-lose situation for him no matter what. He loses lots of time finding the max stable o/c for each card and then rerunning benchmarks, and he loses with people like you who, regardless of the results, won't believe key's methodology anyhow.

 

[Dark4ng3l]

Even if he did it, you'd just dispute and argue those results too so it's a lose-lose situation for him no matter what. He loses lots of time finding the max stable o/c for each card and then rerunning benchmarks, and he loses with people like you who, regardless of the results, won't believe key's methodology anyhow.

No because what we really want to know is how well the thing does as a % of the base card. It it overclocks 15% core and 10% memory for 12% more speed then you can apply that to whatever benchmarks you believe. I mean unless he comes up with the magical asus 50% overclocks I can't really dispute it since I have no idea what kind of relative overclocks are possible on these things.

 

[MegaWorks]

I don't. When a focus member discusses ATI -- they're not going to be believed by some at face value and will be challenged. No matter how nice, respectful they are, if affiliated with nVidia -- it's a way of voicing what nVidia may do wrong in their eyes at them.

/\ what he said.

 

[GaiaHunter]

An interesting review I don't think I've seen posted in here. http://www.hardwareheaven.com/reviews.php?reviewid=950&pageid=1

Has OC 480 - 766/1918, shows physX improvements over GTX 285 and has min framerates and graphs.

 

[ronnn]

heaven does very poor reviews and are basically for sale.

 

[happy medium]

An interesting review I don't think I've seen posted in here. http://www.hardwareheaven.com/reviews.php?reviewid=950&pageid=1

Has OC 480 - 766/1918, shows physX improvements over GTX 285 and has min framerates and graphs.

Thats funny I was just reading that review.
I thought it was very good.

Even had some 3d vision stuff.

 

[clok1966]

Well i for one am not bashing Keys, just questioning how 10 reviews from other sites all show an average of 10% and his shows 15%-35%.

Again, I do appreciate his time, but do not agree with his results. No implication that he did anything "shady" but wish I could understand why the results are so much (well 5% at the lowest may not be huge) different then everybody elses.

As I'm in the market for a new video card I like to read "everything" i can. Its frustrating, not becuase I prefer one card over the other, as i have owned every barnd there is for the last 18 years, but as I want the most BANG FOR MY BUCK. I'm honeslty frustrated with the 6 month cycle on cards for the last 4-5 years. As a bit of a PC "nerd" i like my stuff to be somewhat "up to date" but the $500 video card cycle has really soured me (on both sides of the 2 man war it is nowdays) upgrades. So i have regulated myslef to 1 card a year at most. Hence my involvment in this thread. I care about my money and not about the manufacturer. And when i read 10 reviews that say one thing, and (do i call this a review? guess not, its benchs and he did state nothing but his own off the cuff ones)... err let em me rephrase that, I read 10 different sets of benchmarks that all resemble each other then one that does not, cant blame some of us for questioning it.

There is where the question lies, review sites, or average joe users benchs.. to be honest I like average Joe users benchs as he is "me" so to speak, not sitting on a $5000 computer with new stuff weekly, he wants what i want, best bang for the buck. His setup most likley resembles mine (used day to day and not just for "clean" benchs). SO maybe on the average machine(as in used daily, has crap running in background) these benchs are acurate.. again this get back to "why " they are different, if its true maybe Nv has the better card speed wise even more then we know for the average user.

Oh well, I think with all this, and drivers being new (on nv) i may just wait some more and see in a month. Hopefully prices have dropped a littel on the older stuff, or compatition forces some price cuts (ah.. the good old AMD Intel days when they both preformed within 10% of each other and prices dropped!!!, common ati and NV, fight it out! drop those prices!).

 

[Lonyo]

heaven does very poor reviews and are basically for sale.

I love the way that I clicked in the review, and in the background is a giant ad for the Sapphire HD5870 Toxic, behind the GTX480 review.

If it's for sale, I'm wondering which side it's going to be biased for.

 

[nyker96]

I dont know, at these resolutions the 480 is stomping the 5870 according to Keys. In some of these games it is by over 35%.

/shrug

I wonder if it's because of 480's superior tessellation or 5870 is quite memory limited at those high rez situations which give 480 near 35% advantage.

 

[Dark4ng3l]

If you look at the 2GB 5870 review you will see that the extra GB does absolutely nothing at all in 1 screen situations so it can't be memory limited.

