Quantcast

Jump to content

» «
Photo

Crossfire ATI 6870

38 replies to this topic
Hyp3rion
  • Hyp3rion

    Player Hater

  • Members
  • Joined: 24 Oct 2010

#1

Posted 24 October 2010 - 11:09 AM

Hello,

I am pretty new to GTA IV and my machine, Intel Q9650 | 4gig RAM | GTX460 had ok frame rates but nothing to right home about. I wanted to go SLI but as my board is xfire only I decided to buy 2 of the new 6870 cards. I have run the Unigine Heaven benchmark that is used quire commonly and with full AA, AF and normal tessalation I get and average of 53 fps, this is higher by almost 15fps on one review I saw of the GTX480.
So, why is the performance in GTA 4 no better than my GTX460? I do not even have frull draw distance and detail set above default and yet I only get around 43fps during the benchmark test. Anyone give some usefull pointers?

Regards,

Hyp

luceberg
  • luceberg

    Homeboy

  • Members
  • Joined: 27 Dec 2008

#2

Posted 24 October 2010 - 11:21 AM

Unless the cpu is overclocked, the performance in this game is cpu limited.

No mention of operating system, but Windows 7 64 bit would work best for that system.

Hyp3rion
  • Hyp3rion

    Player Hater

  • Members
  • Joined: 24 Oct 2010

#3

Posted 24 October 2010 - 11:27 AM

Hi, thanks for the reply.

No, my CPU is not overclocked at the moment, it is on it's stock 3ghz although I have had it clocked to 3.8ghz in the past so I will give this a try and see what effect it has on the benchmark test. I am on Win7 x64.


Hyp3rion
  • Hyp3rion

    Player Hater

  • Members
  • Joined: 24 Oct 2010

#4

Posted 24 October 2010 - 12:23 PM

So I have overclocked my q9650 from 3ghz to 3.8 and I am sure it has made some difference. Shame it is not totally smooth.

crackdawg
  • crackdawg

    supreme ruler

  • BUSTED!
  • Joined: 28 Nov 2007

#5

Posted 24 October 2010 - 01:23 PM

Because RAGE engine has it's own rendering pipeline, and only uses texture and framebuffer features of your card and subsystem. This is why I say buy by TFR and VRAM, not what is new or model numbers, or even clocks like all the experts here like to repeatedly suggest.

Sorry to break it to you, but that big 460 means little, my 8800GT is almost as fast as your 460. Buy according to TFR benchmarks and VRAM and you wont waste money, like I say in all my posts, that best buy grade geeks try to criticize but fall way short.

Also look in old posts here, SLI and crossfire are also useless. It doesent even use whole x2 cards.

yojo2
  • yojo2

    ~y

  • Andolini Mafia Family
  • Joined: 06 Dec 2008
  • Poland
  • Most Knowledgeable [Technology] 2013
    Helpfulness Award
    Helpfulness Award [Technology]

#6

Posted 24 October 2010 - 01:40 PM

QUOTE (crackdawg @ Oct 24 2010, 15:23)
Sorry to break it to you, but that big 460 means little, my 8800GT is almost as fast as your 460.

Yeah yeah whatever.
http://www.techpower..._768_MB/31.html

As you can see GTX460 768MB is almost twice as fast as 9800GT (which is more or less equal to 8800GT).

IronHide-AW
  • IronHide-AW

    Long Live AC/DC

  • Members
  • Joined: 22 Jul 2010

#7

Posted 24 October 2010 - 01:57 PM

@Hyp3rion

As said here, and regularly across the entire board, GTA-IV is very CPU intense. Out of curiosity check and post the CPU percent usage reported during the game benchmark at stock speed and also OC speed.

@crackdawg

You have described those details of the RAGE engine so many times you should just create a concise engine details post with all that and then have a mod sticky the sucker smile.gif

It's extremely useful to know generally anyway since it is so specific for this game. Surprised it's not already a sticky somewhere.

Hyp3rion
  • Hyp3rion

    Player Hater

  • Members
  • Joined: 24 Oct 2010

#8

Posted 24 October 2010 - 02:09 PM

@Crackdawg, thanks for the info but to buy a card based on one games performance requirements is nuts. I play many games where a 8800gt would get spanked. As long as I can optimize what I have to play GTA4 as goos as possible that is fine. All my other games perform exceptionally.

