Yeah it's unbelievable how much of a performance difference it makes for the Witcher 3. For me, all micro-stuttering is finally gone. Especially in cities and large towns walking/running around among large crowds was frustrating at times; and it really put you off the game. This no longer is the case. I noticed performance increases in GTA 5 (ever so slightly as reported by Guru3D). Running multiple benchmarks I noticed a small percentage of improvement in all of them (I ran 2-3 hours of benchmarking before updating from 15.11.1 to Crimson), but the most impressive thing is that wherever micro-stuttering seemed to exist, it has now either decreased significantly or vanished completely.

Yeah stuttering has definitively been reduced, framerate in Fallout 4, AC: Syndicate and Batman: Arkham Knight are about the same but they feel smoother with less hitching or how it's called, "juddering" comes up as a term every now and again which I at first thought was a typo of "stuttering" but apparently it's an actual term for something. (AC: Syndicate did unfortunately black screen crash when I tried to downsample it so for now it's stock 1920x1200 but eh aliasing isn't too big a deal for me when playing the game though it can be annoying if you try to take some screenshots of it.)

This review feels kind of incomplete and rather strange not having the fury X...

TDR is a classic for the nvidia drivers. Hence the reason for me to never buy any graphics card that company. In the era of fermi I felt the constant Tdrs BSODs twice. I had to sell my graphics card and never had a problem, since I amd, incidentally always had cards from ATI/AMD. For me the drivers AMD are by far much better and more stable. I think a grace people who continue to say that the nvidia drivers are better. This is just my opinion.. 🙂

Wow man, that was unnecessary. :X

I know they are still broken. 🙁

I know they are still broken. 🙁

I meant your post, meaning you should probably calm down.

This review feels kind of incomplete and rather strange not having the fury X...

I might look into FIJI perf as well, though 380X is GCN based hence the relative perf increment will remain the same at 2~3% The reason I opted the 380X is that it is the more affordable card with 4 GB, hence way more people will buy a 250 bucks card like the 380X opposed to spending the 700 Euro you need to spend on a Fury X here in the Netherlands.

I might look into FIJI perf as well, though 380X is GCN based hence the relative perf increment will remain the same at 2~3% The reason I opted the 380X is that it is the more affordable card with 4 GB, hence way more people will buy a 250 bucks card like the 380X opposed to spending the 700 Euro you need to spend on a Fury X here in the Netherlands.

if i may link it again http://www.overclock3d.net/reviews/gpu_displays/amd_radeon_software_crimson_-_overview_and_testing/2

I'm not sure how much this relates to Windows, but framerate can be a pretty gray topic. The frames rendered isn't the same thing as frames displayed (for example, if your monitor is 60Hz, you can get a framerate beyond 61FPS but you're not going to see it). Catalyst on Linux has been notoriously bad because you might measure 100FPS but looking at your display it seems to be more like 40FPS. The claims AMD made about framerate improvements might have more to do with the frames displayed. From what I heard, there was almost no measured performance difference in Linux at all but there was a significant visual improvement. Maybe Windows is the same way?

Thanks for the review, would have liked to see some frametime testing though considering it's the new flashy 3.0 upgrade. 😉

This driver is almost the same as 15.11.1. But this driver is more about stability in some Gw games,e.g The Witcher 3.

Running with the 15.11.1b Catalysts, W3 has never crashed for me--not once. Hard to see anything more stable than that... Speaking of Crimson, I have reverted to the 15.11.1b Catalysts upon the discovery that the Crimson drivers omit any control of resolution! Yup, with the Crimsons installed you must drop out of them and go to Win10 Display Properties in order to set screen resolution! It's been stripped from the Systray icon as well...Unreal. Grrrr--that is crazy. Screen resolution & refresh rates, rather than being completely unsupported in Crimson, should have been added to the per-game profiles just as fan speeds and overclocking were...! Aside from that, hotkeys don't function even though the ATi external events utility service is still being installed and is running. Let's see...oh, yea...couldn't make the custom res utility work...but ancient CRU in Win10 works just fine--even with the Crimsons! Honestly, this software has all the feel of being made for consoles--and then hastily retrofitted for PCs. Pretty sad. Hope they'll fix it soon. EDIT: OK workaround seems to work fine, instructions below for those that need them: install the 15.11.b's and then use the Device manager to install the 15.30xxxx Crimson drivers into the most recent CCC--instructions below for whomever might need them: Install the 15.11.1b cats, then open up the Device Manager select your GPU driver under "Display Adapter." Right-click on the name of your driver, then select "Update Driver Software" and then select "Browse My Computer," and then "Let me pick from a list of device drivers on my computer," and then select "Have Disk" and then point the program here: C\:AMD\Radeon-Crimson-15.11-Win10-64-bit\Packages\Drivers\Display\WT6A_INF\C02969567.inf and select Open--and the 15.30xxxx Crimson drivers will be installed into the 15.11.1 CCC *and* you will also get the new Custom Refresh Utility--assuming you can get it to work--as well. You should have all hotkey function back, etc. Everything seems to work fine--haven't tested everything, though. This gives you the Crimson drivers and solves all control problems with resolution switching & hotkeys, etc. Note: The specific driver numbers above may vary a bit from your own unless you have the same card that I do--an R9 380 4GB.

