|
[H] does fairly good graphics card reviews. I know they did a bad on their IQ review, but when it comes to performance gains or decreases, [H] is a good resource.
|
#
?
Jan 30, 2014 18:07
|
|
|
|
| # ? May 30, 2024 19:16 |
|
|
At best I can do tests with single/double/triple card in single player mode. I'm not in the mood to gently caress around in multiplayer on bf4. I haven't benched anything in years, what programs would you like me to test with. One of the neat features of my board is that I can physically turn off pci-e slots. So it'll actually be like testing as if I pulled the card out (which ain't gonna happen cause water). I also need to reflash my cards with whatever magical bios set the middle card to 1030mhz.
|
|
#
?
Jan 30, 2014 18:28
|
|
|
veedubfreak posted:At best I can do tests with single/double/triple card in single player mode. I'm not in the mood to gently caress around in multiplayer on bf4. I haven't benched anything in years, what programs would you like me to test with. One of the neat features of my board is that I can physically turn off pci-e slots. So it'll actually be like testing as if I pulled the card out (which ain't gonna happen cause water). I also need to reflash my cards with whatever magical bios set the middle card to 1030mhz. Probably just use fraps, and run through a small chunk of a level that has some indoor/outdoor/lots of big boom explosions. If you;re feeling super frisky, find FCAT 
|
|
#
?
Jan 30, 2014 19:07
|
|
|
If you aren't wanting to record and just want frame rates pop open a console in BF4 and run Render.DrawScreenInfo 1 to show the rates on your screen. Alternatively do PerfOverlay.FrameFileLogEnable 1 to dump the info to a file.
|
|
#
?
Jan 30, 2014 19:30
|
|
|
Where are my drivers ATI!!!!
|
|
#
?
Jan 30, 2014 22:34
|
|
|
veedubfreak posted:One of the neat features of my board is that I can physically turn off pci-e slots. Wait, what board is this? I'm really wattage oriented and I've love to switch between Integrated Graphics / Single GPU / SLI depending on what I'm booting my PC to do.
|
|
#
?
Jan 30, 2014 23:15
|
|
|
http://www.newegg.com/Product/Product.aspx?Item=N82E16813128596
|
|
#
?
Jan 30, 2014 23:31
|
|
|
|
|
#
?
Jan 31, 2014 01:57
|
|
|
It wasn't so bad. It was on sale for 399, I got a 40 dollar discount for buying my 4770k and they already approved my 100 dollar rebate. Sadly if you want more than 2 cards you gotta pay to play.
|
|
#
?
Jan 31, 2014 02:03
|
|
|
MrMoo posted:Thanks, I had a discussion with the Google guys and at least there is a defect with the FPS counter: Google cannot explain the strange paint tiles but there is a complete rewrite from Android, of all places, that apparently fixes or at least improves this performance quirk.
|
|
#
?
Jan 31, 2014 02:08
|
|
|
GeForce Titan Black Edition is semi-confirmed. If you need a compute+gaming card and held off the original Titan, this may be for you. Also, I remember reading another article that confirmed the 790 existence but with just 6GB (2x3GB actually) of VRAM instead of 10GB. Did someone already post that here?
|
|
#
?
Jan 31, 2014 06:06
|
|
|
So, if nvidia isn't going to step up the VRAM in their consumer line because "the bottom line" won't that make AMD the goto for multiple display setups? It just seems like Nvidia is shooting themselves in the foot.
|
|
#
?
Jan 31, 2014 06:16
|
|
|
SlayVus posted:So, if nvidia isn't going to step up the VRAM in their consumer line because "the bottom line" won't that make AMD the goto for multiple display setups? It just seems like Nvidia is shooting themselves in the foot. Unless you are going over 3x1080p then more than 2gb doesn't matter right now http://www.pugetsystems.com/labs/articles/Video-Card-Performance-2GB-vs-4GB-Memory-154/ http://i0.wp.com/alienbabeltech.com/main/wp-content/uploads/2013/08/MainChart.jpg
|
|
#
?
