AMD ATI Radeon HD 4870 X2 @ [H]

Kowan said:
Many do claim they rush right to the benchmarks and don't even look at the test system specs.
Fast food times we live in...
Click to expand...

Hey I only read the first half of your post, can you post the second part again here so that I can read it? :p
 
BigCactus said:
Regardless of what drivers they are using, the benchmarks seem a lot lower then most of the review sites...especially in Grid at 2560x1600.
Click to expand...

Actually playing a game vs. a time demo creates very different results.
 
BigCactus said:
Regardless of what drivers they are using, the benchmarks seem a lot lower then most of the review sites...especially in Grid at 2560x1600.
Click to expand...
Again, who cares? if they use some random benchmark with beta drivers, that has no real relation to how the card will perform with stable, final-build drivers in real gameplay. I hope they like their shiny graphs and wholesale conclusions that mean ABSOLUTELY NOTHING in terms of the way that the game performs in real life on MY system.
 
Modred189 said:
Again, who cares? if they use some random benchmark with beta drivers, that has no real relation to how the card will perform with stable, final-build drivers in real gameplay. I hope they like their shiny graphs and wholesale conclusions that mean ABSOLUTELY NOTHING in terms of the way that the game performs in real life on MY system.
Click to expand...

+1.
 
thanks for the review and for putting the performance of the X2 into perspective. great comparisons made and a really good read.

i wonder how the 260 respins will fair... right now i've caught an interest in 260 sli for the upcoming fall line-up of games.

and a big hello to the ravenous fans of arguments!
 
BigCactus said:
Regardless of what drivers they are using, the benchmarks seem a lot lower then most of the review sites...especially in Grid at 2560x1600.
Click to expand...

I have looked at some of the data around the web on the X2 with much anticipation.
As an ordinary guy, I have some results that are dead on and other data that just doesnt fall in line.
In my hands, I haven't seen a great difference in the Beta 8.54 driver versus the WHQL 8.8 release, save for a few frames in COD4. However, that may well be my style of play and settings.
I get nowhere near the performance in Crysis with TWO X2s that I've seen some report (even [H]) with one X2; regardless of driver version.

Truth be told, I just want to see ATI give me some drivers that let my two cards go batshit on these games.........:eek:.........right now they just seem too held back.
 
young boy said:
Is this solved with 8.8? Read about this problem about a month ago.
Click to expand...

Some people still have problems with that on the 8.8 release and 8.9 beta drivers. I'm experiencing this one of my personal machines with my 4870 X2's and not the other.
 
Dan_D said:
Actually playing a game vs. a time demo creates very different results.
Click to expand...

I think he's talking about the TechReport review, which reported a 94 FPS average for the 4870X2 and 57 FPS for the GTX 260 SLI at 2560x1600 4xAA. And they played the game too, rather than using a time demo.
 
Dan_D said:
Actually playing a game vs. a time demo creates very different results.
Click to expand...

Problem with this is that driverheaven doesn't use time demos in Grid...

dirverheavengrid.jpg


I think your numbers are way off. The fact you guys were not able to enable 4xaa at 2560x1600 in grid on the 4870x2 seems quite strange. Of course I see you are using adaptive for some reason...:confused:
 
BigCactus said:
Problem with this is that driverheaven doesn't use time demos in Grid...



I think your numbers are way off. The fact you guys were not able to enable 4xaa at 2560x1600 in grid on the 4870x2 seems quite strange. Of course I see you are using adaptive for some reason...:confused:
Click to expand...
What drivers did they use?
 
BigCactus said:
Problem with this is that driverheaven doesn't use time demos in Grid...

dirverheavengrid.jpg


I think your numbers are way off. The fact you guys were not able to enable 4xaa at 2560x1600 in grid on the 4870x2 seems quite strange. Of course I see you are using adaptive for some reason...:confused:
Click to expand...

When I try to enable 4XMSAA in-game myself and it does weird things. I can make a whole lap around some tracks with barely any slow down, but when I start the next lap it slows to a crawl, then all of the sudden catches up and BOOM into a wall... I bought my copy of GRID through Steam though if that makes a difference.
 
BigCactus said:
I think your numbers are way off. The fact you guys were not able to enable 4xaa at 2560x1600 in grid on the 4870x2 seems quite strange. Of course I see you are using adaptive for some reason...:confused:
Click to expand...

omg...just please stfu already :rolleyes:
 
Blackforge

Personally I'd get your cpu clock speed up to 3.2 to 3.4. Also make sure you have the latest Grid patch installed. I'd also monitor those temps...and try to bump up the 4870x2 fan speed manually to 50% temporarily to see if the problem goes away. If you don't know how to adjust your video card fan manually, I can tell you how.

Oh and what drivers are you using?
 
BigCactus said:
Sorry, when every other review shows a different story, I'll question the one that doesn't make any sense. So how about you stfu, seriously :p
Click to expand...



First off, driverheaven doesn't claim that the game runs great on a 4870x2, infact they say....


driverheaven said:
what this does not show is that the game suffers from random drops in performance on the X2 which result in the reduction of framerates, sometimes for a few seconds and at other times until the game is restarted.
Click to expand...

