HIS Radeon HD 6970 2 GB

[W1zzard]

wow i don't think i have ever been so wrong in my life my bad

no worries. we're all here to learn.

 

[bear jesus]

no worries. we're all here to learn.

Very true but the silly thing is i poked google to try and gather the dates i listed, darn internet i blame it for my knowledge and lack of it

I would hope that with more time openCL will become more used as my last two cards have been AMD/ATI so cuda has been unavailable to me for years now and i would love to be able to put my cards to some useful GPGPU processing.

 

[Wile E]

Why don't I don't see Need for Speed Shift on that list of Physx games lol..

Pardon the slight change of topic from Cuda to Physx. Maybe, I should not have included Cuda in my initial statements b/c Cuda drivers/code are written by Nvidia and not the software developers.

Because Shift does not use GPU Physx. It is just CPU Physx. It does not use the video card for running any of teh Physx calculations at all. Most Physx games are run on the CPU.

 

[HalfAHertz]

OpenGL is not popular anymore because the api is so cumbersome to write for. After the Kronos group took over it and started slapping "patches" left and right to fill the widening gap between it and Dx, things quickly went downhill. Game devs nowadays need to put out a game as quickly as possible and because writing for Dx is quicker and simpler, most people go that way.
OpenCL on the other hand has been kept pretty tight so far. But I think development has slowed down to a craw because it is trying to cover so many different interests (Nvidia, AMD, Intel) and because the big three will never come to an agreement.

 

[Magikherbs]

Because Shift does not use GPU Physx. It is just CPU Physx. It does not use the video card for running any of teh Physx calculations at all. Most Physx games are run on the CPU.

Then why does Need for Speed Shift install the Nvidia Physx software regardless of your config ? There is no option, it just installs it. heh..

I just remembered how the new Nvidia drivers seem to relieve your cpu of its physics duties. hehh.. How much does that affect things ?

 

[pantherx12]

Then why does Need for Speed Shift install the Nvidia Physx software regardless of your config ? There is no option, it just installs it. heh..

I just remembered how the new Nvidia drivers seem to relieve your cpu of its physics duties. hehh.. How much does that affect things ?

http://img.techpowerup.org/101225/Capture.jpg

It installs it so your cpu can run phsyx....

phsyx isn't something that just runs it needs to be installed XD

 

[wahdangun]

no worries. we're all here to learn.

btw wizz where is HD 6970 CF ?? ,

 

[Magikherbs]

It installs it so your cpu can run phsyx....

phsyx isn't something that just runs it needs to be installed XD

huh.. didn't know it could run Physx without a Nvidia gpu.

My cpu can run physics just fine lol. I was trying to say that EA Games seems to have some vested interests.

 

[pantherx12]

Loads of games run the phsyx software on the CPU actually, just hardly any use GPU phsyx.

Big list on nvidia website : ]

 

[DarthElvis]

i love porn too, but didnt install it and didnt benchmark it.

Well then your leaving out a benchmark that way more than 5% of users could use

Back on topic, 10% -15% more performance with a 580 for 30% more cost. Yeah, ATI failed alright

6970 prices have stabilized here to about the same price as the 570, which is pretty much it's equal.

 

[Super XP]

Well then your leaving out a benchmark that way more than 5% of users could use

Back on topic, 10% -15% more performance with a 580 for 30% more cost. Yeah, ATI failed alright

6970 prices have stabilized here to about the same price as the 570, which is pretty much it's equal.

Elvis fan I see, yes he was the KING, oh and still remains the KING
ATI's Radeon HD 6970 2GB GDDR5 Graphics Card currently holds the best Price/Performance here in Canada IMO. It all depends on the games you want to play. I also like the FACT AMD/ATI's HD 6800 and HD 6900 series scales AMAZINGLY in CrossfireX vs. previous Gen Graphics Cards.

 

[Wile E]

Elvis fan I see, yes he was the KING, oh and still remains the KING
ATI's Radeon HD 6970 2GB GDDR5 Graphics Card currently holds the best Price/Performance here in Canada IMO. It all depends on the games you want to play. I also like the FACT AMD/ATI's HD 6800 and HD 6900 series scales AMAZINGLY in CrossfireX vs. previous Gen Graphics Cards.

