Search before asking a question in the forum. The answer might be waiting for you. Feel free to blog about anything. For instructions on how to use the content creation tools, see the video on the left. If you are an author/ contributor to Tek Syndicate, hi. Log into the Staff Area
@Logan try benching witcher 3 with those settings and new drivers ~ you ought to get more fps from amd cards. (it has everything enabled and maxed out - I'm getting avg. of 55fps on single 290x at 1080p)
I felt I have been one of the only ones blown away by how much of a difference DSR makes over 1080P with filters, since no one else has been saying anything about it. I am glad you are spreading the word.
Along the line of DSR, it would be nice if you guys did benchmarks trying to max out the Vram and see how cards compare. In particular: the Furys, 980TIs, 390s, along with the 4GB higher end cards such as 980, and 970. I am willing to bet that, in some scenarios, the 390 and 390x might be coming up just behind the 980TIs, beating out some of the other cards due to having more than 4GB of ram.
Shadow of Mordor is an obvious choice for this, as well ask ARK: Survival evolved (although ARK uses lots of vram due in part to not being fully optimized yet).
When the 900 series launched, I felt that 4GB might have been a little lean for games a year or two down the road, and then DSR launched, and I was surprised Nvidia did not equip their latest cards to properly handle that feature. Would be nice to see if 4GB does cut it in most cases for the latest generation of games.
@Logan Something I have been meaning to bring up for awhile with Sapphire. I love when the card works 99% of the time, but when they don't, or have weird issues like mine. See below. Going through RMA, is not a pleasant experience.
Sapphire uses a separate company in the US, and if you look at that company's main page. It looks like it hasn't been updated since 1998 (http://www.althon.com/main.htm). My point is, this time around, I don't think I'll be going with Sapphire. I may opt for an Asus branded card or simply the water cooled version instead.
I guess if it were possible to get Ed's feedback what Sapphire is doing, to better the experience with their customers. I am not sure how I can get that feedback to Ed. Besides FB or YT. Just my thoughts. Don't know if it belongs here. I'll be looking forward to the head to head with the 980Ti.
I made a post that touched on the subject, and I know how you feel with discussing VSR/DSR. Even if it is not as good as "the real thing" it is still a great way to improve graphics. Even older 1680x1050 monitors get some new breath.
I just noticed that you jelp saying that the HBM is all on the right. I do believe those are the voltage regulators? the stacked HBM are those 4 little modules above and below the GPU. each one of those is 1GB of RAM.
Yeah, I screwed up there. I thought they were beside the core. They told me not to remove the heatsink because it was "complicated" so I didn't. I saw the images after shooting the video and simply forgot to edit that part out. We work way to fast with this stuff. We on't had this card for a couple days, so it's rapid mode... plus all the other work. Anyway, oops.
Honest mistake. I know you guys work hard... and with no AC? Dear god. Hope i helped. Looking forward to the 980/Fury battle. I am guessing you will be using the EVGA 980 with the ACX 2.0 cooler?
I will finally be buying my first Graphics card next month. You have me really excited to play around with VSR/DSR. No one else talks about that, they just post the same tests and the same results. I appreciate the unique content you guys creating for us. It is enriching.
One thing an amd representative/engineer said about the 4gb limit on the graphics card that I don't quite agree with. He was asked if this would perform worse than 6 or 8gb models and said no. The reasoning he had was that the HBM would act like a cache for the system memory to hold textures and easily swap missing data. While that is conceptually correct, I think of what seems to be the current setup for most pc games on the market.
Usually a game has a choice of low, medium, high, or ultra textures and each game seems to just dump all the textures into the graphics card for a given level/area. This design assumes you have all data needed on the graphics memory. When it actually has to fetch from system memory you can expect a significant stutter from the latency to call system memory. I would expect that was why you saw some the min FPS got into the single digits in 4k or graphics memory intensive games.
In the future I think this could get better but it would probably be through design changes in games and engines. I imagine something like what was done on 360 consoles when games could not require hard drive install. They would show low quality textures for a second or so until the high quality textures were read off disk.
I see logan said he no longer has the fury card anymore. I would like to see a review that explore what would explore what happens to the min FPS in the scenario where there seems to be some fallback to system memory. I think someone out there could do a test the card where they alter the speed of their system memory to compare 1333 to 2400 system memory speeds to find out if that makes a significant impact into min FPS of the card.
If you know of a review like that I would love to see it.