Second off lets see what was actually benchmarked...

Hard says they benchmarked....
HardOCP said:
We have chosen to do our run-through in the Race Day Pro Tuned mode in Okutama on the Grand Circuit with three laps.
Click to expand...

DriverHeaven says they benchmarked.... oh wait, they didn't say what they benchmarked. All we have is this
Then we record with FRAPS across several in game levels recording the averages.
Click to expand...

Which means they are not benchmarking the same things. NO NAME CALLING - Kyle
 
Has anyone tried one of these one UT3, or any other UE3 games? Just wondering since this seems to be the one game that the newer ATI cards don't handle as well as the nVidias. I just can't decide anymore, I think 'Ah, finally this MB/CPU/ ohhh, wait , but what about, nope that wont work' and then have to start over :confused:

I really need to upgrade...
 
DeadSkull said:
Thank you very much for the review, Brent and Kyle.


What kinds of fps do you get with 2xaa and 4xaa? Are they below 25 fps avg as well?
Click to expand...

about 20 ish with 4aa and 23 ish with 2 aa, 25 ish without AA.
And thats all with 16x AF
But as noted, motion blur makes it feel smooth.

even at 15 FPS, it still feels smooth, more like 30 FPS or more.
 
All this talk finally got me to pull my 8800GTX out tonight and install an ASUS 4870X2. The GTX was water cooled and was waiting for my 4870X2 water block. Went to Home Depot and got me a maleXmale hose barb to close the loop. Installing ATI drivers as I type.

(Killed my ES-800 PSU install during all of this as well, but that is for another thread.)
 
BigCactus said:
Sorry, when every other review shows a different story, I'll question the one that doesn't make any sense. So how about you stfu, seriously :p
Click to expand...

I would highly suggest that you disregard anything we say in our reviews. You would be much better off listening to other sites for their opinions.
 
wapos said:
Has anyone tried one of these one UT3, or any other UE3 games? Just wondering since this seems to be the one game that the newer ATI cards don't handle as well as the nVidias. I just can't decide anymore, I think 'Ah, finally this MB/CPU/ ohhh, wait , but what about, nope that wont work' and then have to start over :confused:

I really need to upgrade...
Click to expand...

It handles UT3 fine. I'm not sure which one provides better FPS but UT3 is generally smooth and quite playable at 2560x1600 with 4xAA and 16xAF.
 
wapos said:
Has anyone tried one of these one UT3, or any other UE3 games? Just wondering since this seems to be the one game that the newer ATI cards don't handle as well as the nVidias. I just can't decide anymore, I think 'Ah, finally this MB/CPU/ ohhh, wait , but what about, nope that wont work' and then have to start over :confused:

I really need to upgrade...
Click to expand...

We do not test with UT3 anymore simply because it is not demanding on any current generation hardware.
 
BigCactus said:
Blackforge

Personally I'd get your cpu clock speed up to 3.2 to 3.4. Also make sure you have the latest Grid patch installed. I'd also monitor those temps...and try to bump up the 4870x2 fan speed manually to 50% temporarily to see if the problem goes away. If you don't know how to adjust your video card fan manually, I can tell you how.

Oh and what drivers are you using?
Click to expand...

Doubtful my processor speed is to blame at my resolution (2560x1600 plus take a look at the CPU Scaling topic). I run the 8.8 Official drivers and they run more consistently than the 8.54 did in all of the games I've tried so far. In GRID, you max out all of the details, etc and turn up to 4XMSAA and above and it behaves how I described. 2XMSAA and below runs great (track run through exactly like the [H] review). So, I've been able to reproduce their results and agree with them wholeheartedly at the moment.

I do think there might be a powerplay/driver issue sometimes though, as the loading screens slow to a crawl compared to the actual gameplay/garage in comparison, in fact you can hear the fan speed kick up during the gameplay/garage. I do not appear to be having fan control/temp issues. I am very familiar with a lot these problems as the 3870X2 were even more affected early on... and I had two of them.
 
how can it draw 540 watts under load when the system draw with single 4870 is only 280? theres NO way those numbers can be correct because a second 4870 gpu cant possibly pull 260 watts by itself. :eek:

sorry if this was already answered but I skimmed through the comments and didnt see it mentioned.
 
Uhmm, did I miss something on the article, or this thread?

Are there any issues with microstutterring with either 4870x2 or 260 GTX SLI?

BTW, I actually never seen microstuttering, but would like the author or editor to confirm that both dual gpu solutions are smooth.
 
Dan_D said:
UT3 caps it's FPS to 60FPS by default. You can over ride that if you want but I don't feel like its' necessary. On the 4870 X2 it just sits at 60FPS all the time rarely dropping below 55 even at 2560x1600 with 8xAA and 16xAF. (This was off of one 4870 X2 mind you.)
Click to expand...

That sounds reasonable enough. I was looking at the Gigabyte MA790GP-DS4H review and had pretty much settled on that. I would like to stick with AMD, but was unsure about the 4870 and the 4870x2, but now, maybe I do have a plan.

