The gigabyte offer is like you are already screwed by nvidia. Then we turn around and say, please screw me again, i wasnt screwed enough the first time. Pay once to nvidia not enough, now ask me to pay more again!

pascal definitely

lel now i can start sacrificing playing medium because now working cannot afford 2 buy new GPU every 2 years already from 2 years change 4 years dang...other commitments 2 bare is the main reason...so my target is volta in 2 years time  

tbh both are similar, theyre some sort of adaptive v-sync. just that FreeSync leaves the GPU to do the processing compared to Nvidia's GSync module which is an external unit

prolly explains why Nvidia charges extra for the module

For FreeSync (and the Adaptive Sync standard as a whole), when a game renders at a frame rate above or below this VRR window, the V-Sync setting is enforced. That means on a 60 Hz panel, if your game runs at 70 FPS, then you will have the option to enable or disable V-Sync; you can either force a 60 FPS top limit or allow 70 FPS with screen tearing. If your game runs under the 40 Hz bottom limit, say at 30 FPS, you get the same option: V-Sync on or V-Sync off. With it off, you would get tearing but optimal input/display latency but with it off you would reintroduce frame judder when you cross between V-Sync steps.

And you'd be wrong, G-Sync does handle this better because at no point do you need to enable V-Sync to avoid tearing above max refresh, thus undoing much of the benefit of this tech to begin with.

At higher than monitor refresh rates, the monitor continues to update at its max refresh rate with no input lag, no tearing because the G-Sync module with the onboard DRAM (y'know, the same magic stuffs AMD and their fanboys thought was extraneous and unnecessary) actively compares, holds, and renders each frame using it as a lookaside buffer.

So, any frames above what the monitor can refresh, the G-Sync module holds, compares, and chooses to display new, hold or throws out.

I guess all that pricey, proprietary hardware was justified after all!

But yes it is nice to see AMD finally show us something with FreeSync beyond slidedecks, fixed refresh windmills and empty promises, but now they have shown it, they have also confirmed what many of us expected:

G-Sync isn't going anywhere, its better than FreeSync and wehatever premium Nvidia is charging will be justified.

G-Sync above max refresh has the option to hold a frame and wait for the next frame after monitor refresh, effectively reducing the actual frame rate output of the monitor, because the G-Sync module is dynamically changing and controlling the monitors refresh rate. Unlike typical refresh and V-sync, where the GPU is still slave to the monitor's refresh and the next frame must be rendered, regardless of how old it is, based on the timing of the monitor's refresh.

So in a worst case scenario, with uneven frametimes, you might see 6-7ms of input lag/latency on a 120Hz monitor (8.33ms between frame refresh).

the onboard DRAM on the G-Sync PCB acts as a lookaside buffer that allows the G-Sync module to compare, hold, and send frames to display. 768MB is capable of storing a whole lot of 4-20MB frames.

All that awesome knowledge Tom Petersen dumped on us in interviews on this very site starting to pay off!

Guess that G-Sync module and DRAM wasn't extraneous and unnecessary after all!

There are several PROBLEMS which G-Sync fixes.

Screen tearing - happens if you can't synch monitor and GPU (because normal monitors update at a FIXED interval)

LAG - VSYNC is used to fix screen tearing but this causes lag because there's a delay caused by buffering the GPU output to match the monitors update cycle

STUTTER - happens if you have VSYNC on but are outputting below the refresh rate

*So you can fix screen tearing by enabling VSYNC but then get lag (or lag and stutter) or disable VSYNC but get screen tearing.

G-SYNC fixes all these issues AT THE SAME TIME. The only real issue is staying above 30FPS which isn't a huge deal and even that will be fixed with newer PANELS (a panel limitation, not a G-Sync limitation).

**Above max?
I believe this is where G-SYNC is superior but it's hard to confirm. It's my understanding that G-Sync's added fast memory as a "lookaside" buffer allows G-Sync to stay enabled when the GPU can output more than the monitor's maximum.

Thus the monitor still updates ONLY as directed by the GPU software which keeps things SMOOTH. So you can basically stay capped at 60FPS on a 60Hz monitor this way.