Jan 31, 2014 06:25
|
|
|
Hmmm, here's a question. Is there anything to suggest that the bit coin inflation thing will turn around at any point, or is that a here to stay status quo? As much as I like the idea of Mantle and a bit more memory and performance for down the road gaming [Even texture packs on Skyrim can push the limits of 2gbs. Whatever I'm modding in a year or two may make that look downright simple], a fifty to a hundred dollar mark up on what they were is crazy.
|
|
#
?
Jan 31, 2014 07:02
|
|
|
There's a thing where after enough coins are mined into the economy, the "difficulty" increases to the point where it is like twice as hard to mine coins with the same hardware, which means you're getting them twice as slow or burning double the electricity. When that happens people will give it a rest or try to develop dedicated mining hardware like ASICs and the GPUs will return to normal price. Either that or AMD ramps up production enough to meet demand, or a combination of both.
|
|
#
?
Jan 31, 2014 07:14
|
|
|
Well, Alpha Technology has supposedly been developing a scrypt asic that has a massively better power to hash ratio than any GPU. They're developing it with an Indian company. If something actually comes of that, it might upset the balance, but even the "cheap" one is fairly expensive.
|
|
#
?
Jan 31, 2014 09:56
|
|
|
Phuzun posted:Maybe check the setting that auto-updates the game configurations. Never used that myself, but that sounds like what it says it does. I looked around online and I think part of the weirdness had to do with AC4 forcing 30fps with VSync on if your card can't put out a continuous 60fps. I knew something was weird when I'd turn shadows from low to high with no framerate difference at all. Turning VSync off also isn't an option due to excessive screen tearing. I ended up having to download RivaTuner to force triple buffering and now the game runs on high settings at 45fps in open areas and 60fps everywhere else, this is with a GTX 760.
|
|
#
?
Jan 31, 2014 11:26
|
|
|
exquisite tea posted:I looked around online and I think part of the weirdness had to do with AC4 forcing 30fps with VSync on if your card can't put out a continuous 60fps. I knew something was weird when I'd turn shadows from low to high with no framerate difference at all. Turning VSync off also isn't an option due to excessive screen tearing. I ended up having to download RivaTuner to force triple buffering and now the game runs on high settings at 45fps in open areas and 60fps everywhere else, this is with a GTX 760. Download Nvidia Inspector and force a frame cap at 59/60fps or adaptive V-Sync. Hell everybody download Nvidia Inspector or Radeon Pro, if you're asking "It'd be cool if you could make my GPU do X" chances are those two can do something like that.
|
|
#
?
Jan 31, 2014 13:37
|
|
|
Looks like that Star Swarm demo is out on Steam (Nitrous Engine). Once those Mantle drivers are out it'll be interesting to see what kind of performance delta we see.
|
|
#
?
Jan 31, 2014 14:35
|
|
|
I'm not an expert but from what I read, AC4 has a built in frame cap at 63fps and no native triple buffering option, so with VSync on it'll automatically drop to 30 if your framerate dips even a little. Is there a benefit to setting your framecap to 60 in this scenario?
|
|
#
?
Jan 31, 2014 14:49
|
|
|
exquisite tea posted:I'm not an expert but from what I read, AC4 has a built in frame cap at 63fps and no native triple buffering option, so with VSync on it'll automatically drop to 30 if your framerate dips even a little. Is there a benefit to setting your framecap to 60 in this scenario? Setting it to 60 should prevent tearing from going above 60 FPS, I don't get why it's a 63 FPS cap since that's just asking for most monitors to go on a tearing spree. Speaking of AC4, the options to fix it on nVidia at least seem to be either Adaptive Vsync that can be enabled in the nVidia control panel or using D3DOverrider to force Triple Buffering and Vsync, is there an option that's better than the other or is it effectively the same?
|
|
#
?
Jan 31, 2014 15:12
|
|
|
I had noticeable tearing with only Adaptive VSync despite the higher 50+ framerates, while with D3DOverrider I would average about 40fps in open areas with no tearing. I haven't tried to cap framerates though or try a combination of force triple buffering + Adaptive. I mean the game looks beautiful on every setting but the little fraps number going down to 30 really bums me out!
|
|
#
?
