![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Best way to stress test a new 5870
- Thread starter digicat
- Start date
To test the capabilities of the heatsink, yes; for an overclock no. For overclocking, it's usually best to test with a very graphically demanding game (Crysis, etc.). Remember to observe your performance closely when overclocking the RAM, as it only shows instability as increasing drops in performance, since the ECC corrects artifacts.
Thanks for the info.
Shadowssong
[H]ard|Gawd
- Joined
- Sep 17, 2009
- Messages
- 1,969
Why do you want to stress your new card? Are you unsure of its stability?
I just put the system together and am in the process of a 24 hour Furmark stress test.
Just wanting to be sure quickly. I don't automatically trust a component just because it's fresh out of the box. I'm not gaming a whole heck of a lot (maybe a few hours a week), so I wanted to try to single out any bad components quickly, not a few months down the line when I'm in the middle of a game.
Keep a fire extinguisher ready...
24hrs of furmark and the card hit 88c as it's highest temp.
It's the xfx xxx edition, so it comes with a factory-overlcock of 875/1300.
I ended up RMAing my mobo yesterday (coil whine problem), so no testing or gaming for a week or so.
heatlesssun
Extremely [H]
- Joined
- Nov 5, 2005
- Messages
- 44,154
I like to use the Lost Planet EC benchmark which will run continuously. Pump up the settings and that game can still give a GPU a very good workout.
Make it get married and have kids…..
Play your favorite game for hours on end.
Run benchamrk after benchmark.. Then game for hours..
Record a long play through in Crysis and run it over and over again..
I would test with a bunch of games that use different engines. The one 5870 I had was stable in Crysis, Furmark, you name it but would artifact with 4XAA + on any title that used the Orange Box version of Source
OK, but what if you're NOT a gamer? That is, you want a decent vid card because Adobe uses the vid card in CS 4 to improve performance.
If you're not a gamer, what software or utilities should you run? (Please don't say run a Gaussian blur in Photoshop.
)Thanks
Case: Coolermaster 690
CPU: i7 860
Mobo: Asus p7p55d-e pro
CPU and GPU at stock speeds.
I'm only using 2 of the fans in the CM690 though, the bottom-front and the back exhaust. I'm using the Thermalright Venomous X heatslink with a 120mm fan on it that pulling in air from the unused 5.25 bays (the bay covers have a mesh that lets the air in). Since the 5780 is so long, it pretty much creates a chamber, being fed air from the bottom-front fan.
I'm trying to figure out what to do with the 3rd case fan that came with it. It came factory mounted on the side, but I worry that it'd mess with the air flow from the bottom-front fan.
24hrs of furmark and the card hit 88c as it's highest temp.
It's the xfx xxx edition, so it comes with a factory-overlcock of 875/1300.
I ended up RMAing my mobo yesterday (coil whine problem), so no testing or gaming for a week or so.
Uh, how do you know it was your mobo? I had coil whine on my PSU, my 5870s and my 250 GTS but not ever my mobos.
I took the PSU out of the case, and powered it up. PSU was silent, the noise was in the case. I then moved the graphics card down a slot, and put a big book between the graphics card and the CPU, then pretty much put my head in the case and listened. The noise was loudest on the CPU side than on the GPU side.
Also, I found another message board where a guy was complaining about the same problem with the "deluxe" version of the mobo I had.
Anyways, I wanted to do something, so I started with what I thought the problem was, the mobo. Since I just bought it a few weeks ago, I got an RMA from Newegg and sent it back.
as far as the 3rd fan, you can leave it as a side intake. the stock case fans don't push nearly enough air for the front fan to interfere with the side fan.
i have a cm 690 as well. i've replaced all the stock fans with 1200 rpm scythe slipstreams. they push about 63 cfm. i'm adding another slipstream in the 5.25 drive bays for intake. then i'm adding two scythe kama flex's (fluid bearings). one in the bottom for intake, and one in the top rear for exhaust.
you really shouldn't count on the bottom intake to bring in too much air. it has to pull through the front cover of the case, plus the filter. then the air has to pass over the hard drive cage and however many hard drives you've got in your system.
i have a cm 690 as well. i've replaced all the stock fans with 1200 rpm scythe slipstreams. they push about 63 cfm. i'm adding another slipstream in the 5.25 drive bays for intake. then i'm adding two scythe kama flex's (fluid bearings). one in the bottom for intake, and one in the top rear for exhaust.
you really shouldn't count on the bottom intake to bring in too much air. it has to pull through the front cover of the case, plus the filter. then the air has to pass over the hard drive cage and however many hard drives you've got in your system.
OK, but what if you're NOT a gamer? That is, you want a decent vid card because Adobe uses the vid card in CS 4 to improve performance.
If you're not a gamer, what software or utilities should you run? (Please don't say run a Gaussian blur in Photoshop.
Thanks
OCCT Gpu test. Furmark will tell you nothing because it will continue running even if the card starts producing artifacts. IE my experiences with Furmark with ATI HD 4890 cards showed that its not as sensitive to OC instability as other apps are and that it will continue running despite blatantly apparent artifacts popping up.
I would not be in favour of FURMARK for 24hrs...jeez OVERKILL. I think the best benchmark for stabilty etc is real world tests.....i.e. use your PC in everyday tasks and you will find a problem if one is there.
Remember when you overclock your rig, its not just the CPU you overclock its all the components as they all work together.
Game, game & game some more....BTW when I ran furmark to test my fan was working and to record my temps the max temps I got after 30 mins was indeed 88 degrees. My idle temps are 24/25 degrees.
Remember when you overclock your rig, its not just the CPU you overclock its all the components as they all work together.
Game, game & game some more....BTW when I ran furmark to test my fan was working and to record my temps the max temps I got after 30 mins was indeed 88 degrees. My idle temps are 24/25 degrees.
I've removed the HD mounts, so asside from a single SSD mounted in the top location, the drive bay is open.
inotocracy
Supreme [H]ardness
- Joined
- Jul 25, 2004
- Messages
- 5,625
Best way to stress test it-- do an all nighter, play some games, drink lots of coffee. If it survives, you're probably alright.
revenant
[H]F Squad Captain
- Joined
- Apr 10, 2005
- Messages
- 16,354
I think that's a pretty damn good method... if I don't say so my damn self. er.. your damn self? That's pretty much what I did on my own 5870.. just gamed it hard all weekend and drank a lot of diet pepsi max.. stayed up very late.. played lotro and l4d2 back to back lots.. some dead space and torchlight also.. good times.
i only have a single drive in my cm 690 as well, and have removed all other hdd mounts. still, with the hard drive cage (the metal that the mounts click into) you're getting a lot of airflow restriction. just having that single front intake is a bit insufficient, imho.
I do notice that the air the GPU vents out the back is pretty warm, while the air the CPU vents out is cool. Is that at all a good indicator that the GPU needs more air under it?
Do you think the bottom-front case fan is so hampered that it'd be better to just remove it outright and use the bottom-side?
KayossZero
Limp Gawd
- Joined
- Mar 8, 2003
- Messages
- 420
I've never heard of this, is this something only newer VPU's are doing? My 9800Pro I seem to recall definitely displayed artifacts if the memory OC was too high.
Brand spankin' new, IIRC. It was introduced with the 58xx cards.