The CPU usage now that I have clocked to 3.8ghz is 68%

IronHide-AW
  • IronHide-AW

    Long Live AC/DC

  • Members
  • Joined: 22 Jul 2010

#9

Posted 24 October 2010 - 03:09 PM

He can clear it up himself, but I think crackdawgs only point is if someone is specifically and intentionally trying to boost the GPU performance of GTA-IV, in a very cost effective way, but not trying to get the latest hardware for other games. In that scenario, it would be possible to get much cheaper board that has hardware specific to what GTA-IV uses most and therefore produce the most bang for the buck for this game. At-least, that's my take on what he's been saying.

crackdawg
  • crackdawg

    supreme ruler

  • BUSTED!
  • Joined: 28 Nov 2007

#10

Posted 25 October 2010 - 08:02 AM Edited by crackdawg, 25 October 2010 - 08:04 AM.

QUOTE (IronHide-AW @ Oct 24 2010, 13:57)
@Hyp3rion

As said here, and regularly across the entire board, GTA-IV is very CPU intense. Out of curiosity check and post the CPU percent usage reported during the game benchmark at stock speed and also OC speed.

@crackdawg

You have described those details of the RAGE engine so many times you should just create a concise engine details post with all that and then have a mod sticky the sucker smile.gif

It's extremely useful to know generally anyway since it is so specific for this game. Surprised it's not already a sticky somewhere.

lol I barely get by without being banned, even though 99% of my input here is technically sound contribution. I don't see a sticky happening. Especially when I show the brown nosers with badges how they don't know what they are talking about on a regular bases.

I just remind people that texture fill rate is all that matters and hope they spread the word. RAGE handles most of it's own rendering on CPU along with all the AI and Euphoria scripts and input.

@yojo2:
9800GT and 8800GT TFR = 33.4
460 GTX = 37.8

That's not much of a difference as I pointed out in my post. 460 has a 2GB configuration and 8800 has only 1GB. 460 even produces more heat and consumes more energy despite being 40nm.

Like I've said in other threads if you're going to question my info you should have the data to back yours up, cause I always do for mine. I also don't like department store lurkers or geek squad level geeks trying to correct me.

yojo2
  • yojo2

    ~y

  • Andolini Mafia Family
  • Joined: 06 Dec 2008
  • Poland
  • Most Knowledgeable [Technology] 2013
    Helpfulness Award
    Helpfulness Award [Technology]

#11

Posted 25 October 2010 - 08:06 AM

QUOTE (crackdawg @ Oct 25 2010, 10:02)
Like I've said in other threads if you're going to question my info you should have the data to back yours up, cause I always do.

Well I provided you with benchmark results clearly showing that 9800GT (which is more or less like 8800GT) is far behind from GTX460 in terms of performance. If that's not data you find true then I don't know what is.

crackdawg
  • crackdawg

    supreme ruler

  • BUSTED!
  • Joined: 28 Nov 2007

#12

Posted 25 October 2010 - 08:19 AM

QUOTE (yojo2 @ Oct 25 2010, 08:06)
QUOTE (crackdawg @ Oct 25 2010, 10:02)
Like I've said in other threads if you're going to question my info you should have the data to back yours up, cause I always do.

Well I provided you with benchmark results clearly showing that 9800GT (which is more or less like 8800GT) is far behind from GTX460 in terms of performance. If that's not data you find true then I don't know what is.

That's testing computing power in general, RAGE doesn't use anything but texture instructions and frame buffers. It even uses custom shaders to work with it's lighting system. This is why, once again, TFR is all that matters. Show me any TFR benchmark that proves your side of the argument.

I was going to get youtube videos to show at x1200 resolution, but they are all mismatch configurations and the people are trying to use FSAA and stuff.

My original setup which was 5400+ and a 512MB 8800GT even got good frames with medium and high settings, while driving in MP. This is where cards really get tested in GTA because of TFR.

32+ TFR and 1.3GB+ VRAM seems the be the perfect card formula. Clocks and general SP and core benchmarks don't even reflect in GTA.

yojo2
  • yojo2

    ~y

  • Andolini Mafia Family
  • Joined: 06 Dec 2008
  • Poland
  • Most Knowledgeable [Technology] 2013
    Helpfulness Award
    Helpfulness Award [Technology]

#13

Posted 25 October 2010 - 04:45 PM

http://www.pcgamesha...eviews/?page=11

So, how will you explain that? Here we can see that GTX470 (which, mind you, has actually TFR = 34 - even lower than GTX460) is more than twice as fast as 8800GTX (which has TFR = 36.8 - higher than 8800GT).

crackdawg
  • crackdawg

    supreme ruler

  • BUSTED!
  • Joined: 28 Nov 2007

#14

Posted 25 October 2010 - 05:11 PM Edited by crackdawg, 25 October 2010 - 05:16 PM.

