• 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.

AMD 6800 Series Review

H

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
The crew at Overclockers Club have posted their review of AMD’s new 6800 series video cards. The gang ran the HD 6850 and HD 6870 through a complete set of benchmarks and threw in a little overclocking for good measure. For comparison purposes, you can see our evaluation here.

The HD 68XX series are cards that target NVIDIA's GTX 460 and hit the mark on performance and price. The performance of the HD 6870 was actually better than I was expecting and shows that you can get excellent performance for a sub $250 price tag. All-in-all, the two series delivered excellent performance in the game testing by comparison to a stacked field.
Click to expand...
 
Ehh, I can't stand their benchmark chart methods. Too annoying to try and pick and read, pick and read.

I think TomsHardware has the best bench chart format. And it's done in super easy to use exel charts as well. Can't for the life of me figure out why more sites don't use something similar.
 
I'd like to know why they are including Batman Arkham Asylum with 4xAA in their benchmarks. It gives a huge performance advantage to Nvidia and is very misleading.

The advantage of Nvidia in-game AA, is that the engine can run AA selectively on scenes, whereas Forced AA in AMD CCC is required, to use brute force to apply AA on every scene and object, requiring much more work.

I do like the fact that they include overclocked and stock clocked results for all the cards in their reviews.
It's too bad that most of their benchmarks and methodology are biased towards Nvidia.
 
Nuby1Canuby said:
I'd like to know why they are including Batman Arkham Asylum with 4xAA in their benchmarks. It gives a huge performance advantage to Nvidia and is very misleading.

The advantage of Nvidia in-game AA, is that the engine can run AA selectively on scenes, whereas Forced AA in AMD CCC is required, to use brute force to apply AA on every scene and object, requiring much more work.
Click to expand...

Batman A.A. GOTY edition doesn't have the vendor ID check, so AA is done exactly the same on both Nvidia and AMD. They don't say what version they're running, so we can't make any kind of claim of bias.
 
I hope people recognize the input lag associated with MLAA... I find it annoying enough to turn it off in games.
 
thanks for the heads up ;) Just solidified my purchase choice :D

awaiting the [H] review though
 
sorry, let me rephrase, the more indepth review ;) I already read the launch review....the one with scaling and such si what i'm looking for
 
SicKlown42012 said:
Batman A.A. GOTY edition doesn't have the vendor ID check, so AA is done exactly the same on both Nvidia and AMD. They don't say what version they're running, so we can't make any kind of claim of bias.
Click to expand...

With the GTX 460 beating the 5870 in every single benchmark I would say their benchmark is definitely biased towards Nvidia. I wasn't aware the GOTY edition had enabled in game AA support, that's good to know.
I don't think they purposely picked benchmarks that favor Nvidia. Some sites just seem to choose benchmarks that favor either Nvidia or AMD.
 
Synthetic garbage. At least they added overclocking results. Is HardOCP and Guru 3d the only 2 sites that use real world gaming results? I saw this days ago btw but it's still not telling me what I want to know about the 2 competing gpu's.
 
colinstu said:
Techreport has the scoop too.

Still looks like the GTX 470/480 is the way to go if you want max performance.
Click to expand...

Strange that is the only review I remember seeing that nvidia fared so well at stock frequencies. I notice that they are using catalyst 10.9a and more than likely that is why. They should have used 10.10 beta. Also important to note that their CPU runs at stock 3.2Ghz.
 
Lord_Exodia said:
Strange that is the only review I remember seeing that nvidia fared so well at stock frequencies. I notice that they are using catalyst 10.9a and more than likely that is why. They should have used 10.10 beta. Also important to note that their CPU runs at stock 3.2Ghz.
Click to expand...
Agreed, from reviews I've seen, it should only be OCed 460s that are getting ahead the 6850
 
Lord_Exodia said:
Strange that is the only review I remember seeing that nvidia fared so well at stock frequencies. I notice that they are using catalyst 10.9a and more than likely that is why. They should have used 10.10 beta. Also important to note that their CPU runs at stock 3.2Ghz.
Click to expand...

LOL. Why are they using old drivers?
 
PRIME1 said:
Seems 460 is the way to go for DX11 over the 6800 series.
Click to expand...

So you did catch the part where that review was using older AMD drivers.. hence basically no other review sight had the 460 beating the 6850 much less the 6870
 
bastage said:
So you did catch the part where that review was using older AMD drivers.. hence basically no other review sight had the 460 beating the 6850 much less the 6870
Click to expand...

