• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

bit-tech review 3870x2

they also said they performed the test manually inside the game. no cut seen or flybys
 
lol417 said:
bit-tech review is up since yesterday. According to their review it seems like the 3870x2 is a bit slower or close to the 8800gtx. About the same conclusion as [H]

http://www.bit-tech.net/hardware/2008/01/29/amd_ati_radeon_hd_3870_x2/14
Click to expand...

Even here the results differ from [H]. Check out UT3:

http://www.bit-tech.net/hardware/2008/01/29/amd_ati_radeon_hd_3870_x2/11

http://enthusiast.hardocp.com/article.html?art=MTQ1NCw2LCxoZW50aHVzaWFzdA==

I'm starting to wonder if the card was tested too early. In almost every review, the 3870GX2 came on top:
http://forums.vr-zone.com/showthread.php?t=230799

Also in tests where real time gaming was used as method:
http://www.driverheaven.net/reviews/AMDX2Review/testsystem.php

There are few tests of its video capabilities, but it put all other cards to shame in driverheavens review.

But, the results differ so much its hard to get a real impression of this card. I don't think its any faul play anywhere, but that the drivers are too immature to give good enough indication of its capabilities. CPU might play a bigger role here as well, compared to other cards, so it might explain some of the differences at least.

I hope that [H] will retest it later when it becomes more mature. I like reading their reviews :)
 
I dont know.. their numbers look good to me. The X2 averaged within 10% of the ultra and managed to beat it a few times. I am sure with better drivers it will perform even better. 450 vs 650 ish dollars ? I will take the x2 performance at 450 dollars. That's a price that will come down eventially. Then add in the overclocked versions that hopefully HIS will do with their turbo cool models I am sure it wil get better and better.
 
Tamlin_WSGF said:
Even here the results differ from [H]. Check out UT3:

http://www.bit-tech.net/hardware/2008/01/29/amd_ati_radeon_hd_3870_x2/11

http://enthusiast.hardocp.com/article.html?art=MTQ1NCw2LCxoZW50aHVzaWFzdA==

I'm starting to wonder if the card was tested too early. In almost every review, the 3870GX2 came on top:
http://forums.vr-zone.com/showthread.php?t=230799

Also in tests where real time gaming was used as method:
http://www.driverheaven.net/reviews/AMDX2Review/testsystem.php

There are few tests of its video capabilities, but it put all other cards to shame in driverheavens review.

But, the results differ so much its hard to get a real impression of this card. I don't think its any faul play anywhere, but that the drivers are too immature to give good enough indication of its capabilities. CPU might play a bigger role here as well, compared to other cards, so it might explain some of the differences at least.

I hope that [H] will retest it later when it becomes more mature. I like reading their reviews :)
Click to expand...
i think Kyle is going to run a ingame bench frap and record it... then rerun the recording and frap it again ....

this will clear up alot ... alot of people are puzzled how a custom recorded time demo ingame.. is that much different than live in game frapping
 
POPEGOLD said:
i think Kyle is going to run a ingame bench frap and record it... then rerun the recording and frap it again ....

this will clear up alot ... alot of people are puzzled how a custom recorded time demo ingame.. is that much different than live in game frapping
Click to expand...

Nice! I hope he'll get to the bottom of this. I wonder also why the ingame testing done on other sites provide different results then on [H]. Is there some major difference in their setups, or is the ingame testing too unreliable to get consistent results?

Perhaps, as you mentioned in your other thread, that CPU might play bigger role with the 3870X2 then with the other cards? That the increased AI of the real time gaming is larger then the time demo's and therefore plays bigger role? Would be worthy to know if a CPU upgrade would turn the tables when comparing GFX cards. :)

Edit: I was also wondering if ATI powerplay can cause different results with clocking down the processor. Is it possible to turn it off?
 
Tamlin_WSGF said:
Nice! I hope he'll get to the bottom of this. I wonder also why the ingame testing done on other sites provide different results then on [H]. Is there some major difference in their setups, or is the ingame testing too unreliable to get consistent results?

