Sign in with
Sign up | Sign in
Your question

Which video card to get?

Last response: in Graphics & Displays
Share
March 27, 2010 3:16:51 AM

I am looking to upgrade from my 9800gt. I am displeased with the 9800gt since I can't seem to find a way to control the fan speed properly. Something internal to the card overrides RivaTuner and after 70*c it goes into auto mode and doesn't ramp up to 100% until 92*c. I could get another cooler for it but would rather just go with a better card.

My mobo is a Gigabyte EP45-UD3P and it supports X-fire. What would be the best bang for my buck in the $300 range? I am not stuck on Nvidia. I'm almost thinking ATI if I decide to go xfire in the future. Suggestions on cards in my price range?

Edit - I would also like a card that is fully configurable. No auto low level stuff where it stays in 2d mode and something I can accurately control the fans on without the card taking over.

Thanks,
Aaron

More about : video card

a c 130 U Graphics card
March 27, 2010 3:21:43 AM

2x HD5770 for $300 is literally the best for your money at the moment.
Related resources
March 27, 2010 3:34:10 AM

So just buy two of those and run them Xfire and I should be a lot better off than my crappy 9800gt? What about fan control and stuff? Does the ATI software manage that?

Oh and I run 1680x1050 res

EDIT - and this will be running on W7 64

Thanks a lot for all of the help
a c 130 U Graphics card
March 27, 2010 3:39:00 AM

Yes, CCC (Catalyst control center) manages everything. There are additional programs that can assist you in overclocking/voltage/fan control, though CCC does the basics (OC/Fan).

One of those already offers more than 35% increase in performance, and since they're some of the better-scaling cards, two'd offer much more.
March 27, 2010 3:41:15 AM

shadow187 said:
Yes, CCC (Catalyst control center) manages everything. There are additional programs that can assist you in overclocking/voltage/fan control, though CCC does the basics (OC/Fan).

One of those already offers more than 35% increase in performance, and since they're some of the better-scaling cards, two'd offer much more.



And W7 64 compatibility is there?

Is Furmark a good benchmark program to use so I can see the b4 and after?
a c 130 U Graphics card
March 27, 2010 3:43:52 AM

Yes, ATI is only bad with Linux.

Furmark is the ultimate stressing system. It's like trying to test how fast a hamster can run by torturing it until it can't run no more. It stresses your video card like nothing else can, and is only used to discover max GPU temps at load, and max GPU power draw.

Use something like 3DMark Vantage.
March 27, 2010 3:45:08 AM

shadow187 said:
Yes, ATI is only bad with Linux.

Furmark is the ultimate stressing system. It's like trying to test how fast a hamster can run by torturing it until it can't run no more. It stresses your video card like nothing else can, and is only used to discover max GPU temps at load, and max GPU power draw.

Use something like 3DMark Vantage.



Ok. I will get 3DMark Vantage and those two cards and see what happens. Thanks again for the help!
March 27, 2010 4:22:25 AM

Here is what I got with RivaTuner off and no OC.

March 27, 2010 4:27:51 AM

The XFX cards come with a double lifetime warranty. Whatever "double lifetime" means........just something to think about.....
a c 130 U Graphics card
March 27, 2010 12:16:45 PM

I'm not exactly sure how 3dMark works. I know it's a great tool for benchmarks, that's about it, :) .

As for JohnieReb; DoubleLifeTimeWarranty simply means that at any point in your life you can return the card for a replacement. Now if you sell the card, the person you sold it to will also have warranty, which is a very good deal.
a c 376 U Graphics card
March 27, 2010 2:40:57 PM

Overclocking an HD5850 is a better call than 2 HD5770s. If you bump up the voltage the card can OC an insane amount, usually over 40%. It also has a full 256-bit memory bus while the 128-bit bus of the HD5770s tends to be an issue when you put that card in crossfire, affecting minimum frame rates which is important for smooth game play. With the HD5850 it will also leave the other slot free as an upgrade path and you will have no crossfire scaling/compatibility issues. It will also use less power and give off less heat plus your case will have better airflow with just one card.
That said for your resolution a single HD5770 should be quite good so you may want to just get one and perhaps add another at some point in the future. If you are going to buy 2 upfront though the HD5850 is a better choice.
March 27, 2010 3:52:59 PM

I think I may wait for the 5850 2 gb cards to come out
a c 376 U Graphics card
March 27, 2010 4:13:13 PM

Especially at your resolution I don't think the extra gig will make much of a difference really, if any.
a c 235 U Graphics card
March 27, 2010 4:39:03 PM

JohnieReb said:
The XFX cards come with a double lifetime warranty. Whatever "double lifetime" means........just something to think about.....


means you can sell the card used and the buyer gets a lifetime warranty
March 27, 2010 5:11:38 PM

jyjjy said:
Especially at your resolution I don't think the extra gig will make much of a difference really, if any.