At 1680x1050 8800 gets 30.6 FPS avg and 480 72.7. 480 is also 4.2 TFR over the 460. I'm not sure how my posts are inconsistent? 8800GT and the 460 aren't even there, but the data still validates my statements.

yojo2
  • yojo2

    ~y

  • Andolini Mafia Family
  • Joined: 06 Dec 2008
  • Poland
  • Most Knowledgeable [Technology] 2013
    Helpfulness Award
    Helpfulness Award [Technology]

#15

Posted 25 October 2010 - 05:32 PM

8800GTX has higher TFR than GTX470, but performs noticeably worse - that's what I said. What's in your opinion the reason behind this?

_corTEC
  • _corTEC

    Snitch

  • Awaiting Authorisation
  • Joined: 01 Jun 2010

#16

Posted 25 October 2010 - 09:13 PM

Jeez, TFR! Wow! My TFR is 37.5, when I clock the card it's 43.5. Woo-f**kin'-hoo.

I still want an EVGA GTX460 SuperClocked. And yes, 8800GT was a ground-breaking card, but that was four years ago. GET OVER IT.

Also look for these when buying GPUs - memory bandwidth, CUDA cores and ROPs. Google 'em if you don't know all about them, cos they are making or breaking today's GPUs.

http://rapidshare.co...134278/CARD.gif

Stinky12
  • Stinky12

    No title

  • Members
  • Joined: 14 Oct 2010

#17

Posted 25 October 2010 - 09:29 PM

8800GTX and GTX470 are from different generations, where as the GTX 470 architecture is much superior than
8800GTX no matter if it has a higher texture fill rate.


Warlord.
  • Warlord.

    :O

  • Leone Family Mafia
  • Joined: 03 Apr 2004

#18

Posted 25 October 2010 - 09:37 PM

QUOTE (yojo2 @ Oct 25 2010, 23:32)
8800GTX has higher TFR than GTX470, but performs noticeably worse - that's what I said. What's in your opinion the reason behind this?

Explain this please. crackdawg.

Jigglyass
  • Jigglyass

    Homie

  • Members
  • Joined: 18 Dec 2008
  • None

#19

Posted 26 October 2010 - 02:17 AM

This game loves shaders, it will anally rape them. Only reason fermi beats the 5870 is because it has more unified shaders which run faster.

Chevyboy
  • Chevyboy

    Playa

  • Members
  • Joined: 02 Dec 2008

#20

Posted 26 October 2010 - 03:27 AM

QUOTE (Warlord. @ Oct 25 2010, 21:37)
QUOTE (yojo2 @ Oct 25 2010, 23:32)
8800GTX has higher TFR than GTX470, but performs noticeably worse - that's what I said. What's in your opinion the reason behind this?

Explain this please. crackdawg.

He's been proven wrong thus decided to crawl back under that rock of his. I learned not to argue or provoke him a long time ago as all he wants to do is argue illogical topics then disappear when proven wrong.

Long story short, I wouldn't expect to see him in this thread again.

crackdawg
  • crackdawg

    supreme ruler

  • BUSTED!
  • Joined: 28 Nov 2007

#21

Posted 26 October 2010 - 06:44 AM

QUOTE (Chevyboy @ Oct 26 2010, 03:27)
QUOTE (Warlord. @ Oct 25 2010, 21:37)
QUOTE (yojo2 @ Oct 25 2010, 23:32)
8800GTX has higher TFR than GTX470, but performs noticeably worse - that's what I said. What's in your opinion the reason behind this?

Explain this please. crackdawg.

He's been proven wrong thus decided to crawl back under that rock of his. I learned not to argue or provoke him a long time ago as all he wants to do is argue illogical topics then disappear when proven wrong.

Long story short, I wouldn't expect to see him in this thread again.

When have I ever hidden? You don't even know me kiddy, go enlighten someone else.

"my 8800GT is almost as fast as your 460" - crackdawg

Something that also discredits that sites benchmark:
470 GTX "March 26, 2010"
8800 GTX "November 8, 2006"
"Geforce 197.13/17(Q)"

This is my last post in this thread. too many noobs hording..

_corTEC
  • _corTEC

    Snitch

  • Awaiting Authorisation
  • Joined: 01 Jun 2010

#22

Posted 26 October 2010 - 06:59 AM

Could we at least agree that AMD/ATi, or whatever the hell they call themselves, are inferior in every way to nVidia?

GTA3Freak-2001
  • GTA3Freak-2001

    Run out of Cake.

  • Members
  • Joined: 25 Nov 2001

#23

Posted 26 October 2010 - 07:30 AM

So now we have fanboyism to go with our idiotic comments! dozingoff.gif

crackdawg
  • crackdawg

    supreme ruler

  • BUSTED!
  • Joined: 28 Nov 2007

#24

Posted 26 October 2010 - 10:18 AM Edited by crackdawg, 26 October 2010 - 10:29 AM.

lol I actually support any console over PC because it's cheaper and more efficient. Nvidia has always had more stable drivers. ATI just has better performance for the money. It's really a waste of money in any case unless you use it for other things than games, which most of these people obviously don't..