I might look into FIJI perf as well, though 380X is GCN based hence the relative perf increment will remain the same at 2~3% The reason I opted the 380X is that it is the more affordable card with 4 GB, hence way more people will buy a 250 bucks card like the 380X opposed to spending the 700 Euro you need to spend on a Fury X here in the Netherlands.

Be interesting to see if you get the core going up and down constantly, causing massive performance drops like quite a lot of us are getting, its no fun at all when your cards are dropping to around 350 odd all the time, your frames just plummet. 🙁

I might look into FIJI perf as well, though 380X is GCN based hence the relative perf increment will remain the same at 2~3% The reason I opted the 380X is that it is the more affordable card with 4 GB, hence way more people will buy a 250 bucks card like the 380X opposed to spending the 700 Euro you need to spend on a Fury X here in the Netherlands.

I agree, and this goes to a pet peeve of mine for the last 20-30 years (Sheesh! Am I that old already? Aghhhhh!) Sometimes videocard reviews seem tailored to show off cards other than those advertised, or do things like compare a $1500 card to a $600 card, or even, a $300 card...! My personal preference is that reviews cap the products tested against each other to the price of the card that is the subject of the review. That way it doesn't present the reader with statistics he doesn't need to see. In the reviews of flagship GPUs, they should be shown only against other flagship products--all of the approximate same price. If one card costs $1500, say, and the competition doesn't make a product that expensive, then imo such a review should contain performance data on the review product alone and nothing else (seeing that there is nothing else in its price range, etc.) And then, maybe at the end of the year, around Christmas, for instance, a hardware site could throw all of the review results for the entire year into one gigantic grab-bag bar chart...just for fun... 🤓

Maybe it would of been a better idea to have this driver ready for the Fury X launch. Would of probably made a bigger splash. Obviously game enhancements for newer games such as Fallout 4 would not have been present at that time but releasing such a dramatically new GUI with performance increases (3% is better than none) and improved features like frame pacing and finally shader cache. All this coupled with a side by side release of Fury X it would of had AMD in the headlines for months not weeks. Nvidia just released the 980Ti and totally stole the lime light away from AMD's Fury X in a heartbeat. At least with drivers like these people would of been talking and comparing them even more. If I was AMD at this moment I would just release the FURY X with a triple slot fan cooler and shave $50 off the price tag (maybe stretch to $100 if possible). By releasing new drivers and rereleasing your top tier GPU with a cooling solution that would fit into most cases (unlike a cooler that requires mounting) would draw more people in. In addition getting more game coupon bundles too would be awesome and just the icing on the cake.

I was hoping AMD would've integrated shadowplay like functionality in their new software. Shadowplay is what won me over to getting GTX 980s. I can do 2560x1440@60fps on to an SSD with very little CPU usage. Is the functionality even in the pipeline for AMD?

I was hoping AMD would've integrated shadowplay like functionality in their new software. Shadowplay is what won me over to getting GTX 980s. I can do 2560x1440@60fps on to an SSD with very little CPU usage. Is the functionality even in the pipeline for AMD?

You can use Raptr. I don't think they'll combine them.

Thank you very much for review! Small mistype: Page6 Conclusion: Fallout V -> Fallout4

Why isn't Fiji on the benchmarks? Not that I'm expecting a significant difference, but it makes absolutely no sense to exclude Fiji from this analysis...

My guess. Because this is a driver review, not a GPU review.

I have told all friend with A(md)TI that Crimson were better... but i had very bad return on game that is +/- 1year and older (all are on W10 64 and the hardware were R9 series (all exept fury) a R7 270X all on i3/i5): -In some Bioware and EA game it make all light fx deseapear (specialy in the new extension of SWTOR ) -random square black artefacts in wargaming's games and more in all the case going back to last beta solve everything (but add back the stall in witcher 3 lol) my point is it could have to do with the AA... but i am green right now, so cannot test.

True, 2fps with every driver adds up indeed. What nvidia have done lately besides those many TDR crashes? :P

Guess they have to start somewhere which is awesome. I'm building my brother an AMD based system with a 380x and get to finish it up next week. For people who have single card setups and have performance issues sounds like a serious problem. Issues like stuttering/jiddering sounds horrible. I will be able to see first hand how the AMD flag has been flying and can not wait. An issue I have is why when on a subject of one specific brand or topic people have to bring up some "other" companies possible shortcomings? Is it to make them feel better inside about their personal choices in life? Just saying...

Hope it gets better...