Where is everybody getting that they scale better in Crossfire than previous generations? I haven't seen this anywhere. It all scales about evenly, including SLI setups, when the cards are of similar capabilities singly.

 

[bear jesus]

Where is everybody getting that they scale better in Crossfire than previous generations? I haven't seen this anywhere. It all scales about evenly, including SLI setups, when the cards are of similar capabilities singly.

I think everyone is referring to the fact that the 68xx and 69xx cards scale better than past mid/high end ATI/AMD cards, if you take a look at the TPU reviews of the 68xx and 69xx cards you can see some games/res scale up to 100% where as past cards were often well below 80%.

Just as an example on cryisis at 2560x1600 the 5850 got 16.7fps and 5850 crossfire got 25.9fps, 6950 got 20.9fps and 6950 crossfire got 41.2fps

 

[Wile E]

I think everyone is referring to the fact that the 68xx and 69xx cards scale better than past mid/high end ATI/AMD cards, if you take a look at the TPU reviews of the 68xx and 69xx cards you can see some games/res scale up to 100% where as past cards were often well below 80%.

Just as an example on cryisis at 2560x1600 the 5850 got 16.7fps and 5850 crossfire got 25.9fps, 6950 got 20.9fps and 6950 crossfire got 41.2fps

One test is hardly a conclusive result, however.

At 2560x1600, if I did my math right, overall scaling is at 69% for 6950 Crossfire, it's listed at 68% for 5870 Crossfire. I'd call 1% within the margin for error, or even attributable to driver update since the Crossfire 5870 test (run on a beta between Cat 9.10 and 9.11).

 

[bear jesus]

One test is hardly a conclusive result, however.

At 2560x1600, if I did my math right, overall scaling is at 69% for 6850 Crossfire, it's listed at 68% for 5870 Crossfire. I'd call 1% within the margin for error, or even attributable to driver update since the Crossfire 5870 test (run on a beta between Cat 9.10 and 9.11).

I was just picking a result that showed what i was talking about i know it alone proves nothing.

With the 6870/50 reviews there was a very long winded chat in the comments about how the overall scaling results were much lower than they really were according to the numbers in the results as most people were coming up with over 80% scaling and removing a couple games that had negative scaling/no scaling at all it went into 90%.

I should probably take a closer look at the 6950 crossfire review though as i admit i only skimmed through it

 

[Wile E]

Yeah, I meant 6950. fixed.

And going by the tpu review, the 6800 cards still scale at around 60% at 2560x1600.

 

[bear jesus]

Yeah, I meant 6950. fixed.

And going by the tpu review, the 6800 cards still scale at around 60% at 2560x1600.

Are you using the overall result at the end that w1zzard said was wrong or adding results together yourself?

I have been referring to the results that forum members got doing the math them self as they saw that the end result of overall scaling were wrong.

Such as this post here

I'm unsure if the issue has been fixed but this is what w1zzard had to say about them

 

[Wile E]

I'm going be relative performance charts @ 2560x1600. Didn't bother looking at 1920.

 

[Over50]

just a look at 68xx and 69xx

ATI/AMD making new cards to replace some old in the market and helping to reduce cost of nVidia GTX 460, GTX 470, GTX 480 and GTX 570 then possibly GTX 580.

the 68xx are more like the 57xx replacements.
While the 69xx are to compete with nVidia GTX 570 and GTX 580.

Still the ATI HD 5870 and HD 5970 are great gpu's.

Here we're seeing ATI /AMD using 2GB over the 1GB and 564MB also DDR5.
Not only that but ATI gpu's best configured for CrossfireX to get that Scaling ratio.

So depending on your system setup .. Do you prefer loading up PCI x16 /2.0 Slots with large ATI CrossfireX cards or a single nVidia card or 4 carded SLI.

The ATI 68xx series are throw backs or kick backs to mid-level performance and
the 69xx series are the new High End performers.

From what I have been reading from Reviews and various Forums only make sense that ATI /AMD is attempting to keep up with nVidia. Unfortunately, nVidia will always be one step ahead of ATI.

So trying is better than giving up ... we should appreciate ATI/AMD's efforts in keeping nVidia
GPU's at somewhat affordable prices. Otherwise we would be paying $500 to $1000 for those Top cards all the time. That's competition rather than one having a fully established monopoly. Thank you ATI !!!! can't wait till the GTX 580 is marked down some so I can buy it.