Don't think ti wil make any diffrence at any res, if the 4890 2gb is an indicator.
a c 376 U Graphics card
March 27, 2010 7:25:53 PM

Yeah, maybe once you crossfire and move up to 2560x1600 it'll make some difference. Or perhaps with eyefinity.
March 27, 2010 8:11:19 PM

It doesn't help with 2560x1600.

It might help with Eyefinity, but even that's not a given.

Honeslty, for non-eyefinity, more than 1gb per GPU is just marketing bullcrap.
a c 130 U Graphics card
March 27, 2010 8:28:29 PM

Oh really? Care to look at the GTX480 vs HD5870 benchmarks builder?
a c 376 U Graphics card
March 27, 2010 8:44:58 PM

What does the GTX 480 have to do with anything?
a c 130 U Graphics card
March 27, 2010 8:48:39 PM

Just showing him that extra memory helps.
a c 376 U Graphics card
March 27, 2010 8:49:47 PM

Well comparing two different cards with entirely different architectures certainly wont accomplish that in any meaningful way.
March 28, 2010 3:15:01 AM

I've ordered the dual HD 5770's. Will see how they work out.
March 28, 2010 2:21:27 PM

shadow187 said:
Just showing him that extra memory helps.


If there's one thing Honors Chem has taught me, It's that the best experiments have multiple variables.

/sarcasm
March 28, 2010 2:30:14 PM

builderbobftw said:
If there's one thing Honors Chem has taught me, It's that the best experiments have multiple variables.

/sarcasm



I have 4gb of Corsair memory. Seems to be ok. I don't use more than 50% of that at any time.
March 28, 2010 3:17:34 PM

Where talking about Graphics memory lol.
March 28, 2010 3:20:28 PM

builderbobftw said:
Where talking about Graphics memory lol.


Not sure what GPU memory I use with what I have at any given time. 1gb of mem currently. What can I use to monitor that?
March 28, 2010 3:22:17 PM

Huh, Prehaps GPUZ?

I'm not sure such a utilty exists, but if it does, GPU-Z is the one.
March 28, 2010 3:25:49 PM

Should I monitor with a benchmark or under normal gaming?
March 28, 2010 3:27:13 PM

Try a demanding game at max AA and Res.

( Try low settings, Highest possible res, 24AA, 16AF in Crysis or something like that)

That would show you the absoulte worst case senario for you're memory.
March 28, 2010 3:27:43 PM

I ran a furmark fire test and memory never went over 170
March 28, 2010 3:52:03 PM

22fps and 650 mb in Wow in Dalaran running around on Ultra.

Sys CPU at 50% and sys mem at 50%
a c 235 U Graphics card
March 28, 2010 3:55:20 PM

Dalaran FPS changes with the server in game population so its not a good FPS benchmark

what exactly are you trying to find out?

March 28, 2010 4:01:30 PM

ct1615 said:
Dalaran FPS changes with the server in game population so its not a good FPS benchmark

what exactly are you trying to find out?


Well the original point to this post was to find a card that I could control properly. The 9800gt I have has a mind of its own. Fan speeds and OC can't be set in stone. It does what it wants to.

I decided on going dual HD5770's. Just curious now how to compare what I have now to the new ones I should have next week. I have 3dmark tests saved and may run a few more to get a good baseline before the new cards get here.
a c 235 U Graphics card
March 28, 2010 6:11:16 PM

freakaccident said:
I decided on going dual HD5770's. Just curious now how to compare what I have now to the new ones I should have next week. I have 3dmark tests saved and may run a few more to get a good baseline before the new cards get here.


1) download Fraps
2) get a game(s) that does not rely on a broadband connection, single player FPS / RTS (Far Cry2 works great)
3) set game levels to a taxing setting (less then 20FPS)
40 set Fraps to capture min/max/avg for 120 seconds, run a constant benchmark 3x in a game

once you install the 5770s do the same thing, then compare

3dmark is really nothing more then a pissing contest, ATI/Nvidia constantly improve their drivers to up 3Dmark scores but if you are a gamer then you want to know "how well will it run my game" and not every card will run a game the same way.

