Sign in with
Sign up | Sign in
Your question
Closed

Crossfire - Why isn't my second card doing much?

Last response: in Graphics & Displays
Share
October 13, 2011 2:13:33 PM

I've noticed that my second 4870 only gets utilised from about 8-14% when gaming where as the first one gets utilised 100% (according to GPU-Z). This tells me that its not really being put to use and improving things. I am having to turn settings down in games when i could probably be running them a lot higher if the second card was doing more. I have crossfire x enabled and catalyst ai set to standard. im running an asus crosshair formula iv mobo which runs both cards at 16x, i also have a corsair 1000w psu. crossfire mode is certainly active as i am able to play battlefield bad company 2 on maximum settings at a big res very smoothly and i know i couldnt do that before, however i just dont think its being fully utilised from the gpu-z readings. i know some games dont support crossfire, but at the moment im playing metro 2033 which does. i can run it at high settings with 1920x1080 just fine, when i run it on ultra high it plays very smoothly about 70% of the time but it does get a bit choppy when in heavy combat. when ive ran it on ultra ive checked gpu-z readings after and my second cards max load was 14%. i cant figure out why this is happening. ive had the same thing with benchmarks that ive ran - second card not being utilised much. does anyone know why this is? thanks.

More about : crossfire card

October 13, 2011 2:19:45 PM

When you check your cards use, are you playing the game in windowed mode? In windowed mode, crossfire doesn't get used.
Score
0
October 13, 2011 2:49:44 PM

no im playing full screen at 1920x1080. cheers for that tip though, i didnt actually know that!
Score
0
Related resources
October 13, 2011 3:52:01 PM

ive also just noticed in gpu-z it says crossfire is enabled on one card but not the other. i dont know if thats normal...?
Score
0
October 13, 2011 4:07:22 PM

Your using a crossfire bridge right? Also if you are I had the same problems last week my crossfire bridge some how went bad you might want to check it
Score
0
October 13, 2011 4:11:51 PM

ET900 said:
ive also just noticed in gpu-z it says crossfire is enabled on one card but not the other. i dont know if thats normal...?


I see the same thing in GPU-z, which I didn't think was normal. But I ran a highly intesive game of SC2 in fullscreen mode, with the AMD monitoring utility on my other monitor. Both of my cards were being utilized at about 50-60% each.

As mentioned, your issue could be a bad bridge.
Score
0
October 13, 2011 4:13:03 PM

There was a post previously that had a similar problem....

Apparently the bridge was on the wrong spot on the video cards...something about 1a and 1b or something so only 1 card was being used, and only about 20% of the other card. Double check your Xfire bridge (make sure they are connected on the right spots).
Score
1
October 13, 2011 4:18:28 PM

thanks for the replies. i am using the bridge that came with my motherboard which i bought new and have had for a matter of weeks. catalyst control centre says crossfire x is enabled and i have definitely noticed a performance increase in bad company 2. i actually just tried using the bridge which came with my 4870 and it made no difference :/  crossfire does appear to be enabled, its like something isn't set/optimized right and so the second gpu is only partially being used.

without trying to sound patronising i have been gaming on pc for some years now and generally dont need to ask much in forums as far as hardware and software setups go. its just that this is my first ever crossfire rig so ive ran into problems ive never encountered before, its unfamiliar territory you could say...
Score
0
October 13, 2011 4:19:56 PM

Chainzsaw said:
There was a post previously that had a similar problem....

Apparently the bridge was on the wrong spot on the video cards...something about 1a and 1b or something so only 1 card was being used, and only about 20% of the other card. Double check your Xfire bridge (make sure they are connected on the right spots).


hey thats a good tip man, i didnt really consider that as i read a few crossfire tutorials that all said it doesn't matter which connectors you put the bridge on! i am going to give that a shot now and post back my findings!
Score
0
October 13, 2011 4:49:47 PM

well i tried putting the bridge on the other connectors and played metro 2033 with every setting as high as possible @ 1920x1080 res. the game went sketchy at times which obviously means the computer was struggling yet the second card never went above 11%. im really stumped with this! :/ 
Score
0
October 13, 2011 5:55:11 PM

ok another update if anyones able to make sense of any of this.... i found a third crossfire cable/bridge. all 3 of these cables were new and had never been used before... i tried all cables on both conectors with catalyst ai disabled, on standard and on advanced. nothing made any difference apart from the fact it felt like i was only using one card with catalyst ai disabled (ive also read that thats one of the things disabling it does).