Jan 31, 2014 15:58
|
|
|
exquisite tea posted:I had noticeable tearing with only Adaptive VSync despite the higher 50+ framerates, while with D3DOverrider I would average about 40fps in open areas with no tearing. I haven't tried to cap framerates though or try a combination of force triple buffering + Adaptive. I mean the game looks beautiful on every setting but the little fraps number going down to 30 really bums me out! I am almost 100% sure that Adaptive Vsync takes priority over everything. 99%ish. Pretty sure.
|
|
#
?
Jan 31, 2014 16:23
|
|
|
I have an ASUS 660 TI I'm trying to extend it's life a bit, well, reduce its life by extending its performance probably. I have it overclocked using the ASUS GPU Tweak utility at 1140 MHz core clock (at least that's what it's set to, from what I understand it only does 13 mhz increments) and a 6550 effective memory clock. When I watch the graph though that actually comes to 1202 MHz core clock. Frankly it does very well for what I want it to do with the exception of AA. I wouldn't turn down a little more FPS but I'd like to focus on the AA. From what I've read the AA performance is poor due to the 192 bit bus. My temperatures at this overclock are low, 40-50 *C. But it is unstable at the next increments, and rock solid where it's at now. I've been reading on how to unlock the voltage via bios mod. I swear I've read that I should never, ever do this somewhere but I think it might have been for cpu overclocking. It seems that it's a popular thing for these 660 ti's though. If it's not some terrible idea I'd like to do it, but should that allow me to overclock memory further? Most seem to focus on the core clock and I'm not sure if that's because the memory bottoms out for reasons other than voltage or not. Or will focusing on one or another make any difference at all, as in they will both max out to the same numbers regardless if I focus on one or another?. People seem to hit ~1330 core clock easy enough after the simple bios mod. I'd be happy if that's what I could get out of this since reference is 967, but I have a feeling memory speed might be more important here. Also I don't recall maximum voltage limit being an option in ASUS Gpu Tweak, just minimum. If the bios mod raises the max by say, 50% would it just go wild and attempt to use all of that when it thinks it needs it?
|
|
#
?
Jan 31, 2014 16:40
|
|
|
Your ASUS 660Ti runs that cool under load with that overclock? Mine gets up to 72C under load at the stock settings even with the fancy custom cooler that came with it. I would redo the thermal paste on my card but wouldn't that void the warranty?
|
|
#
?
Jan 31, 2014 16:56
|
|
|
spasticColon posted:Your ASUS 660Ti runs that cool under load with that overclock? Mine gets up to 72C under load at the stock settings even with the fancy custom cooler that came with it. I would redo the thermal paste on my card but wouldn't that void the warranty? It's an ASUS Direct CUII if that makes any difference. I've read it's one of the better coolers. But yeah I've never seen above say 60 at most after hours and hours of gaming at those clocks, it's usually 40-50. And I always watch it since GPU Tweak has to be on all the time. The crazy part is the fan speed almost never, ever hits full speed. It's so infrequent it has started me in the past because it sounds so foreign I think something might be going wrong for a second. I've just left it on auto for the time being. I'm also going to replace my paste if I do this bios mod since I've only heard bad things about all factory paste, but especially on video cards. And then fan settings might come into play for sure. I'm sure it would void the warranty though but that's never been a huge deal for me. Chances are I'm the one who's going to break it anyway Edit: I do remember seeing it hit 60* with Furmark. Ignoarints fucked around with this message at 17:14 on Jan 31, 2014 |
|
#
?
Jan 31, 2014 17:12
|
|
|
That's the same cooler mine has so that probably means the thermal paste on my card dried out or they applied too much at the factory. At those temps I was thinking you were using water cooling or something fancy like it.
|
|
#
?