FREE-SYNC however as a limitation of the monitor hardware (no proprietary hardware with lookaside buffer) seems forced to DISABLE the asynchronous method and go back to the normal fixed update by the monitor.

This is going to be a HUGE ISSUE for Free-Sync. So you have to stay in the RANGE to make it work (i.e. above 40FPS or below 60FPS). Not as big a deal for 30FPS up to 144FPS.

So basically G-SYNC seems to work almost perfectly and Free-Sync is problematic especially on 60Hz monitors.

(Worse is playing a game with 60FPS average on a 60Hz monitor. You'd keep going above and below the 60FPS mark meaning FREE-SYNC would turn on and off.)

Don't flame me if this is incorrect, but I've done a lot of reading and I think it's correct.

(Also not sure how AMD can fix this without introducing a proprietary scaler with lookaside buffer like NVidia's solution.)

Why you want to sell bro? Are u affected by the problem? What game you playing use more than 3.5GB?

If you want to play Shadow of Mordor on ultra then you’re going to need a pretty high end card as apparently there is a 6GB VRAM requirement, meaning that users may need a very specific GPU in order to play at full detail settings. To make things worse, that’s the recommended requirement for 1080p, rather than 4K, a resolution that you would expect to eat up that much VRAM.

Flagship cards from both Nvidia and AMD currently only ship with 4GB of VRAM, with the exception of the GTX Titans and some specific custom made GTX 980 and AMD R9 290x cards.

Conclusion

Our product reviews in the past few months and its conclusion are not any different opposed to everything that has happened in the past few days, the product still performans similar to what we have shown you as hey .. it is in fact the same product. The clusterfuck that Nvidia dropped here is simple, they have not informed the media or their customers about the memory partitioning and the challenges they face. Overall you will have a hard time pushing any card over 3.5 GB of graphics memory usage with any game unless you do some freaky stuff. The ones that do pass 3.5 GB mostly are poor console ports or situations where you game in Ultra HD or DSR Ultra HD rendering. In that situation I cannot guarantee that your overall experience will be trouble free, however we have a hard time detecting and replicating the stuttering issues some people have mentioned.

The Bottom line

Utilizing graphics memory after 3.5 GB can result into performance issues as the card needs to manage some really weird stuff in memory, it's nearly load-balancing. But fact remains it seems to be handling that well, it’s hard to detect and replicate oddities. If you unequivocally refuse to accept the situation at hand, you really should return your card and pick a Radeon R9 290X or GeForce GTX 980. However, if you decide to upgrade to a GTX 980, you will be spending more money and thus rewarding Nvidia for it. Until further notice our recommendation on the GeForce GTX 970 stands as it was, for the money it is an excellent performer. But it should have been called a 3.5 GB card with a 512MB L3 GDDR5 cache buffer.

The solution Nvidia pursued is complex and not rather graceful, IMHO. Nvidia needed to slow down the performance of the GeForce GTX 970, and the root cause of all this discussion was disabling that one  L2 cluster with it's ROPs. Nvidia also could have opted other solutions:

Release a 3GB card and disable the entire ROP/L2 and two 32-bit memory controller block. You'd have have a very nice 3GB card and people would have known what they actually purchased.
Even better, to divert the L2 cache issue, leave it enabled, leave the ROPS intact and if you need your product to perform worse to say the GTX 980, disable an extra cluster of shader processors, twelve instead of thirteen.
Simply enable twelve or thirteen shader clusters, lower voltages, and core/boost clock frequencies. Set a cap on voltage to limit overlclocking. Good for power efficiency as well.
We do hope to never ever see a graphics card being configured like this ever again as it would get toasted by the media, for what Nvidia did here. It’s simply not the right thing to do. Last note, right now Nvidia is in full damage control mode. We submitted questions on this topic early in the week towards Nvidia US, in specific Jonah Alben SVP of GPU Engineering. On Monday Nvidia suggested a phonecall with him, however due to appointments we asked for a QA session over email. To date he or anyone from the US HQ has not responded to these questions for Guru3D.com specifically. Really, to date we have yet to receive even a single word of information from Nvidia on this topic.