so i think its fair to say that its extremely unlikely to be a problem with the bridge! the highest ive ever seen the max load on the second card was 14%. Now this is all based on GPU-Z readings. However, i just tried amd system monitor with metro 2033 and battlefield bad company 2. Whilst playing each game i 'alt tabbed' out suring the action and the graphs on amd system monitor appeared to show both cards working pretty hard - about 96%. Do you guys think that GPU-Z is giving false readings? You would've thought amd would know how to make a working monitoring tool for their hardware...right?
Score
0
October 13, 2011 5:59:02 PM

Looks like you may have found a bug. BTW i don't think GPU-Z is made by AMD.
Score
0
October 13, 2011 6:01:48 PM

ET900 said:
ok another update if anyones able to make sense of any of this.... i found a third crossfire cable/bridge. all 3 of these cables were new and had never been used before... i tried all cables on both conectors with catalyst ai disabled, on standard and on advanced. nothing made any difference apart from the fact it felt like i was only using one card with catalyst ai disabled (ive also read that thats one of the things disabling it does).

so i think its fair to say that its extremely unlikely to be a problem with the bridge! the highest ive ever seen the max load on the second card was 14%. Now this is all based on GPU-Z readings. However, i just tried amd system monitor with metro 2033 and battlefield bad company 2. Whilst playing each game i 'alt tabbed' out suring the action and the graphs on amd system monitor appeared to show both cards working pretty hard - about 96%. Do you guys think that GPU-Z is giving false readings? You would've thought amd would know how to make a working monitoring tool for their hardware...right?


I'd think AMD's is more likely accurate. What's the difference in FPS between crossfire and not?
Score
0
October 13, 2011 6:15:22 PM

@Chainzsaw - yeh i reckon its probably a bad reading with gpu-z. i was getting readings of 104.5c on the MEMIO reading on my gpu to which i think (and am hoping) may be false. btw i knew that gpu-z isn't made by amd, i just worded that sentence badly sorry :/ 

@bystander - yeah you would think so wouldn't you, im hoping so. i haven't tried metro with crossfire off but i certainly have with battlefield and on the highest settings its pretty choppy to play about 50% of the time, im guessing around 25-35fps tops. with crossfire on its almost constantly smooth, like 45-60fps, with the odd drop here and there. i think i mentioned it earlier, or maybe in a different thread that i was seeing a noticable performance increase with crossfire on, i just wasn't sure it was working anywhere near to its full potential because of the gpu-z readings and was wondering if i should be getting more out of it.

i think a good idea would be to test with another program. anyone know of something else i can monitor gpu usage with? thanks all :) 
Score
0
October 13, 2011 7:56:34 PM

+1 to Chainzsaw for posting about the connector bridge labels. Beat me to it!

This may be a dumb question but have you done a clean install of the 11.9's and the 11.9 CAP1's?

GPU-Z seems to be accurate, I wouldn't worry to much about that.

Also; check your BIOS for something along the lines of "Turbo USB mode" and turn that off and check your results.
Score
0
October 13, 2011 8:08:25 PM

COrehound@ I'm curious what effects the turbo usb mode does for gaming? i've seen it in my bios and never thought anything about it, would i get better performance from disabling it?
Score
0
October 13, 2011 8:24:06 PM

Turbo USB shares bandwith with one of your PCI-E 16x lanes effectively halving it. Usually this isn't an issue with a single GPU, but when running crossfire/sli it will make one PCI-E slot run at 4x instead of 8x. At least this is how I understand it, I could be wrong. Please correct if so.
Score
0
October 13, 2011 8:27:59 PM

for monitoring i use MSI afterburner.

It does not matter where the crossfire bridge is connected,as long as it is,I always ran both just for looks.
Score
0
October 13, 2011 8:29:15 PM

C0rehound said:
+1 to Chainzsaw for posting about the connector bridge labels. Beat me to it!