Perhaps, as you mentioned in your other thread, that CPU might play bigger role with the 3870X2 then with the other cards? That the increased AI of the real time gaming is larger then the time demo's and therefore plays bigger role? Would be worthy to know if a CPU upgrade would turn the tables when comparing GFX cards. :)

Edit: I was also wondering if ATI powerplay can cause different results with clocking down the processor. Is it possible to turn it off?
Click to expand...

I'm seriously starting to think ati hasn't optimized their 64 bit drivers yet..
 
Tamale said:
I'm seriously starting to think ati hasn't optimized their 64 bit drivers yet..
Click to expand...

That might be a plausible explaination, since most reviews are in 32-bit, while [H]'s in 64-bit. I tried to ask Bennett several time about running a test in 32-bit, but no response. :mad:
 
Tamlin_WSGF said:
That might be a plausible explaination, since most reviews are in 32-bit, while [H]'s in 64-bit. I tried to ask Bennett several time about running a test in 32-bit, but no response. :mad:
Click to expand...

it could be, but I would guess that rather large portions of code are identical between the two platforms.
 
While I love a good conspiracy. There are so many factors in Testing hardware its impossible nearly to have identical results. I think Kyle's approach is fine if it were combined with custom timedemos and the same settings used accross the board.

The way its set up now its more like a game review instead of GPU review. Looking at 1920x1200 with 4AA 16 AF accross the video cards for example, give us an idea on how well that GPU can compare in performance vs the other. Having one with extreme textures, medium shadows, and another with high textures, high shadows doesnt really show the whole story.

I dont know how much work goes into making sure the level they choose shows their sponsor "nvidia" in the best possible light either. I am not saying it DOES happen I am just saying we have no way to know it doesnt. That's why I take the over view of all the websites and their numbers and that gives me a fairly over all performance value of the product being reviewed.

My own summary for the X2 is that it offers GTX or Close to GTX performance in most games if the driver is right. With newer drivers I am sure It will get better with time. After all nvidia has had a long time to optimize the GTX drivers.
 
kllrnohj said:
it could be, but I would guess that rather large portions of code are identical between the two platforms.
Click to expand...

On driver level its not that easy to port code unfortunately. If it were that easy, we'd have 64-bit driver support in a snap instead of running buggy drivers in 64-bit as many early adopters of 64-bit Vista/XP have done earlier.

If it were a program, it would be easier. Still, even with porting of software to 64-bit, its not optimised for 64-bit.
 
FlyinBrian said:
While I love a good conspiracy. There are so many factors in Testing hardware its impossible nearly to have identical results. .
Click to expand...

Its not about conspiracy for me at least, though for some it might be. My two recent cards were Nvidia and before that one ATI and before that Nvidia again. I don't care about the ATI vs. Nvidia war as much either. But, since I was looking for a new card and tired of waiting for the new G100 from Nvidia, the large deviations between [H] and the rest made it hard to make a descition. I did order the 3870X2 today, but I would have preferred if the results didn't differ so much. Kinda makes it a shot in the dark, since I rely maybe a bit more on [H]'s review.

I agree that many factors needs to be taken into account and that FPS comparison doesn't always match. But, they are using within the same review the same settings, so conclutions should be closer anyway. If not, the whole testing is void and worthless.

Most sites deemed the 3870X2 as the new top card, both those with time demo's and those with ingame testing using fraps. Usually, it would mean majority rules, but for me it means I wish to know why.
 
Tamlin_WSGF said:
Its not about conspiracy for me at least, though for some it might be. My two recent cards were Nvidia and before that one ATI and before that Nvidia again. I don't care about the ATI vs. Nvidia war as much either. But, since I was looking for a new card and tired of waiting for the new G100 from Nvidia, the large deviations between [H] and the rest made it hard to make a descition. I did order the 3870X2 today, but I would have preferred if the results didn't differ so much. Kinda makes it a shot in the dark, since I rely maybe a bit more on [H]'s review.

I agree that many factors needs to be taken into account and that FPS comparison doesn't always match. But, they are using within the same review the same settings, so conclutions should be closer anyway. If not, the whole testing is void and worthless.