 

[Zebo]

techreport has ati's new 5870 in 480 review. Not quite fast enough.... needs 1000mhz and 2gb to trade w/ 480..

Course I think both are silly when you can buy 2 5850's for a little more and be a whole lot faster. (see firingsquad review... quieter, less power and 20-50% faster)

 

[ronnn]

I love the way that I clicked in the review, and in the background is a giant ad for the Sapphire HD5870 Toxic, behind the GTX480 review.

If it's for sale, I'm wondering which side it's going to be biased for.

They vary.

 

[ArchAngel777]

The reality is that even though nVidia was late, at least they produced a faster part. R600 arrived 6 months late and still wasn't fast enough. So, in that regard, nVidia delivered better than ATI did.

However, I believe that the nVidia glory days are over. Where can they go next? Die Shrink? As far as I can tell they don't have any room to move. ATI on the other hand, in theory, could produce a monolithic chip in the next 4-6 months and dominate the high end for quite a while. I just don't see where nVidia can go at this point.

But, for the record, I am very impressed with the minimum frame rates of this GTX 480. If I had plenty of money, I'd pick one up and slap an A/M Watercooled cooler on it and be happy. But as it stands now, my GTX 280 (in RMA, at the moment) is about as loud as I can handle and from the tests, this 480 is quite a bit louder.

Anyway, lets give credit where credit is due.

nVidia has some really awesome minimum frame rates (honestly, I care far, far, far more about that than the average frame rate).
nVidia is faster by at least 15%

But, the down side
nVidia is loud...
nVidia is power hungry...

And if money is no object, where one could simply spent a lot of money on an A/M cooler, then I would glady upgrade to a GTX 480. But, IMO, if a video card doesn't come with the complete package stock, then I am not likely to buy it. Because spending $500 on a card, only to spend another $150 for a block just isn't worth it to me.

So nVidia is on the top again slightly. But I do not believe they are going to hold on to this lead long at all.

 

[NoQuarter]

The reality is that even though nVidia was late, at least they produced a faster part. R600 arrived 6 months late and still wasn't fast enough. So, in that regard, nVidia delivered better than ATI did.

However, I believe that the nVidia glory days are over. Where can they go next? Die Shrink? As far as I can tell they don't have any room to move. ATI on the other hand, in theory, could produce a monolithic chip in the next 4-6 months and dominate the high end for quite a while. I just don't see where nVidia can go at this point.

I disagree with the 'at least it's faster', the nVidia part was only faster because they knew how high they needed to overclock it to beat the 5870. If they released it at a reasonable clock (ie not running super hot and loud, specialty cooling) or not having 6 months to analyze the 5870 it wouldn't have performed well enough to beat it so they threw on a badass cooler and jacked it up.

But I certainly don't think the nVidia glory days are over, the Fermi architecture itself has a lot of promise and it's going to be a very strong base for their next 2 generations of chips.

 

[ArchAngel777]

I disagree with the 'at least it's faster', the nVidia part was only faster because they knew how high they needed to overclock it to beat the 5870. If they released it at a reasonable clock (ie not running super hot and loud, specialty cooling) or not having 6 months to analyze the 5870 it wouldn't have performed well enough to beat it so they threw on a badass cooler and jacked it up.

But I certainly don't think the nVidia glory days are over, the Fermi architecture itself has a lot of promise and it's going to be a very strong base for their next 2 generations of chips.

I agree with you on why it was faster, but ATI had well over 6 months to get R600 faster than G80, and it wasn't... In fact, it was dominated all over. R600 was a complete failure. That ran hot, AND it didn't perform... Go figure. The GTX 480 on the other hand isn't a complate failure, or a failure at all. It just didn't live up to the hype. So, regardless of whether or not nVidia did an impressive thing by releasing a card 6 months late that bested the competition, the point is more that at least they did that, unlike ATI. And hey, this is coming from someone who really doesn't like nVidia. But I own what I deem the best card for the money... I won't let RED Versus GREEN determine my purchase decisions.

 

[Keysplayr]

Well i for one am not bashing Keys, just questioning how 10 reviews from other sites all show an average of 10% and his shows 15%-35%.

Again, I do appreciate his time, but do not agree with his results. No implication that he did anything "shady" but wish I could understand why the results are so much (well 5% at the lowest may not be huge) different then everybody elses.