This may be a dumb question but have you done a clean install of the 11.9's and the 11.9 CAP1's?

GPU-Z seems to be accurate, I wouldn't worry to much about that.

Also; check your BIOS for something along the lines of "Turbo USB mode" and turn that off and check your results.


i think i did a clean install after installing the second card but tbh im not 100% on that. im gonna have to go do that now to rule it out!

if gpu-z is accurate then i have more to worry about because it means my crossfire setup isnt working properly and my gpu is getting way to hot :( 

i dont think i have that setting in the bios but the only thing like that is what appears to backwards compatability support for usb 2.0 - i have the option of high speed which is what set to by default and is currently on, and full speed which according to the bios is actually slower! :/  what affect will changing this setting do if its what you're on about? i didnt think usb settings would come into play here...
Score
0
October 13, 2011 8:30:22 PM

oh i just read your other reply. shall i try changing that setting to full speed mode instead of high speed?
Score
0
October 13, 2011 8:31:27 PM

Try full speed instead of high speed. That could be the same thing as Turbo USB. I see your motherboard supports USB3.0 and that could be how it manages that. Like I said, if it is sharing bandwidth with your PCI-E lane then it could be hogging power from your GPU.
Score
0
October 13, 2011 8:34:31 PM

Earnie said:
for monitoring i use MSI afterburner.

It does not matter where the crossfire bridge is connected,as long as it is,I always ran both just for looks.



yeah i have been using that to. its just that gpu-z is the only program ive used that shows the extra 3 temp readings. its mainly the memio one that im worried about and not sure what other program will display that temp...

C0rehound said:
Try full speed instead of high speed. That could be the same thing as Turbo USB. I see your motherboard supports USB3.0 and that could be how it manages that. Like I said, if it is sharing bandwidth with your PCI-E lane then it could be hogging power from your GPU.



ok thanks i will do. reinstalling the latest drivers and corssfire profiles to... will post back when ive tested again :) 
Score
0
October 13, 2011 9:17:35 PM

are both of your cards running x16?
Score
0
October 13, 2011 9:40:23 PM

Read your motherboard manual carefully and look for anything (like USB Turbo) that may affect the slot one of your graphics cards are in.

Benchmarking:
Find either a benchmark that gives you an average FPS (in-game benchmark) or else eye-ball the FPS in a short scene that you can repeat for comparison.

VSYNC needs to be off or it locks your framerate (desirable, just not for testing).

Supported games vary, but you should see several games a getting 65% boost (depends on cards, game, CPU etc). Some games don't support Crossfire at all, on the other hand the write game/card/etc combo can get close to 90% boost!

Example:
Batman AA benchmark:
1 card - 100FPS
XFire - 165FPS

(Of course you'd want VSYNC back ON again when playing).

Good luck.
Score
0
October 13, 2011 9:41:40 PM

OK, I remember a Tom's article that in part compared (the difficulties of) installing CF vs SLI. What vid card differences would actually work, card and driver installation sequence, whether the bridge was even required lol, etc. I just looked back through a year or so of graphics articles, and can't find it.

The upshot of my recollection is "here's what I would try":

1) Verify both vid cards are fully functional when used alone in either slot. This involves four tests (2 cards, 2 slots). if one of these tests (using sat Furmark) fails, either a slot or a card problem may have bene detected. If all is OK . . .

2) Wipe drivers. Install one card in primary slot, install drivers. Ensure its running. Install the second card, enable CF, see if the cards are now being used properly. If one card still underused . . .

3) Wipe drivers, install both cards at the same time (with bridge), and install drivers. Check if the problem is resolved. If not . . .

4) Check the BIOS version and timings of each card. If they are different, investigate whether you can update one of the BIOS's and/or get the timings the same. (Alternatively you can do this first if you wish - if BIOS and timings are the same there's nothing else to do in this step.)