Most sites deemed the 3870X2 as the new top card, both those with time demo's and those with ingame testing using fraps. Usually, it would mean majority rules, but for me it means I wish to know why.
Click to expand...


It depends on what card you have Now.. Obviously if your using a GTX your not going to notice much real world performance if any. If your going from an older card to this system then you should see speed increases as long as your not CPU bottlenecked of course.
 
FlyinBrian said:
It depends on what card you have Now.. Obviously if your using a GTX your not going to notice much real world performance if any. If your going from an older card to this system then you should see speed increases as long as your not CPU bottlenecked of course.
Click to expand...

My card is a 8600GT 512mb, so I would get a good upgrade anyway. Was trying to hold out until the g100 came as mentioned. :)
I'm using a c2d E6600 OC'd to 2997mhz.

That doesn't mean this don't matter. If Bennett figures out why the difference, it might effect my usage of this or even make me buy hardware that supports the card better. Since I'm often asked about my opinion elsewhere, the truth is important to find.

Its a shame about all the war that has come between the sites because of the contraversis between [H]'s review and the rest. I'm sure Bennett will rise above this bickering and do what he does best: get to the truth of this and figure out what makes the difference regardless of result.
 
Well I would be interested in your experience when you get yours in. My next build will be a Quad core and either one of these X2s. I am like you, i go with whoever gives me the best performance for the money. I have been using Nforce boards and GPU's the last 4 years so I am itching for a new ATI solution. I have always thought their image quality was a bit better than Nvidia's but that is purely subjective.
 
kllrnohj said:
The articles used different bot amounts (12 v. 23) and different maps (shangri la vs. sandstorm) - that could account for the differences...
Click to expand...

Outcome should still be the same, though framerate could differ. Otherwise, such tests are meaningless if you cannot even spot the differences between GFX cards with them.

If site a, used different machine and map, there would be FPS differences compared to site b, but both should show which card that performs best in that particular game if any of those tests are worth something.
 
FlyinBrian said:
Well I would be interested in your experience when you get yours in. My next build will be a Quad core and either one of these X2s. I am like you, i go with whoever gives me the best performance for the money. I have been using Nforce boards and GPU's the last 4 years so I am itching for a new ATI solution. I have always thought their image quality was a bit better than Nvidia's but that is purely subjective.
Click to expand...

I think we are alike there. :) The card is inside the machine and the logo, be ATI or NVIDIA, matters little. What I see on the screen is what counts.

Its nice that there are different companies, so that is said. Gives us healthy competition, more advances and better prices. :D

I chose ATI this time because of impatience, but I also considered the upcoming CrossfireX and the possibility to run dual X2 if the g100 or r700 doesn't bring anything significant. The reviews showing its video capabilities are also nice, since it really outperforms any other cards on HD:
http://www.driverheaven.net/reviews/AMDX2Review/movies.php

I'm thinking of getting a blue-ray player on my computer, so this might be useful at least.
 
if review sites wrote more than a few sentences at the end of their benchmarking graphs we would have a far better picture of how these cards perform, and consequentially how and why results differ so much among the various review sites. maybe the readers who prefer the 15 chart-one sentence reviews see something that i don't.
 
Nice, Yeh you can get a blue ray drive now for like 200 ish I think. From zipzoomfly.com I am also gona do this. But before I do anything I have got to get a HD monitor.
 
rennyf77 said:
if review sites wrote more than a few sentences at the end of their benchmarking graphs we would have a far better picture of how these cards perform, and consequentially how and why results differ so much among the various review sites. maybe the readers who prefer the 15 chart-one sentence reviews see something that i don't.
Click to expand...

I enjoy that the reviews have different focus points. Its easier then to supplement one review with the other. One of the best reviews of the 3870X2 is the one from driverheaven. I'm not speaking of the results and conclution, but technically. Its balanced with having many games and tests (as compared to just a couple of games) and they write interesting. Best of all, I like the function where they include full bmp of the game they reviewed. Its nice then to click on the picture and then run an F11 (full screen) of the browser to see how it would look on my screen. :D
 
FlyinBrian said:
Nice, Yeh you can get a blue ray drive now for like 200 ish I think. From zipzoomfly.com I am also gona do this. But before I do anything I have got to get a HD monitor.
Click to expand...