It's particularly amusing since AMD provided the new drivers to reviewers, so it's no excuse for them not having them. However, (through some magical means), they managed to obtain the new Nvidia drivers. Ironic.
 
Lord_Exodia said:
Strange that is the only review I remember seeing that nvidia fared so well at stock frequencies. I notice that they are using catalyst 10.9a and more than likely that is why. They should have used 10.10 beta. Also important to note that their CPU runs at stock 3.2Ghz.
Click to expand...

Yeah their results are weird because the 6850 is ties alot or beats the 6870 in some of their tests. Something got messed up on their end.
 
bastage said:
So you did catch the part where that review was using older AMD drivers.. hence basically no other review sight had the 460 beating the 6850 much less the 6870
Click to expand...

He's a known Nvidia troll, don't even acknowledge him and he will go away.
 
bastage said:
So you did catch the part where that review was using older AMD drivers.. hence basically no other review sight had the 460 beating the 6850 much less the 6870
Click to expand...

"The drivers used in this test will be the 10.9 Catalyst drivers for the old ATI lineup, the latest launch driver for the HD 68XX series and 260.89 Forceware drivers from NVIDIA for the GTX 480, 470, 465 and GTX 460 and 450. "

They used 10.10 for the 6850 and 6870, look at the GPU-Z screenshots too.
 
PersonalJ said:
"The drivers used in this test will be the 10.9 Catalyst drivers for the old ATI lineup, the latest launch driver for the HD 68XX series and 260.89 Forceware drivers from NVIDIA for the GTX 480, 470, 465 and GTX 460 and 450. "

They used 10.10 for the 6850 and 6870, look at the GPU-Z screenshots too.
Click to expand...
Agreed, but something is up with their results. I'm looking at Guru3D and the 6850 fares much better in every benchmark. The 6850 beats the 460 most of the time in their tests.
 
TheCapulet said:
Ehh, I can't stand their benchmark chart methods. Too annoying to try and pick and read, pick and read.

I think TomsHardware has the best bench chart format. And it's done in super easy to use exel charts as well. Can't for the life of me figure out why more sites don't use something similar.
Click to expand...

Wow those graphs are terrible :eek::eek::eek:

I like TechPowerUp.com for review graphs. Always very easy to read without staring at the things for 5 minutes while trying to determine which shade of off-green if the card you are looking for. Also I like their relative performance charts.
 
Lord_Exodia said:
Synthetic garbage. At least they added overclocking results. Is HardOCP and Guru 3d the only 2 sites that use real world gaming results? I saw this days ago btw but it's still not telling me what I want to know about the 2 competing gpu's.
Click to expand...

HardwareCanucks also uses real world gaming results. Where they differ from HardOCP is they then compare their real world gaming results, with available in-game benchmarks or time demos for both vendors. If they are within a few percent of each other they will use them.
The main difference is that they don't recommend the settings for maximum graphics quality that give you the best game play experience. You have to guess if a 60 FPS average is leaving some quality settings on the table, or if a 30 FPS average will give a laggy game play experience. It does have the advantage of giving you more benchmarks but I'd rather have a recommended game play setting then a whole bunch of benchmarks.
 
colinstu said:
Techreport has the scoop too.
Still looks like the GTX 470/480 is the way to go if you want max performance.
Click to expand...


From what I've seen the GTX 470 at stock speeds is on par with a 6870 but the GTX 470 has the advantage when they are both overclocked. The 5970 wipes the floor with the GTX 480 when they are both overclocked. Now that the 5970 is selling for $550, it is competitive with the GTX 480 from a price to performance standpoint.
 
Nuby1Canuby said:
From what I've seen the GTX 470 at stock speeds is on par with a 6870 but the GTX 470 has the advantage when they are both overclocked. The 5970 wipes the floor with the GTX 480 when they are both overclocked. Now that the 5970 is selling for $550, it is competitive with the GTX 480 from a price to performance standpoint.
Click to expand...

I expect the 480 to drop drastically in price when 69xx comes out. I'm hoping the 68xx series performance will increase (if marginally as more drivers appear.
 
Nuby1Canuby said:
From what I've seen the GTX 470 at stock speeds is on par with a 6870 but the GTX 470 has the advantage when they are both overclocked. The 5970 wipes the floor with the GTX 480 when they are both overclocked. Now that the 5970 is selling for $550, it is competitive with the GTX 480 from a price to performance standpoint.
Click to expand...

Except it's Crossfire.