5) What mobo and BIOS version are you using? Have you checked for any mobo switches or BIOS settings that would change the operation of your PVIe slots (eg, from X16-X4 to X8-X8.)
Score
0
October 13, 2011 9:44:15 PM

use a different benchmark
Score
0
October 13, 2011 10:13:07 PM

Ok so i have reinstalled the graphics drivers and crossfire profiles, reinstalled all motherboard drivers and updated the motherboard bios. i changed that usb setting also... i just did some more gameplay tests and have discovered a bug with gpu-z! you can select which gpu you want to see the stats for in the drop down box in gpu-z. when you come out of a game the only gpu that appears to have been in heavy use is whichever one you left selected before you loaded the game. so i have tried with one selected, then the other when gaming and both gpu's get high load percentages (though it'll only say one of them did - the one you left selected). so i guess its fairly safe to say that my crossfire is probably working ok?

odoyle_rules said:
are both of your cards running x16?


as far as i can tell yes. my motherboard does support 2 running at this speed and gpu-z says they are :) 



photonboy said:
Read your motherboard manual carefully and look for anything (like USB Turbo) that may affect the slot one of your graphics cards are in.

Benchmarking:
Find either a benchmark that gives you an average FPS (in-game benchmark) or else eye-ball the FPS in a short scene that you can repeat for comparison.

VSYNC needs to be off or it locks your framerate (desirable, just not for testing).

Supported games vary, but you should see several games a getting 65% boost (depends on cards, game, CPU etc). Some games don't support Crossfire at all, on the other hand the write game/card/etc combo can get close to 90% boost!

Example:
Batman AA benchmark:
1 card - 100FPS
XFire - 165FPS

(Of course you'd want VSYNC back ON again when playing).

Good luck.



thanks for the tips man :)  i am gonna have to run some more benchmarks to see whats up. if im getting like 50% boost on games that support corssfire then i'll just accept it as working :) 


Twoboxer said:
OK, I remember a Tom's article that in part compared (the difficulties of) installing CF vs SLI. What vid card differences would actually work, card and driver installation sequence, whether the bridge was even required lol, etc. I just looked back through a year or so of graphics articles, and can't find it.

The upshot of my recollection is "here's what I would try":

1) Verify both vid cards are fully functional when used alone in either slot. This involves four tests (2 cards, 2 slots). if one of these tests (using sat Furmark) fails, either a slot or a card problem may have bene detected. If all is OK . . .

2) Wipe drivers. Install one card in primary slot, install drivers. Ensure its running. Install the second card, enable CF, see if the cards are now being used properly. If one card still underused . . .

3) Wipe drivers, install both cards at the same time (with bridge), and install drivers. Check if the problem is resolved. If not . . .

4) Check the BIOS version and timings of each card. If they are different, investigate whether you can update one of the BIOS's and/or get the timings the same. (Alternatively you can do this first if you wish - if BIOS and timings are the same there's nothing else to do in this step.)

5) What mobo and BIOS version are you using? Have you checked for any mobo switches or BIOS settings that would change the operation of your PVIe slots (eg, from X16-X4 to X8-X8.)



haha :/  i have tried both cards individually with fresh driver installs and both work fine. i have tried them both ways round in crossfire to though im not sure i installed fresh drivers each time for that! one card has a slightly newer bios though but im scared to flash the bios on a gpu :/  i hear so much about it putting cards to bed lol...

im using an asus rog crosshair formula iv with the 2002 bios (was 1902 bios earlier). i do have buttons for all sorts of things to do with overclocking. i did read about them in the manual and chose what ones i wanted enabled. i haven't done anything in the way of overclocking though. i cant find any settings for pci-e in the bios so i cant tell on that level about the lane width in use. ive been relying on software for that - gpu-z. i'll dig around the bios again and see if ive missed any settings! appreciate the reply mate :) 


wishmaster12 said:
use a different benchmark



do you have any suggestions?
Score
0
October 13, 2011 10:53:01 PM

You probably know this, but the Crossfire update is a separate from the main Driver and CCC update.

It's called "ATI Catalyst Application Profiles" and is usually about 0.5MBytes. It's right below the main AMD driver and has the same number (latest is v11.9).
Score
0
October 13, 2011 11:16:56 PM

From the the FPS readings and the AMD monitoring tool's readings, I do not think you need to troubleshoot. It's working, GPU-Z is just not giving you good readings. 65% boost in performance is great for crossfire on that card.
Score
0
October 13, 2011 11:34:50 PM

photonboy said:
You probably know this, but the Crossfire update is a separate from the main Driver and CCC update.