In Norway, the selection is small and its still expensive. I have a HD monitor, but desided to wait a bit until the choice is bigger before buying a Blueray. Hopefully there comes some models soon with burning capabilities for blueray as well. :D
 
Tamlin_WSGF said:
Outcome should still be the same, though framerate could differ. Otherwise, such tests are meaningless if you cannot even spot the differences between GFX cards with them.

If site a, used different machine and map, there would be FPS differences compared to site b, but both should show which card that performs best in that particular game if any of those tests are worth something.
Click to expand...

Except that different maps can cater to the strengths of the card. One map might be more shader intensive or be more spacious than another, etc... Especially in UT3, where despite being on the same engine, the maps can vary greatly in graphical load and look. My card tore it up in heatray, but got bogged down in shangri la... I was just throwing it out there as a possibility more than anything.

I still think the whole thing is being way overblown. Both cards are great, they are around the same price, and they trade blows in performance - who really cares? You can't go wrong either way, but I personally would lean towards the 3870X2 because most of the time I'm not gaming, and therefore the 3870X2 will overall use less power and put out less heat for me. I'm not seeing the "conspiracy" or problem here - every review has liked the card that I've seen...
 
I dno, People are doing too much heresay on this card- Like driver optimisations sure they help but when the 8800 GT came out it rocked from the start and no one was looking for any driver upgrades to justify a purchase...I dont think thats a good way of thinking to buy a card.

Also the other thing people arent taking into consideration is how far these cards can overclock. The Ati card doesnt overclock very well- if pretty much at all...825 is stock and most reviewers have gotten it to 850-860, secondly the memory pretty much isnt worth overclocking - bit tech has the only review ive seen even gotten a higher memory clock.

Where as the 8800 Ultras come out around 675/1110 in most cases from reviewers...

Now some can say you shouldnt overclock there cards, but who doesnt Overclock there GPU's.....

The 8800 ultra is about 600 and the hd 3870X2 is about 500+ here in Canada.
 
i think bioshock in this review is interesting

as they increase the res all the single cards take a huge drop in performance.

The ultra goes from 91.7/46.0 to 45.8/24.0 Basicly looses half its performance.

The x2 goes from 85.5/45.o to 61.8/35.0 .

This game it scales very very well

Then look at Quake wars. IT does the same thing , however this time performance sucked from the onset .


I wonder how much drivers will improve this .

I also really would like to see a review with the ram clocked to the same as the 3870.

Right now its running 1.80ghz effective vs 2250 effective. Thats over 400mhz effective speed diffrence.

I doubt anyone can get the ram to run that fast on the x2 , however it be interesting to see what a 100mhz diffrence will bring. I wonder if its bandwidth starved and they did this for the x2 with gd4 card priced slightly higher.
 
btdvox said:
I dno, People are doing too much heresay on this card- Like driver optimisations sure they help but when the 8800 GT came out it rocked from the start and no one was looking for any driver upgrades to justify a purchase...I dont think thats a good way of thinking to buy a card.

Also the other thing people arent taking into consideration is how far these cards can overclock. The Ati card doesnt overclock very well- if pretty much at all...825 is stock and most reviewers have gotten it to 850-860, secondly the memory pretty much isnt worth overclocking - bit tech has the only review ive seen even gotten a higher memory clock.

Where as the 8800 Ultras come out around 675/1110 in most cases from reviewers...

Now some can say you shouldnt overclock there cards, but who doesnt Overclock there GPU's.....

The 8800 ultra is about 600 and the hd 3870X2 is about 500+ here in Canada.
Click to expand...


Well the GT wasnt really super new so I doubt they had to do a lot of tweaking on the drivers especially since its just a single gpu. I think if ATI can get more developers to optimize for their hardware we will see even better performance. I am not being fanboyish as I been using nvidia stuff for the last 4 years. But seeing 800 dollar ultras really made me hope ATI can compete and so my next build will consist around their hardware.
 
gljvd said:
i think bioshock in this review is interesting

as they increase the res all the single cards take a huge drop in performance.