Which just begs the question why wouldn't you grab 2 470GTXs for less than $500 and wipe the floor with the 5970? Or even grab 2 6870s (if you don't plan on OCing) for similar results.

I've never really gotten the whole idea of comparing Xfire/SLI to a TRUE single card solution. The 5970 isn't single GPU, so comparing it to a 480 or 5870 seems disingenuous. Compare it to a 460/470 SLI or 6870 Xfire (all 3 of which seem like better purchases to me).
 
In case anybody missed it in the Overclockersclub review, they used a Sapphire Toxic HD 5850 2GB card. It comes with an overclock of 765 core and 1125 memory. They ran it at 905 core and 1155 memory for their overclocked results. Not your typical 5850.
 
Nuby1Canuby said:
With the GTX 460 beating the 5870 in every single benchmark I would say their benchmark is definitely biased towards Nvidia. I wasn't aware the GOTY edition had enabled in game AA support, that's good to know.
I don't think they purposely picked benchmarks that favor Nvidia. Some sites just seem to choose benchmarks that favor either Nvidia or AMD.
Click to expand...

I think that’s because the Vanilla HD 5870 was competing against a factory OCed GTX 460 (EVGA FTW) which has also beaten the GTX 470, you can see similar results in anandtech review and the readers didnt like that though they also included a standard GTX 460 unlike this review and it’s hard to read graphs.
 
Vagrant_Zero said:
Except it's Crossfire.

Which just begs the question why wouldn't you grab 2 470GTXs for less than $500 and wipe the floor with the 5970? Or even grab 2 6870s (if you don't plan on OCing) for similar results.

I've never really gotten the whole idea of comparing Xfire/SLI to a TRUE single card solution. The 5970 isn't single GPU, so comparing it to a 480 or 5870 seems disingenuous. Compare it to a 460/470 SLI or 6870 Xfire (all 3 of which seem like better purchases to me).
Click to expand...

Not everybody has a motherboard that supports crossfire or SLI. The 5970 is the fastest card AMD makes and the 480 is the fastest card Nvidia makes.
Has anybody ever done a review with 2 overclocked 5970s versus 2 overclocked 480s? That would be interesting since the 5970s are monsters when overclocked but the 480s are better at dual card scaling (nuclear power plant not included). Most reviewers run the 5970s at their stock underclocked speeds. It leaves a lot of performance on the table and makes the cards look much slower then they really are. I think it was a bad decision on ATI's part, to underclock their 850 core 1200 memory chips to 725 core 1000 memory, to save power on an enthusiasts card.
The best bang for the buck would probably be 2 overclocked 6850s for $360.
 
firas said:
I think that’s because the Vanilla HD 5870 was competing against a factory OCed GTX 460 (EVGA FTW) which has also beaten the GTX 470, you can see similar results in anandtech review and the readers didnt like that though they also included a standard GTX 460 unlike this review and it’s hard to read graphs.
Click to expand...


The Overclockersclub review, also had the overclocked 460 beating the overclocked 5870 and 6870 in all resolutions in Batman. I'm not sure if the game engine just favors Nvidia at 4xAA, or if their version wasn't the GOTY edition. I can't find any other BAA GOTY reviews, to compare AA between Nvidia and AMD for a baseline.
 
Same approximate performance ---> 6870 and GTX 460 (1GB version)

Prices today @ Newegg ---> $239 vs $169 (AMIR for the Zotac 1GB)

Not a tough decision, IMHO.
 
I've been out of the PC tech scene for about 2 years; my last purchase was a 4870.
Seeing these cards, for that price, with those specs. Amazed. The 6850 is so tiny! What happened to the 10"+ cards?

Definitely might have to pounce on these, but I have a lot of catching up to do.

Same approximate performance ---> 6870 and GTX 460 (1GB version)

Prices today @ Newegg ---> $239 vs $169 (AMIR for the Zotac 1GB)

Not a tough decision, IMHO.
Click to expand...
Am I crazy or blind?
Anand shows the 6850 beating the 460 1GB almost always.

460 1 GB is something like $220 on NewEgg while the 6850 is around $180.
I'm seeing the opposite of what you just posted. Though I spent maybe 10 minutes looking at one review and NewEgg.
 
Last edited:
WOOT, big brown truck of awesome just delivered my Sapphire 6870. Now I can finish my latest build. Specs in sig.

I upgraded from one the first Phenom quad cores, a Foxconn mobo and a Sapphire 4870 512mb.
 
You must log in or register to reply here.
Back
Top