It's called "ATI Catalyst Application Profiles" and is usually about 0.5MBytes. It's right below the main AMD driver and has the same number (latest is v11.9).



yeah i did install that also :)  thanks.

bystander said:
From the the FPS readings and the AMD monitoring tool's readings, I do not think you need to troubleshoot. It's working, GPU-Z is just not giving you good readings. 65% boost in performance is great for crossfire on that card.


yeh i thought that i had a pretty nice boost on bad company 2, i can really feel it when the second card is disabled so i felt it was working somewhat. its just that gpu-z led me to believe i wasnt getting the full potential out of my setup and i can never rest when i feel thats the situation! i like to get what ive paid for! :p  i have reported this bug to gpu-z dev so hopefully they'll get this fixed in the next release :) 

really appreciate all the help offered here from everyone!! im off to play some games :D 
Score
0
October 14, 2011 4:01:38 AM

scaling on the 4000 and 5000 series cards was not the greatest,maybe 50% depending on games.now the 6000 series cards are a different story,depending on the application/game,they scale great,any where from 80-100%.A great and cheap upgrade into the 6000 series is to crossfire 2-6870's,on some games they will even beat the 590 and 6990 dual gpu cards.i have done a lot of research into this as I am running 2 5870's,which is beat out or matched by the dual 6870's due to the difference in scaling.
Score
0
October 14, 2011 4:40:27 PM

Earnie said:
scaling on the 4000 and 5000 series cards was not the greatest,maybe 50% depending on games.now the 6000 series cards are a different story,depending on the application/game,they scale great,any where from 80-100%.A great and cheap upgrade into the 6000 series is to crossfire 2-6870's,on some games they will even beat the 590 and 6990 dual gpu cards.i have done a lot of research into this as I am running 2 5870's,which is beat out or matched by the dual 6870's due to the difference in scaling.


im happy with the boost on battlefield at least, thats what i play most. if the 6870's match your 5870's at times then is it really worth the upgrade? i was thinkin of crossfiring some 5 series cards next year sometime. i reckon my 4870's will probably play any game going on high settings at 1920x1080 with 35fps+. i find it quite good on the wallet to stay behind by a year or 2 on hardware but still get good performance. i would cry if i paid like £300-£400 for a sweet 6 series card only to find that it starts struggling a couple years down the line and i can only sell it for a fraction of what i paid! i bought a second 4870 for 50 quid the other day and it was worth every penny for the boost it gave :) 
Score
0
October 14, 2011 5:09:02 PM

Good read for sure. Either I'm lucky or my eyes aren't the best as I haven't noticed any such issues. And believe me I looked into this stuff before going Crossfire. Call me an addict, I had a Voodoo2 SLI setup back in the day! :) 

IMO almost anything at the "enthusiast" level is going to be less then perfect. You don't buy a Corvette C7 for the high quality interior trim. ;) 
Score
0
October 14, 2011 5:10:32 PM

yeah thats cool, read most of it :)  i prefer having a single card setup but if i sold my 4870 for 50 quid and put the 50 i spent on the other 4870 into the pot id have 100 quid. there's no way i could buy a 100 quid card and get better performance. i think crossfire/sli can be quite nice if your graphics card is a little dated and you just want a bit of a boost before getting a decent new single card at some point within the next year :) 
Score
0
October 14, 2011 5:10:38 PM

photonboy said:
http://www.tomshardware.com/reviews/radeon-geforce-stut...

Anyone interested in Crossfire or SLI should read this article. Don't flame me, I just said you should read it and am not expressing an opinion on the matter.


Ya, I've seen it. I've read it and believe it. I do, however, have 1 issue with it. That only tested the 6870 cards. I've seen a lot of complaints from those with 6870's in crossfire and far fewer with any other AMD card. While it might be an issue with all AMD cards, it would have been nice to have seen some tests on other cards.