We slowly wonder though why certain US press is always so much prioritized and is cherry picked … Nvidia ?

970 is not highest end. Titan is. Never settle for less if you really serious into going ultra everything. better to collect money and get the best. Shadow of Mordor came with preset setting on different GFX. they auto detect best setting for it. If you change those means U are going beyond the card limit. Even played ultra for AC4 BF.

Back to topic of misinformation yes I believe NVIDIA is at fault for not doing proper test or intentionally misled consumer. I believe this case is as car recall they should just recall all the 970 and come out with new hardware config which reflect the 4gb.

Let me clearly state this, the GTX 970 is not an Ultra HD card, it has never been marketed as such and we never recommended even a GTX 980 for Ultra HD gaming either. So if you start looking at that resolution and zoom in, then of course you are bound to run into performance issues, but so does the GTX 980. These cards are still too weak for such a resolution combined with proper image quality settings. Remember, Ultra HD = 4x 1080P. Let me quote myself from my GTX 970 conclusions “it is a little beast for Full HD and WHQD gaming combined with the best image quality settings”, and within that context I really think it is valid to stick to a maximum of 2560x1440 as 1080P and 1440P are is the real domain for these cards. Face it, if you planned to game at Ultra HD, you would not buy a GeForce GTX 970.

My EVGA GTX980 also shows a slowdown on the Nai benchmark:
RUN 1 RUN 2
I've not noticed any frame spikes when going over 3.5GB in games but I've mainly been playing DA:I which seems to cap out at 3GB. I just tried SoM (1920x1200 with everything on Ultra including textures) and it mainly hovered around 3.3GB but there were no obvious hitches when it occasionally hit 4GB.

yep, pretty much sums up similar to what ive read so far

GSync is superior due to the module itself

well, rmb what they said. high quality IPS panel dont exactly come cheap

not to mention the response rate is unacceptable for gamers

frankly, GSync to me is good. but for a very specific purpose

wont be getting it at all, unless the monitor+module itself is affordable

4k tv is getting cheaper and cheaper

But 1440p and 4k monitor is expensive zzzz

Moogle Stiltzkin

Nvidia Can bring down the cost of Gsync Module by finalizing it to ASIC rather than FPGA

dude gsync is working right now and optimal fully functioning...

If u were Nvidia ... just have to see how freesync works with a mediocre AMD driver team.

Just waiting for AMD as usual if there is a issue to blame the scaler vendors.

AMD likes to show off brute gpu power where Nvidia try to reduce power consumption, trust me my electricity bill is different when i started using gtx970

previously i always use ordered dithering but madvr crashes. this issue once a while happened suddenly.

now i using error diffusion - option 1
used colored noise - ticked
change dither for every frame - ticked

bro, what kinda bug u mean?

i watching .mkv 720p H.265 don't see vertical lines uisng madVR v0.87.13 + MPC-HC.1.7.8.20.x86 + LAVFilters-0.63.0-75 everything 32bit

let me test few more times. later i feedback to u.

I'm using option 1, no vertical scan line bug for me on any source (driver 347.25 and .52). Be it 8bit or 10bit.

cyberbeing-
I just received a response from NVIDIA.

They were able to reproduce the issue, and have assigned the bug to their development team for investigation.

No, not the same. Please don't use Niyawa's config anymore, it's outdated and his settings doesn't work correctly after Madvr ver 0.87.10.

Just use the latest KCP beta, enable the highest preset option and then change dithering to option 1. Also, define your monitor as an 8 bit panel under Devices > properties.

Alternatively use - https://imouto.my/madvr/
Though I don't like that he uses Lanczos for chroma and image upscaling. If you follow his guide, make sure you change it to Jinc (3 taps).

I have SLI 970s. But previously on a 560Ti, I did not have this bug either.

Just use the latest KCP beta, enable the highest preset option and then change dithering to option 1. Also, define your monitor as an 8 bit panel under Devices > properties.