I've always thought fanboys were just competitive consumers..I don't own or plan to buy most of this stuff, just have access to it because of work. I do own a 8800GT from a couple years ago and some CPUs. At the end of the day the TFR&VRAM formula still holds, even the data posted here to attempt to discredit me to the rest of the herd reflects it despite inconsistencies.

Warlord.
  • Warlord.

    :O

  • Leone Family Mafia
  • Joined: 03 Apr 2004

#25

Posted 26 October 2010 - 01:51 PM

QUOTE (crackdawg @ Oct 26 2010, 16:18)
lol I actually support any console over PC because it's cheaper and more efficient. Nvidia has always had more stable drivers. ATI just has better performance for the money. It's really a waste of money in any case unless you use it for other things than games, which most of these people obviously don't..

I've always thought fanboys were just competitive consumers..I don't own or plan to buy most of this stuff, just have access to it because of work. I do own a 8800GT from a couple years ago and some CPUs. At the end of the day the TFR&VRAM formula still holds, even the data posted here to attempt to discredit me to the rest of the herd reflects it despite inconsistencies.

Ah so the site's not credible. That explains it.

Stinky12
  • Stinky12

    No title

  • Members
  • Joined: 14 Oct 2010

#26

Posted 26 October 2010 - 02:00 PM

So what crackdawg is saying TFR and VRAM is more important than any other aspects of the card?
If that is the case, assuming here are the cards

card A: TFR: 56.8 and 2GB of VRAM 64bit
card B: TFR: 30.5 and 1GB of VRAM 128bit

What card will be the best?

OverTheBelow
  • OverTheBelow

    OTB "Oaty-Bee"

  • Zaibatsu
  • Joined: 30 Oct 2008

#27

Posted 26 October 2010 - 02:05 PM

*hides behind his 62.6 TFR*.

Chevyboy
  • Chevyboy

    Playa

  • Members
  • Joined: 02 Dec 2008

#28

Posted 26 October 2010 - 02:19 PM

QUOTE (crackdawg @ Oct 26 2010, 06:44)
This is my last post in this thread. too many noobs hording..

And yet you continue to post rolleyes.gif

You are the only one rambling about who knows what now, have been proven wrong several times with various links then either proceed to ignore them completely and continue your senseless rambling or claim the sites are not credible.

I only see one "noob" in this thread.

I guess you can't respond to this since you aren't posting in this thread anymore sad.gif

mkey82
  • mkey82

    Keep riding hard, son

  • Members
  • Joined: 14 Feb 2008

#29

Posted 26 October 2010 - 02:36 PM

QUOTE (_corTEC @ Oct 26 2010, 08:59)
Could we at least agree that AMD/ATi, or whatever the hell they call themselves, are inferior in every way to nVidia?

Shall we make a poll?

a) ATI/AMD
* bad drivers and lousy support
* rather respectable price/performance ratios, albeit depending on the series, 5000 series sucked in this regard (IMO)

b) Nvidia
* they manufacture fake cards and show them off as the real stuff
* they cripple their omnipotent proprietary technologies in software (to make it run really bad or not run at all)
* they bribe game devs to f*ck up games for ATI cards
* despite ATI has lousy drivers, Nvidia is the one who managed to fry cards with a botched driver release
* oh yes, the all mighty rebrands, I'm a big fan of those

+1 for the red team, even though if they start to rebrand my next card is going to be a Larabee (fat chance) or I'm gonna put 15 GMA's in SLI/CF biggrin.gif

yojo2
  • yojo2

    ~y

  • Andolini Mafia Family
  • Joined: 06 Dec 2008
  • Poland
  • Most Knowledgeable [Technology] 2013
    Helpfulness Award
    Helpfulness Award [Technology]

#30

Posted 26 October 2010 - 03:14 PM Edited by yojo2, 26 October 2010 - 03:18 PM.

QUOTE (crackdawg @ Oct 26 2010, 08:44)
Something that also discredits that sites benchmark:
470 GTX "March 26, 2010"
8800 GTX "November 8, 2006"
"Geforce 197.13/17(Q)"
Erm... and how is that discrediting this benchmark?
QUOTE (_corTEC @ Oct 26 2010, 08:59)
Could we at least agree that AMD/ATi, or whatever the hell they call themselves, are inferior in every way to nVidia?
Both ATI/AMD and nVidia card have their pros and cons.




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users