They did a small sample of 6950's and mentioned it was not as big of a problem and made the assumption that it's only midranged and below cards that have issues, but it is possible it's just the 6800 cards.
Score
0
October 14, 2011 5:15:36 PM

C0rehound said:
Good read for sure. Either I'm lucky or my eyes aren't the best as I haven't noticed any such issues. And believe me I looked into this stuff before going Crossfire. Call me an addict, I had a Voodoo2 SLI setup back in the day! :) 

IMO almost anything at the "enthusiast" level is going to be less then perfect. You don't buy a Corvette C7 for the high quality interior trim. ;) 


ive just noticed frame rate drops when things get hectic, but ive noticed that on any setup!! i remember getting a voodoo 2 back in the 90's :)  think it was 12MB or something haha... i got it to play the second rainbow 6 game! forgotten the name of it....

you're right though man. ive always found faults in every pc ive had. its only cos playing new games will push your system in certain areas and you're like man i need to upgrade this now....then this....etc it never ends!! thats why ive just got into using hardware from a year or 2 back which you can get hold of at very reasonable prices secondhand and still get good performance :) 
Score
0
October 19, 2011 2:15:12 AM

My,second Gpu isn't doing much either I have 6970s in supposed crossfirex but while playing the witcher2 or crysis full screen the 1st gpu would go to 80% activity and 60C while my 2nd stays either idle 35C 0%Activity,sometimes even 0C.Gpu-z says crossfireX enabled so does catalys,t says cfx enabled ,yet only detects one gpu and i tried playing with and without CF and there is no difference in FPS.
Did you find a fix for your problem?
Score
0
October 19, 2011 9:52:49 AM

well to be fair, the second card probably isn't needed to play those games in most scenarios, hence the frirst card only being utilized 80%. but you would expect a crossfire setup to balance the load between the 2 cards wouldn't you!? i would :/  i realized that gpu-z wasn't working properly though. i dont know if you read my post about it but it only refreshes readings for the card you have left selected in the drop down menu. its clearly a bug with gpu-z because i ran amd hardware monitor and fur mark which said both cards were being heavily utilized. the performance increase in battlefield bad company has been huge to so clearly crossfire is doing something.

so i would recommend running amd system monitor then while in the thick of gaming quickly "alt-tab" out and check what it says. you'll probably see the graphs up quite high.

also you say no fps difference with crossfire on. what fps are you getting and are you using vsync?? vsync caps the frames per second to your monitors refresh rate, for example: 60hz refresh rate on the monitor with vsync on = 60fps max!

hope this helps, let me know :) 
Score
0
October 19, 2011 7:53:13 PM

Ok Idid what you suggested with crossfire enabled a got an average 105FPS while just looking at scenery and walking around minimum of 45FPS and a max of 187FPS,while in combat I got a minimum of 33Fps ,average of 40Fps and Max of 70Fps.

With crossfire disabled, I got an average of 66FPS,Min of 40Fps, and max of 88FPS while looking at scenery ,but in combat I got 43FPS minimum,53FPS Average, and 58 Max. So in combat I get better fps with crossfire disabled,while in just walking walking around and looking at scenery I get 5 less frames minimum,39 less on average and almost 100 less Frames on max,.

So CF is working fine,also both where at 50-60% activity each with Crossfire enabled and with it disabled 80-100% activity. 55C Top and 47C bottom CF enabled and 60C-70C Disabled, after reinstalling drivers for a 3rd time xD. The load isn't balanced as you said sometimes top 60% and bottom 20-40% it shifts as I play.
Score
0
October 19, 2011 9:26:11 PM

sounds like things are working alright :)  some games just dont work well with crossfire/sli so dont worry that something is wrong if crossfire doesn't give u a huge boost in all your games.
Score
0
October 20, 2011 6:16:09 PM

i7-860, 8GB DDR3 1600MHz, HD5870, U2711 2560x1440, Win7 Premium x64

**I've got a LONG LIST of tips that I think are useful for gamers. Please don't flame me for the length. You don't need to read it. I know I would have liked to know all these things a few years ago.

I don't have Crossfire, but I believe this summarizes things for Crossfire and general game tweaking:

1) If you're getting an increase in frame rates in any game then the basics are set up correctly.

2) Crossfire varies from NOT SUPPORTED to as high as 90% in some situations. (65% seems to be the average across some samples).

3) I believe Crossfire can be set up on a per-game basis?