The ultra goes from 91.7/46.0 to 45.8/24.0 Basicly looses half its performance.

The x2 goes from 85.5/45.o to 61.8/35.0 .

This game it scales very very well

Then look at Quake wars. IT does the same thing , however this time performance sucked from the onset .


I wonder how much drivers will improve this .

I also really would like to see a review with the ram clocked to the same as the 3870.

Right now its running 1.80ghz effective vs 2250 effective. Thats over 400mhz effective speed diffrence.

I doubt anyone can get the ram to run that fast on the x2 , however it be interesting to see what a 100mhz diffrence will bring. I wonder if its bandwidth starved and they did this for the x2 with gd4 card priced slightly higher.
Click to expand...

Yeh I am hoping HIS comes out with their turbo cool model fairly soon with factory overclocks.
 
^^ Yeah i think we all knew Nvidia got greedy cuz the 8800 GTX was doing so well and just put a figure on the 8800 ultra- but i see them for 600 bucks at ncix so thats not bad for that card, i see 3870x2's for about 500+ on Ncix, i dont really buy from any other store
 
Tamlin_WSGF said:
I enjoy that the reviews have different focus points. Its easier then to supplement one review with the other. One of the best reviews of the 3870X2 is the one from driverheaven. I'm not speaking of the results and conclution, but technically. Its balanced with having many games and tests (as compared to just a couple of games) and they write interesting. Best of all, I like the function where they include full bmp of the game they reviewed. Its nice then to click on the picture and then run an F11 (full screen) of the browser to see how it would look on my screen. :D
Click to expand...

Driver heavens review is great but there is something really wrong with there Quake Wars benchmark, This game is know to be Nvidia oriented and if you tkae a look at ANY other review you'll see a HUGE difference between a 8800 GTX/Ultra and a 3870X2.
 
btdvox said:
Driver heavens review is great but there is something really wrong with there Quake Wars benchmark, This game is know to be Nvidia oriented and if you tkae a look at ANY other review you'll see a HUGE difference between a 8800 GTX/Ultra and a 3870X2.
Click to expand...


There not the only ones showing x2 beating gtx in quake wars. www.hothardware.com also shows a win there.
 
Tamale said:
I'm seriously starting to think ati hasn't optimized their 64 bit drivers yet..
Click to expand...

I'll say it again, video drivers aren't made like that. They are built from common files, and packaged differently for 32-bit and 64-bit. The drivers are not coded as two separate entities. Performance and optimization and bug fixing that goes on in 32bit also ends up in 64bit.
 
Sr7 said:
I'll say it again, video drivers aren't made like that. They are built from common files, and packaged differently for 32-bit and 64-bit. The drivers are not coded as two separate entities. Performance and optimization and bug fixing that goes on in 32bit also ends up in 64bit.
Click to expand...

Either provide links or quit sounding like a broken record. A quick review of the release notes for the latest AMD drivers shows a couple fixes that are specific to the 64-bit version of Vista, which indicates that the drivers are coded separately and that performance fixes may be optimized better for one than the other.

If you look at the Nvidia Forceware 169 release notes, you will see a completely different set of fixes for Vista 32 than for Vista 64. Again, this shows that the two drivers are coded separately. Even if the same fixes are attempted on both drivers, it is likely that the coding was done better on one than the other. Guess which version AMD likely put more time into optimizing??
 
axel said:
Either provide links or quit sounding like a broken record. A quick review of the release notes for the latest AMD drivers shows a couple fixes that are specific to the 64-bit version of Vista, which indicates that the drivers are coded separately and that performance fixes may be optimized better for one than the other.

If you look at the Nvidia Forceware 169 release notes, you will see a completely different set of fixes for Vista 32 than for Vista 64. Again, this shows that the two drivers are coded separately. Even if the same fixes are attempted on both drivers, it is likely that the coding was done better on one than the other. Guess which version AMD likely put more time into optimizing??
Click to expand...