Jan 31, 2014 17:16
|
|
|
spasticColon posted:That's the same cooler mine has so that probably means the thermal paste on my card dried out or they applied too much at the factory. At those temps I was thinking you were using water cooling or something fancy like it. Yeah it's completely stock hardware that I haven't touched, I'd take a look for sure since that's a huge difference. Or if you're in warranty I guess you could tell them its getting way too hot and throttling if you can stand the wait. Edit: Have you tried overclocking it with the asus utility or have you not because the temps were already dumb?
|
|
#
?
Jan 31, 2014 17:21
|
|
|
I haven't tried overclocking yet because of the high temps and I can't decide if I want to just leave it alone until the card actually starts overheating and throttling, go ahead and redo the thermal paste then overclock if the temps go down, or try to haggle a replacement out of ASUS because the card works as is right now. Edit: Isn't the customer service at ASUS supposed to be terrible? spasticColon fucked around with this message at 17:40 on Jan 31, 2014 |
|
#
?
Jan 31, 2014 17:36
|
|
|
spasticColon posted:I haven't tried overclocking yet because of the high temps and I can't decide if I want to just leave it alone until the card actually starts overheating, go ahead and redo the thermal paste then overclock if the temps go down, or try to haggle a replacement out of ASUS because the card works as is right now and it isn't throttling at least not yet. The difference is more than noticeable at the clocks I'm at even now. I'd be pretty pissed if I had to go back to stock. I guess the dilemma is void the warranty or not, and if you don't void the warranty would they even replace it. I guess long run plan would be seeing if ASUS replaces it first then if they don't you can mess with the paste. But that leaves you out of a video card that otherwise worked and I know how that blows. But if I were you and seeing just how different it is, I'd recommend going for it. The performance difference is significant. Edit: I just read that people are able to warranty it even with aftermarket heatsinks attached. However it is expressly written that it voids the warranty. If that thread is a good indication, reapplying paste to solve a heating issue with an otherwise stock card would be the tamest tampering of them all. http://www.tomshardware.com/answers/id-1755678/asus-gtx-670-dc2-warranty.html Ignoarints fucked around with this message at 17:44 on Jan 31, 2014 |
|
#
?
Jan 31, 2014 17:41
|
|
|
Don't discount the possibility that it's just a substantially better chip than his. Or the zillion variables we don't know about, including important ones relating to airflow, that could account for the temperature delta. Overclocking is so far from an exact science that it's more guesswork and prayer than anything else, but there's a lot involved in this particular setup that we can't really speak to one way or another, and it may be a bad idea to jump to the conclusion that his card definitely has a thermal paste problem and yours doesn't when there are plenty of other explanations for the differences you're discussing. FactoryFactory has a tiny form factor case with restricted airflow, and as a result, when he bought my GTX 680, his temperatures were like 30ºC higher than mine under load. And that's the same card sample, too, which eliminates a whole mess of variables. Just be careful assigning causal relationships to things prematurely and definitely be careful if warranties are on the line for a relatively small performance difference. I don't know what Asus' policies are regarding card modification, so I can't really speak knowledgeably there, but you might want to figure that out before doing anything.
|
|
#
?
Jan 31, 2014 17:52
|
|
|
Ignoarints posted:The difference is more than noticeable at the clocks I'm at even now. I'd be pretty pissed if I had to go back to stock. I guess the dilemma is void the warranty or not, and if you don't void the warranty would they even replace it. I guess long run plan would be seeing if ASUS replaces it first then if they don't you can mess with the paste. But that leaves you out of a video card that otherwise worked and I know how that blows. Well when I play a game or run a benchmark GPU-Z reports that it hits 1084Mhz on the GPU then it settles back to 1071 after a few minutes so I guess its throttling slightly but its not going all the way back to the stock speed of 915 or stock boost of 980. Which is really weird because I'm not doing any kind of tweaking or overclocking of the card.  Did you get a factory overclocked version of the card?I'm just going to leave it alone unless it starts overheating or outright breaks then I'll RMA it. By the time the warranty runs out I'll be looking to upgrade anyway.
|
|
#
?