4) If a game can get MORE than 60FPS with full settings it MAY be best to DISABLE the second card in regards to stutter. Experiment to see.

5) Look at frame rates using FRAPS or similar. FRAPS also has a benchmark tool (hit the FUNCTION KEY assigned and wait, say, two minutes (can change this) and get the average. Toggle in SOFTWARE between Crossfire and single-card setup.

6) Some motherboards can lower the bandwidth on graphics card slots when a feature, like TurboUSB is enabled. See your motherboard manual. This STILL may not matter if the card doesn't need the bandwidth. Don't guess, if you WANT the feature enabled benchmark before and after enabling this feature.

7) VSYNC should be OFF when benchmark.

8) VSYNC should be ON when playing if you get above the recommended refresh rate for the monitor (usually 60Hz so 60FPS).

9) VSYNC achieves TWO main things:
a) stops the computer from working harder than needed (100FPS on a screen that shows 60FPS creates heat and noise).
b) Eliminates screen TEARING. In some games it is VERY noticeable.

10) VSYNC and Anti-Aliasing are two important features that sometimes need to be FORCED on using a tool like RadeonPro.

11) When I adjust for the BEST quality, my first concerns are getting VSYNC working and the best Anti-Aliasing I can support (depending on other quality settings impact). If I can't get full quality I play around with SHADOWS and other settings with VSYNC OFF and I aim for about 65FPS 90% of the time.

Once satisfied (using FRAPS benchmarks are eye-balling it), I turn ON VSYNC, hide the FPS indicator, and just play the game.

12) Metro 2033 is an example of a game that supports DX11 but at a great cost. IMO, it's better to run a game with 60FPS VSYNC and High-Quality DX9/10 then to play at 30FPS using the best DX11 settings.

Game visuals is a combination of solid frame rates and quality graphics. Again, I prefer a solid FPS to a jerky but slightly prettier picture.


13) **Forcing VSYNC and AA:
This requires some experimentation, but here are a few examples to get started. It can make a HUGE difference in many games.

Example #1: Mass Effect 1, 2
- 4xAA Supersampling (Supersampling is far better than normal. Choose the best quality you can support while keeping your frame rates high, preferably 60FPS if possible.)

Example #2: Deus Ex (1)
- Add DeusEx to RadeonPro. "Deus Ex.exe" in Steam (on "E" drive): E:\STEAM\steamapps\common\deus ex\System
- 1366x768 res chosen (highest supported, most games I choose 1920x1080)
- 8xAA forced in RadeonPro
- Toggle F9 until FPS is HIDDEN once I'm sure I get a solid 60FPS
- (DeusEx needs to set the Display to 32-bit TWICE as the first time doesn't stick.)

Example #3: The SIMS 3
- Always ON (DOUBLE VSYNC)
*This game on my HD5870 has screen tearing and stutter with the normal settings. I don't even know if it supports VSYNC. If you can get 60FPS with full quality settings and still get VSYNC that's ideal.

Anyway, for most people if they can get this working, Double VSYNC (30FPS synched) is by far the best option.

Example #4: (generic, DOUBLE VSYNC)
Those people with less powerful systems may wish to try this option to force VSYNC at 30FPS. Perfect for those getting between 32 and 45FPS on average who want to eliminate screen tearing and stabilize frame rates. *It really varies between games; basically try it and see if you like it.

Example #5: (generic settings for unsupported VSYNC and Anti-Aliasing):
- 4xAA (override application settings)
- VSYNC ON

Example #6: (VSYNC @50Hz)
- My monitor also supports 50Hz. I have a few games that can sync at 60FPS bet frequently drop into 50Hz. These games work much better if I just use RadeonPro to force 50Hz VSYNC.

I had to use "ATITrayTools" to force VSYNC on for Witcher 1 (RadeonPro would not work). Awesome! Witcher 1 had a lot of screen tearing on my rig.

The Force Unleashed 1 and 2:
Not supported. TFU1 is locked to 30FPS and has no AA. Not even sure if VSYNC works properly. Sigh...