The 64-bit ATi drivers use large amounts of the same code as the 32-bit ones. How do I know this? I just checked. Download the drivers, and run diff on some of the DLLs - you will see they are identical. You will also see that the 64-bit driver install comes with 32-bit DLLs. This, of course, makes a lot of sense, as most games are still 32-bit, and will need 32-bit calls.

Just because the release notes mention per-platform specific bug fixes doesn't mean the two are completely separate. That is why there are things called libraries, so that different programs (in this case different drivers) can share code :rolleyes:
 
kllrnohj said:
The 64-bit ATi drivers use large amounts of the same code as the 32-bit ones.
Click to expand...

My point is that there are still "large amounts" of code that are different between the two drivers. And that is the code that we should be focusing on when assessing performance differences. Naturally, identical libraries would not lead to performance differences so those libraries don't enter into the picture here.
 
POPEGOLD said:
i think Kyle is going to run a ingame bench frap and record it... then rerun the recording and frap it again ....

this will clear up alot ... alot of people are puzzled how a custom recorded time demo ingame.. is that much different than live in game frapping
Click to expand...


I think we have gotten to the bottom of how to do this with Crysis. Still working on COD 4 as it does not like single player demos being recorded for some reason.

Also worth noting, I did speak with AMD PR this morning, and still not reponse as to our 3870 X2 results being "wrong."
 
Tamlin_WSGF said:
Nice! I hope he'll get to the bottom of this. I wonder also why the ingame testing done on other sites provide different results then on [H]. Is there some major difference in their setups, or is the ingame testing too unreliable to get consistent results?
Click to expand...

The bottom line is that we will only be able to prove what HardOCP is doing. There is little we can do to prove or disprove other sites results. And to be frank, I don't care what any other sites say. All we need to know is that we are doing things right at the [H], the rest I have no control over. We want to be able to look in the mirror in the morning and know we have done the right thing by our readers. And that is what we are going to live by around here. PR folks and site owners can get in a huff because they don't like our attitudes or the way we present ourselves sometimes, but those are thier proplems, not ours.
 
Tamlin_WSGF said:
I tried to ask Bennett several time about running a test in 32-bit, but no response. :mad:
Click to expand...

Windows Vista 64-bit is our current test environment for all our hardware reviews. We do motherboards in XP 32-bit and Vista 64-bit, but we do not do the testing twice. One section is done in one OS and another section in the other OS.

We have been testing with Vista 64-bit in video card evaluations for months now. I do not see that changing.
 
FlyinBrian said:
I dont know how much work goes into making sure the level they choose shows their sponsor "nvidia" in the best possible light either.
Click to expand...

NVIDIA has never spent one dollar with us in advertising. ATI however has in the past, but not in recent years. We cut them off about 4 years ago due to them not paying their invoices for months.

To answer your statement, no work goes into making sure any company is placed in the best possible light. We evaluate video card experience on a playing field that we think the high end gamer expects to be able to use the product at.

BFGTech is our longest running advertiser of NVIDIA based graphics cards. And you can ask them what their ad dollars buy, and they will tell you "ads." We call it like we see it.

We have given ATI their fair share of awards this year. But the product experience earned them, not ad dollars (which is $0).
 
axel said:
Nothing wrong with that. But before jumping to conclusions and claiming that the disparities in the test results between [H] and other sites are due to "real world" vs. "canned" testing, it would be wise to eliminate the OS as the reason first. If you don't have time to test in Vista 32, fine, but then you shouldn't be making unsubstantiated claims either as was done in your Editor's Note addendum to the 3870X2 review.
Click to expand...
A valid point. There are differences between the way this site benches cards and some others, but there were some wild accusations made by Kyle and others here that seemed to paint the entire industry with a broad stroke. There could be other factors such as driver inconsistencies, CPU choices, OS, etc.

I think I hear Kyle typing "We benchmark differently, and we stand by our results and methods, and if you don't like it you can leave"
 
You must log in or register to reply here.
Back
Top