Jan 31, 2014 17:59
|
|
|
Agreed posted:Don't discount the possibility that it's just a substantially better chip than his. Or the zillion variables we don't know about, including important ones relating to airflow, that could account for the temperature delta. Yeah you're definitely right, I just have little care for warranties for things like this and typically just go for it since there is so little to lose, to me. But I know a lot of people do care for good reason. I have great temps but still applying paste anyways, even being in warranty, just based on really good results on the internet. If you want some case info spastic I have a fairly basic Antec 300. I recently mounted 2 intake fans in the bottom front because of a major cpu and motherboard change, but before that I had two factory exhaust fans on the back and top. The numbers didn't change either way for the video card. The video card has space all around it and isn't cramped. spasticColon posted:Well when I play a game or run a benchmark GPU-Z reports that it hits 1084Mhz on the GPU then it settles back to 1071 after a few minutes so I guess its throttling slightly but its not going all the way back to the stock speed of 915 or stock boost of 980. Which is really weird because I'm not doing any kind of tweaking or overclocking of the card. My "base boost clock" is 1058. I wasn't even aware there were different 660 ti's from asus until just now, but the rest are lower. Even with factory settings the clock goes up and down on its own constantly, although always the same amount. For example I have my core clock set to 1140 but it'll go up to 1202 all the time. Edit: Haha I didn't know there were 7 variations. If your core clock is 1084 though that is the "best" one Ignoarints fucked around with this message at 18:10 on Jan 31, 2014 |
|
#
?
Jan 31, 2014 17:59
|
|
|
I'm curious, what's the voltage on your GPU under load? Mine is 1.175V under load. Edit: And the 1084Mhz on my card is only the boost clock and it goes back down to 1071 or even 1058 after 5-10 minutes of heavy load on the GPU.
|
|
#
?
Jan 31, 2014 18:21
|
|
|
spasticColon posted:I'm curious, what's the voltage on your GPU under load? Mine is 1.175V under load. Sorry I'm at work I don't remember. I'll know a little later today if you're interested ill post it here, along with bios mod results.
|
|
#
?
Jan 31, 2014 18:27
|
|
|
I might go ahead and overclock the VRAM on my 660Ti in order to compensate for the gimped memory bus. But the individual VRAM chips on my video card don't have a heatsink or heatsinks so would I have to buy VRAM heatsinks before safely overclocking? I wouldn't overvolt I would just overclock until I hit the wall on the stock voltage.
|
|
#
?
Jan 31, 2014 21:18
|
|
|
So my old GTX580 began artefacting and then crashing during a game, then I couldn't even get through BIOS with white lines on everything and it BSODing. I bought a 770GTX for Saturday delivery... but I also put my 580GTX in the oven for a bit and now it's working again. Should I cancel the 770 order and hope this 580 lasts a while or just upgrade?
|
|
#
?
Jan 31, 2014 21:26
|
|
|
uhhhhahhhhohahhh posted:So my old GTX580 began artefacting and then crashing during a game, then I couldn't even get through BIOS with white lines on everything and it BSODing. I bought a 770GTX for Saturday delivery... but I also put my 580GTX in the oven for a bit and now it's working again. Should I cancel the 770 order and hope this 580 lasts a while or just upgrade? A mad man would say use the 580 for PhysX. But that 580 could fall over at any time, and the 770 will be a decent upgrade. Totally up to you whether you feel like saving the money until it falls over again.
|
|
#
?
Jan 31, 2014 22:16
|
|
|
Anecdotal evidence for the AMD / nVidia software bonfire: https://forums.geforce.com/default/topic/573811/nvidia-streamer-service-may-be-freezing-your-pc/ That started happening to me last week. On top of that GeForce Experience itself often took up most of a core when nothing else was going on, removed that piece of poo poo for now.
|
|
#
?
Jan 31, 2014 22:38
|
|
|
|
| # ? May 30, 2024 19:16 |
|
|
What is "the AMD/nvidia software bonfire" and where does AMD come into play via that link? edit: I think I get it, you are saying burn the extra software that comes with graphics drivers in general, in which case I agree. beejay fucked around with this message at 23:07 on Jan 31, 2014 |
|
#
?
Jan 31, 2014 23:00
|
|