I'm also using a "10-bit" monitor, but madvr only recognizes 8bit (despite it saying "& above"). It doesn't matter though, as long as you set it to 8-bit then it'll work fine.

Yes, switch to KCP. Like I said, his settings does not work anymore for both madvr and even reclock (try using AC3 encoder to real time encode all your source to DD 5.1 with his reclock settings, you'll either get sync delays or no sound). madvr had A LOT of changes going from 0.86.x to 0.87. Previous settings are pretty much void by now.

*EDIT*
I used MPC-BE for a while, it's okay. But one thing I didn't like was that it had a bug where if I switched to my TV, the player will somehow cause massive framerate drops WITH the exact same settings. This is despite both displays being 8 bit, so the only difference is the size and panel, both of which doesn't do anything to change media player performance. I switch back to my monitor and wham, no frame drops. Sometimes the bug will stick until I restart the computer though.

You have to use the normal KCP version. The MPC-BE no longer works with changing audio in LAV Splitter since v1.4.3 when they forced a filter that's called "Audio Switcher" always on; When this filter is on the player will ignore those lav splitter settings

Just use the normal KCP version or go complain to MPC BE team. I don't care much for MPC BE.

I don't know if they have fixed this bug or not, I'll have to check. But I run my MPC pretty much without menus or borders, so having a black skin or not makes no difference to me.

KCP is basically just madvr + LAV filters + reclock + xy/vs sub rolled into one. It's made easy, their highest preset is higher than that of Niyawa's old highest madvr settings. So by default on their highest preset, it already handles low res content pretty well IMO.

I don't use ffdshow anymore, I think Jinc with 3 taps is pretty good already. NNEDI3 is even better but too bad until today, Nvidia cards cannot work properly with it.

i'm using 6+2 bit monitor. try https://github.com/nikola/htpc-updater/releases/tag/0.8.1

check whether got issue. my suggestion is delete everything & install again. i tried b4 it works.

followed this method >>> http://wiki.mikejung.biz/MadVR

maybe ur gpu suitable option "2"

MadVR Dithering and Error Diffusion

MadVR has added a few new dithering options. Error Diffusion Option 1 and Error Diffusion Option 2. Both of these new options are very demanding, and require a DX11 GPU to use. If you stick with the ordered or random options you will save yourself about 4ms of render time. If you have a GTX 970 or better then I highly recommend enabling Error Diffusion Option 1

I tested both of the options under the main dither choices and there is very little performance gain / loss if you enable or disable the options. From what I could tell there was maybe a 1ms gain / loss for render time if you check the options so only enable them if you feel like the video looks better. I have use colored noise and change dither for every frame enabled.

The GTX 970 can handle Error Diffusion Option 2 without any issues. Even combined with Smooth Motion the render times are under 15ms for 1080p content.

I tested both of the options under the main dither menu and there is very little performance gain / loss between Error Diffusion 1 and Error Diffusion 2, at least if you have a newer Nvidia GTX 970 or GTX 960. You must have a DX11 GPU to enable the highest two options, so if you have an older or not very powerful GPU, stick with either ordered or random dithering.

I chose to go with Error Diffusion Option 2 for Dithering, with both of the options enabled. Personally I like option 2 over option 1, but everyone has their own preferences, so choose whatever option looks better to you. In terms of performance, the fastest option is to use "None", Random, Ordered, Error Diffusion Option 2, and finally Error Diffusion Option 1. If you have a higher end GPU like the GTX 970, GTX 980, or any kind of SLI setup then you should go with Error Diffusion Option 2.

When playing a 1080p video I noticed that Error Diffusion Option 1 cost about 1ms more than Error Diffusion Option. By that I mean the average frame render time was about 6.10ms with option 1 enabled, and 5.10ms with option 2 enabled. While 1ms may not seem like a lot, it can mean the difference between dropped frames, or no dropped frames, so keep that in mind if you start to notice slow playback.

With smooth motion in madVR there really is no need for reclock. Also it just adds one big headache to the whole thing.