 

[bear jesus]

I'm going be relative performance charts @ 2560x1600. Didn't bother looking at 1920.

I suck at trying to make a point, i was really not paying attention to what i was linking to res wise but the results were wrong for every resolution including 2560x1600.

 

[Thatguy]

ATI/AMD making new cards to replace some old in the market and helping to reduce cost of nVidia GTX 460, GTX 470, GTX 480 and GTX 570 then possibly GTX 580.

the 68xx are more like the 57xx replacements.
While the 69xx are to compete with nVidia GTX 570 and GTX 580.

Still the ATI HD 5870 and HD 5970 are great gpu's.

Here we're seeing ATI /AMD using 2GB over the 1GB and 564MB also DDR5.
Not only that but ATI gpu's best configured for CrossfireX to get that Scaling ratio.

So depending on your system setup .. Do you prefer loading up PCI x16 /2.0 Slots with large ATI CrossfireX cards or a single nVidia card or 4 carded SLI.

The ATI 68xx series are throw backs or kick backs to mid-level performance and
the 69xx series are the new High End performers.

From what I have been reading from Reviews and various Forums only make sense that ATI /AMD is attempting to keep up with nVidia. Unfortunately, nVidia will always be one step ahead of ATI.
So trying is better than giving up ... we should appreciate ATI/AMD's efforts in keeping nVidia
GPU's at somewhat affordable prices. Otherwise we would be paying $500 to $1000 for those Top cards all the time. That's competition rather than one having a fully established monopoly. Thank you ATI !!!! can't wait till the GTX 580 is marked down some so I can buy it.

What kind of bullshit is that ? right and the 5xxxx cards went unanswered for how long ?

oh thats right, they didn't.

the 6xxxx cards were 2 different releases, TSMC screwed them on a node and the 68xx cards were refreshs and the 69xx cards were a pretty significant architectural change.

I mean its such a shame the the 69xx cards kick ass at high resolution. I mean what nvidia gonna do when they drop the dual gpu 6990 for the 580 price ?

BTW 580 is a nice card, but no reason to spout bullshit in a forum.

if anything the market is neck and neck with one always edging out the other, this round is a win for AMD/ATI on per/perf/watt and they will again have the single GPU crown with the 6990, nvidia can't compete with crossfire 6970's.

its such a shame that ATI/AMD just can't compete. tragic really.

 

[Wile E]

I suck at trying to make a point, i was really not paying attention to what i was linking to res wise but the results were wrong for every resolution including 2560x1600.

I calculated using this chart:

 

[bear jesus]

I calculated using this chart:

http://tpucdn.com/reviews/ASUS/Radeon_HD_6950_CrossFire/images/perfrel_2560.gif

My only problem is i do not know if the script has be fixed, i would hope so by now but once again without doing the math it seams lower than i would expect due to the following results at 2560x1600

game |single| dual
avp |22 |44
bc2 |37 |76
bf2 |27 |56
cod4 |80 |151
crysis |20 |41
dawn of war| 88| 170
far cry 2| 55 |111
hawx |85 |169
metro| 14| 25
riddick| 39|75
stalker| 26| 51

I know I'm missing a bunch of results but with so many around 100% 41% seams a bit low, to be honest though I'm too lazy to do the math so could we just leave it at most people are looking at the fact many games now scale to 100% where as they were lower with past AMD/ATI cards.

Weather the relative performance at the end is right or not a lot of people mainly look at the large amount of perfect scaling games and thus say they scale much better.

I feel we are going a little off topic here though so i hope i have explained myself well enough so i can be a quiet bear for now

 

[Wile E]

It's not 41% scaling, a single is 41% slower than Crossfire, but that is approximately 70% scaling for dual cards compared to single. 41/59 = .6949 and change.

It's within a few percentage points of that with all cards of recent generations.

 

[bear jesus]

It's not 41% scaling, a single is 41% slower than Crossfire, but that is approximately 70% scaling for dual cards compared to single. 41/59 = .6949 and change.

It's within a few percentage points of that with all cards of recent generations.

I must admit i had never thought of the graphs like that, but that now makes me wonder if everyone else was wrong when they were working out the scaling on the 68xx card's reviews although that confuses me as to why w1zzard agreed they were wrong..... i think i should be quiet as I'm just confused now