Thanks Dan
 
cannondale06 said:
well that would make sense but the GTX260 SLI numbers must be a typo also because theres no way it could be drawing 260watts more than a single GTX280. why is nobody else even questioning these numbers because they are clearly a mistake? :confused:
Click to expand...

Please be patient, Brent will address your question when he returns. Also, if the mistake is so clear, please tell me what it is. He did the power numbers on this article, so I am out of the loop on it, but I have a little experience. My numbers from the preview are below.

BUT, to compare, and these are totally different systems...Also, I am not sure exactly how Brent got his number either. If there was a big CPU pull, I an see where a single 4870x2 SYSTEM would pull 540w at the wall. Remember he is running a top end quad core overclocked to 3.66GHz. And those things soak up the power when OCed. Also keep in mind that is wattage AT THE WALL. So with a good PSU you can figure 540*.8 = 432 actual power draw from the system PSU. You have to remember PSU efficiency.

To quote from the article, again.

To measure power consumption on these video cards, we used a Kill-a-Watt in-line power monitoring device from P3 International. Please note that the power usage numbers listed below reflect total system power usage. The Kill-a-Watt device sits between the wall and our power supply, so the power readings shown reflect the total amount of power our test computer used.
Click to expand...

So if a the mobo and processor use 150w under load and the GPUs 268 (as is high spec) it is already at 418wattage. This is not including any other system components.

1215976304wcOJlico3z_9_1.gif


Your concerns will be further looked into when Brent returns as he did the testing.
 
Finally got my 4870x2 installed and running. Had to pull my 6600GT out that supports monitors 3 and 4. ATI driver did not like being around that. Now that I have it all working, I will likely be able to reinstall the 6600GT, or at least that has been my experience in the past.

Another thing that prompted me to not wait for the 4870x2 water block was a NV BSOD I got earlier today trying to run Far Cry. Suckage. Figured if I was going to clean up the driver mess I might at well start new.
 
good read. i thought it would takes gtx280 SLI to match HD 4870x2 because 4870x2's multi gpu scaling runs so well, i was wrong, gtx260 SLI matched HD 4870x2 perfectly. thanks [H].
 
From what I read of the article, the GTX280 seems to be on the same level as the 4870X2. I don't know the situation in the US, but in Australia the GTX280 can be had for $200 less that the 4870X2. Isn't that much better value? Then think of the SLI benefits ...
 
What is interesting to me, more so than any review can give, is how most people that I've read on these forums that have gone from the GTX280 to a 4870x2 feel the switch was worthwhile.
 
With the way they are testing the cards, you would get the impression that they are close in performance. In reality, they are not. It's nice to see a different approach maxing the highest possible IQ for each card...but this also in a way skewers results to make the cards look equivalent to each other when your comparing fps numbers side by side.

Fact is the majority...even those with 2560x1600 monitors are going to max in games settings...and use box 4xaa/16xaf.

Also Power draw readings at load are definitely incorrect as well, and 4870CF was left out in this review.. The review is incomplete and flawed at best.
 
nösferatu;1032985937 said:
good read. i thought it would takes gtx280 SLI to match HD 4870x2 because 4870x2's multi gpu scaling runs so well, i was wrong, gtx260 SLI matched HD 4870x2 perfectly. thanks [H].
Click to expand...

lol, I dont even understand where people got the notion that a 4870 x2 could match gtx 280 sli. Hell a single hd 4870 is equal to gtx 260 so unless sli scaling completely ****s up in a game, gtx 280 sli reigns supreme.
 
cannondale06 said:
how can it draw 540 watts under load when the system draw with single 4870 is only 280? theres NO way those numbers can be correct because a second 4870 gpu cant possibly pull 260 watts by itself. :eek:

sorry if this was already answered but I skimmed through the comments and didnt see it mentioned.
Click to expand...

That's what it was, I've double checked, I used total system power at the wall and used Grid to gauge peak power as I noticed it was demanding the most GPU activity with the 4870 X2.
 
I am tired of you trashing HardOCP in every post. I can take criticism, but you are obviously here to troll. Take it elsewhere. You no longer have access to our video card forums. - Kyle
 
Brent_Justice said:
That's what it was, I've double checked, I used total system power at the wall and used Grid to gauge peak power as I noticed it was demanding the most GPU activity with the 4870 X2.
Click to expand...

This is one of those areas where using a real world game as a metric is not a good idea I think. Let me talk to ATI on this and see what feel is the best way to load the GPU and memory bus on the card without generating too much of a CPU load.

The current ATITool "fuzzy cube" seems to be reaching into across both 4870x2 GPUs, but I was specifically told that 3D in a window would only be accelerated by one GPU. Maybe this has been "fixed." Got some mails into ATI now.

I will work on getting a better system setup to get power numbers.
 
BigCactus said:
I am tired of you trashing HardOCP in every post. I can take criticism, but you are obviously here to troll. Take it elsewhere. You no longer have access to our video card forums. - Kyle
Click to expand...

yay!!! <claps>

thank you, it will make my days surfing the video card forums so much more efficient now w/ out having to go through all the wasted bandwidth that B.C. created.

thanks again Kyle!
 
You must log in or register to reply here.
Back
Top