As I'm in the market for a new video card I like to read "everything" i can. Its frustrating, not becuase I prefer one card over the other, as i have owned every barnd there is for the last 18 years, but as I want the most BANG FOR MY BUCK. I'm honeslty frustrated with the 6 month cycle on cards for the last 4-5 years. As a bit of a PC "nerd" i like my stuff to be somewhat "up to date" but the $500 video card cycle has really soured me (on both sides of the 2 man war it is nowdays) upgrades. So i have regulated myslef to 1 card a year at most. Hence my involvment in this thread. I care about my money and not about the manufacturer. And when i read 10 reviews that say one thing, and (do i call this a review? guess not, its benchs and he did state nothing but his own off the cuff ones)... err let em me rephrase that, I read 10 different sets of benchmarks that all resemble each other then one that does not, cant blame some of us for questioning it.

There is where the question lies, review sites, or average joe users benchs.. to be honest I like average Joe users benchs as he is "me" so to speak, not sitting on a $5000 computer with new stuff weekly, he wants what i want, best bang for the buck. His setup most likley resembles mine (used day to day and not just for "clean" benchs). SO maybe on the average machine(as in used daily, has crap running in background) these benchs are acurate.. again this get back to "why " they are different, if its true maybe Nv has the better card speed wise even more then we know for the average user.

Oh well, I think with all this, and drivers being new (on nv) i may just wait some more and see in a month. Hopefully prices have dropped a littel on the older stuff, or compatition forces some price cuts (ah.. the good old AMD Intel days when they both preformed within 10% of each other and prices dropped!!!, common ati and NV, fight it out! drop those prices!).

You tell me. My system was configured identically for each card. Two separate but identical hard drives (one for each OS containing the Nvidia, or ATI card).
Same exact drivers for mobo, sound, etc. installed on each in the exact same order.
I took pains to make sure the system was configured EXACTLY alike.
Every bench I took, I wrote down every setting on a checklist to make sure the settings were the exact same for each cards bench session.
If you were sitting here next to me while running the benches, you'd be like "I don't think we could have run a more exactly accurate test if we tried.". So, you tell me.

I'm sure AT had a different setting somewhere than I did. I totally maxxed anything and everything at 16x10 and 19x12. If you've noticed, I have listed every option for every game tested at the top of each page. Basic options and advanced options alike if applicable. Not sure if AT did this as well. Usually they just list the most known. As do other sites.

Another thing. Are you factoring that 10% across the board on 10 other sites? Because I only did DX11 stuff.

 

[SirPauly]

It's pretty simple to compare to some degree considering there must be 5870 owners here in this forum. The title that stands out to me is Stalker and if someone has the title and a 5870 bench the title with the same exact stalker settings as Keys and compare.

Sorry for my ignorance, don't have this title but when using the Ultra preset does it automatically set HDAO on Ultra or is this a separate setting?

 

[Keysplayr]

It's pretty simple to compare to some degree considering there must be 5870 owners here in this forum. The title that stands out to me is Stalker and if someone has the title and a 5870 bench the title with the same exact stalker settings as Keys and compare.

Sorry for my ignorance, don't have this title but when using the Ultra preset does it automatically set HDAO on Ultra or is this a separate setting?

I don't have this title either. It is the Stalker: CoP Benchmark.
And yes, you would need to manually set HDAO on Ultra.

 

[SirPauly]

So, it would be easy for many to compare then, thanks!

 

[Janooo]

Keys, you screwed up Stalker. No site shows close to 50% difference as you do, not to mention results going the opposite direction as AT.
Just be a man and admit it.

 

[SirPauly]

Testing and benching is much more productive than finger pointing. Keys stands behind his numbers, imho -- bench with the same settings and share them -- there has to be a lot of 5870 owners.

 

[Keysplayr]

Keys, you screwed up Stalker. No site shows close to 50% difference as you do, not to mention results going the opposite direction as AT.
Just be a man and admit it.

I think out of the two of us, I am the one being the man. You're approach to discrediting my numbers here could have been a lot more mature. I thought you really wanted to find a solution, but my initial apprehension that you were just playing a game seems more accurate.
If you can explain to me HOW it is even possible that I screwed up Stalker benches, I'm all ears and willing to correct a mistake, but please don't try and take the "man up" road when you yourself are just beating around the bush with what you really mean to say but are afraid to.
Regards...