14) AUDIO and EAX:
a) If your card and game both support EAX you will need to run Alchemy to set up your game for the best sound. I found in general that copying the LINK (from Windows Explorer) to where the main EXE file is worked best.

b) The game may show up automatically (game must be RUN at least once). You must enable it so it is on the RIGHT side of Alchemy.

c) I reinstalled Max Payne 1 and oddly my audio is working just fine! I didn't add it to Alchemy like I needed to before. Did STEAM fix this? I added the rest in anyway as an example, and also if it still isn't working for others.
Here's one example of a game I couldn't play for a long time until I solved the issue:

Example: May Payne 1 (STEAM)
- unplayable due to sound issues
- try to add to Alchemy but it says its already there? (Is there an internal list? If so it may not work with Steam and gets screwed up? Game isn't viewable in Alchemy either way.)

Max Payne 1 SOLUTION (STEAM):
When adding to Alchemy name it anything except "Max Payne", I called it "Max Payne (steam)". I then added it by using the Game Path E:\STEAM\steamapps\common\max payne

... then moving it to the RIGHT side to enable it.

15) STEAM
I really like Steam. Here's some really quick tips:
a) If you have lots of games you can view only by Favorites.
b) You can "verify the integrity of the game cache"; this will add missing files or replace corrupted ones. Saved me a lot of time with a corrupted hard drive that I copied to a new one.
c) You can MOVE the entire Steam folder. Instructions at Steam. Read carefully. VERY HANDY if you need to reinstall Windows.

16) Recommended drive setup:

a) 120GB SSD (Windows, applications etc)
b) 2TB WD Green hard drive (STEAM folder, *non-Steam games folder, Windows backups, Downloads, Media files... )

*If a game is installed to a second hard drive you must manually choose the location. I just created a "GAMES" folder on my 2TB hard drive (E-drive) then point my games at that when installing.

For STEAM games, you setup STEAM the first time and it handles everything after that. I created STEAM at "E:\STEAM" on my hard drive.

(Did you know that a Solid State Drive will boot games quicker but the speed advantage over a hard drive often isn't amazing. The worst case scenario I saw was a 120-second game boot which broke down like this: 2 seconds for the SSD and 118 seconds for the CPU/Memory/Graphics to configure the game. In other words if the hard drive took 10x as long it would have taken 138 seconds instead of 120 seconds.)

Finally...

17) XBox 360 controller for PC:
*Strongly recommend the WIRED version. Why?
- USB receiver for the XBox 360 prone to failure. Mine did. You can NOT buy it separately.
- Cost. Costs more plus you need batteries (rechargeables like Eneloop are best).
- alternate Logitech F710 wireless. (problems with Vibration being too low. replacement same. apparently a design flaw?)

Save yourself the hassle and just get the standard WIRED version of the XBox 360 PC controller (same as the XBox version?).

Do I need a game controller? Now a lot of games for PC that support it, most noticeably Batman Arkham Asylum and Batman Arkham City, also Lego series, Assassin's Creed, Darksiders, Prototype, Fallout 3/New Vegas and of course...

SKYRIM!!!
Score
1
October 20, 2011 6:44:05 PM

some of that seemed a little random lol but i appreciate the effort you put into writing it all. some useful tips for newbies around the beggining :) 
Score
0
June 8, 2012 2:04:59 AM

do both cards have the same vram on them if not one card will bottle neek its self
Score
0

Best solution

June 26, 2012 12:57:31 PM

O.K., so I finally broke down and created a user ID after all these years viewing TH.com as a visitor, just for this thread. I think you have misunderstood a minor detail that is causing your false reading. - when you are playing a game (in full screen) X-fire is enabled. - when you alt-tab out, X-fire is disabled. (at a rate far faster than we humans comprehend). GPU-Z is showing you what's going on right now, on this display. (which isn't in X-fire mode, because it's not in fullscreen, because you are viewing GPU-Z). You must use a program like "Afterburner" (ATI/nVidea) or "EVGA Precision" (nVidea) that have time graphs, to see the GPU loads for the last minute or so. (back when you were still X-fired in fullscreen). ;) 
Share
June 26, 2012 1:08:49 PM

Check the date!

This topic has been closed by Pyree
Score
0
June 26, 2012 2:19:07 PM

Best answer selected by Pyree.
Score
0
!