Your 9800GT can beat a 4850 in some games that over whelming prefer Nvidia cards even if the 4850 has a much higher 3dmark score. Of course, the 4850 is the better all around card.
March 28, 2010 6:34:59 PM

Yep, the 4850 is comparible to the 9800GTX+/GTS 250.
March 30, 2010 12:06:55 AM

Thank you so much for all of the help.

I will try Fraps with Crysis.

I'm running a 60gb SSD so space is at a minimum but I think can jam Crysis on it.
March 30, 2010 12:32:19 AM

Why put Crysis on it?

Won't change FPS, just load times.
March 30, 2010 1:55:52 AM

builderbobftw said:
Why put Crysis on it?

Won't change FPS, just load times.


I'm not sure I understand what you mean. I ran Crysis past the stupid 45,000 minute intro then saved at a certain point then I chose a 120 second Fraps benchmark to run during game play from that save point. 12fps average.

With the new cards use the same save to start and see what I get. Why won't this work?

Load time? My pc boots in under 20 seconds after the bios is done bsing around.
a c 235 U Graphics card
March 30, 2010 2:07:57 AM

freakaccident said:
I'm not sure I understand what you mean. I ran Crysis past the stupid 45,000 minute intro then saved at a certain point then I chose a 120 second Fraps benchmark to run during game play from that save point. 12fps average.

With the new cards use the same save to start and see what I get. Why won't this work?



it will work, not sure why he asked that.

FYI Free Crysis benchmark tool

http://www.crymod.com/filebase.php?fileid=280&lim=0

* i think he meant why did you put Crysis in the SSD since it wont increase game FPS, just load times
March 30, 2010 2:13:18 AM

ct1615 said:
it will work, not sure why he asked that.

FYI Free Crysis benchmark tool

http://www.crymod.com/filebase.php?fileid=280&lim=0

* i think he meant why did you put Crysis in the SSD since it wont increase game FPS, just load times



Ah ok I must have misunderstood. I try to run all of my "games" loaded on the SSD for fast load times.

Thanks for the link to the Crysis benchmark tool!
March 30, 2010 2:40:26 AM

Well I can't really do much until I get the new cards. I can't get this POS 9800gt's fans to work properly. The Crysis benchmark got the card to 106c but the fan never went over 35%. If I use Riva to manually set it to 100% it drops right back down after it reaches 67c.
a b U Graphics card
March 30, 2010 2:42:21 AM

freakaccident said:
Well I can't really do much until I get the new cards. I can't get this POS 9800gt's fans to work properly. The Crysis benchmark got the card to 106c but the fan never went over 35%. If I use Riva to manually set it to 100% it drops right back down after it reaches 67c.


You might have bad drivers. I heard something about one of the recent Nvidia drivers overheating GPUs due to low fan speeds.
March 30, 2010 2:54:30 AM

Bluescreendeath said:
You might have bad drivers. I heard something about one of the recent Nvidia drivers overheating GPUs due to low fan speeds.


I've tried downgrading drivers and using the newest ones with no luck. I've tried every trick that I could find on the net. I'm not the only one either. There are posts complaining of the same issues with this card on this forum. I think it's ridiculous that they purposely build in "safety" measures because the public is too "stupid" to be able to handle controlling the settings. Granted it is a 9800gt and it is a cheap card. $199 when I bought it so not cheap compared to the two I have coming in the mail but damn it is frustrating.

I pray that these two ATI cards will allow me to do what I want to. When I am playing a game I want the darn fans to run fast enough to keep the cards at a low temp. Not wait until they are overheated before speeding up the fans.

I am going to try to rig up the 9800 fan so that it is off the cards control. Ill post back in a few.

OH and my second card is an 8600gt and I have NONE of the same problems. Riva controls the card the way I want to AND they both use the same darn driver. Go figure. Maybe I should do some test on the 8600gt to see how much "worse" it really is.

Thanks for allowing me to vent.

Aaron
March 30, 2010 3:08:54 AM

Ok so I don't have any way to connect the gpu fan to my PS to make it run at 100% without splicing wires. I removed the stock fan and cleaned the poor cooling paste off and applied some Arctic Silver 5. Doubt it will make a diff but good for now I suppose. Should have the Gigabyte ATIs by Wednesday.

Thanks,
Aaron
!