To Ryzen reviewers, in case you read this, along with other games please benchmark cities skyline with 1million citizen, gta 5 heavy situation with lots of ai and destruction, arma 3(singleplayer and multiplayer different), battlefield 1, the division, witcher 3, ark survival and rust multiplayer with lots of people(rust may have 255+ players in a server...) total war warhammer with 20000 units :P and maybe even some mmo like planetside 2 or guild wars/wow.

for vr titles- robot repair, adr1ft, project cars, nvidia funhouse etc.

gpus should be titan xp, titan xp sli, 1060gtx 480 etc. ; competition should include sandy bridge i5 i7 or older because many gamers want to upgrade from sandy bridge/ivy bridge.

coolers should be 212 evo as it's the most popular air cooler, water cooler h80 or h100i or cm seidon cheap water coolers and of course amd stock rgb. overclocking should be done with 70-90$ closed casing with moderate airflow. (although I don't think any reviewers have hands on am4 bracket for coolermaster devices. sad!)

maybe use heat detector gun to check temperature of VRM?

importantly it should have sata 3 m.2 ddr4 and usb3/3.1 pciE 3.0 performances as amd had severe trouble with these things in the past.

motherboard should be cheaper x370, expensive x370, cheaper b350 all of them.

xfr in depth.


and there is something else everybody is forgetting- igpu, although ryzen won't have igpu but try ryzen with 150$ or 200$ mobo, because intel is selling i5 as mainstream cpu for 7 years with minimal cpu improvement but with some gpu improvement, I want to see whether or not asus motherboards' mgpu comes close with intel's iGpu, if it comes close then we might have a real question here- what are they doing up there at intel hq with billions of dollar r&d budget every year?


maybe 'review' some motherboard security features as some motherboard comes with built in wifi/bluetooth?

EDIT:

if possible include unreal engine project kite, showdown vr, elemental demo and unity project adam, cryengine scene sponza, cryengine 5.0 gamesdk scene, some complex blender/3dsmax/maya scene etc.
Besiege with tons of big units in a scene, Space engineers with lots of planets and stuff....

EDIT2.1:

As adoredTv has said earlier, reviewers always miss game version numbers, and I can even name reviewers who miss dx version number or amd driver version lol.

Btw Linus Sebastian if you are listening, big fan, but with all due respect if you bench like only 3 games with this... :@ stop benching 4 games with 2 new gpus that's ABSURD. I am still mad at pascal vs polaris reviews back in June... 4 games... really.... reallyyyyyyyyy :@ :@

EDIT2.2:

I know reviewing hardware is tough job but I hope this times reviewers do a better job than they did in 8 months ago, I did not like pascal and polaris reviews, those data seemed untrustworty (from many reviewers). For example, if i remember correctly, Gamernexus found 1070 weaker than 980nonTi LOL. Although they later fixed that benchmark, they probably had problems with boost clock if i can remember properly. I did not like many other reviews either as some reviewers found higher fps in the same games (they used similar hardware and os) even with games with benchmarking tools, for example Guru3d benchmark average results are highers than almost all other reviewers, I liked techspot techpowerup h/w unboxed more. (PS: don't start the hardocp argument again, we already have many debates about hardocp already)

To make matters worse one youtube reviewer reviewed polaris with older driver but when the commenters pointed him out he posted bad replies which is kinda unacceptable for a youtube reviewer, another youtube reviewers with lots of sub and he is kinda famous in tech world, he posted a youtube review with some little misinformation but when commenters pointed them out he replied 'roughly' (passive aggressive mode). SAD!

Lots of mistakes have been done last year, Tech source found 1050Ti better than legendary r9 470, wut. Although later they fixed it. Lots of mistakes have been made folks I am telling you.

EDIT 2.3:

I mentioned some names here to make not to make another argument, I mentioned those names to make things better. We already had extensive arguments over those topics previously let's not start that again.
Morale of the story: Mistakes have been made, maybe lot more than usual last year, let's not stick to the past let's fix these stuff.


EDIT2.4:

Speaking of review quality and stuff, you may argue that nvidia gpu boost 3.0 technology is new and games might show different frame rates for different reviewers, it's kind of luck. Okay, but I remember Nvidia claiming 67c at 2.1 Ghz or something like that? Maybe I remember a wrong silde it's been 9months+ but still, reviewers have to do better fact checking.  http://www.hardocp.com/article/2016/05/18/geforce_gtx_1080_most_bizarre_secret_paper_launch_ever

Speaking of temperature and fan speed, change fan curve and noise test for air coolers and amd rgb cooler.

Also, check product longevity and advanced VRM stuff (I dont know how to do that ask actually hardcore overclocker), and bios stuff(xmp and other memory clock stuff).  Actually after nvidia Founder edition (read-founder edition uber crappy devices) you just cant trust no one.

EDIT 3:

Speaking of commenters, I can see some good comments here thanks guys. From comments-

Disable some cores in processor so we can assume performance of 1400x 1600x etc. , change clock speed if necessary.

STREAMING BENCHMARK, STREAM USING SHADOWPLAY/RELIVE AND external video capture hardware, streaming is really cpu intensive work. Also delay streaming by more than 5 mins, more than 5 minutes really one cpu intensive sh**. In some games like- Rust raiding party or ark, or Squad/Arma/Proj Reality or MMOs you need to delay streaming or the whole team will be screwed because of you. Total war warhammer streaming(with enough delay) with 20000 units on screen will kill your CPU unless you have 20 cores and 40 threads xD

Speaking of BIOS- overclocking some of the heaviest haswell E was difficult in some of the cheapest x99motherboards etc. So I guess in depth overclocking, bios profile, other bios stability stuff and xmp /xmp alternative ram timing is necessary.

Need FCAT or something like that... we like how pcper and hardocp does things with frametime curve and stuff...

Speaking of detailed stuff check power consumption properly, if possible check power consumption through 8 pin aux cable, don't forget reference 480rx is not technically a PCIE device anymore because of the pre bios update power consumption bug. #neverforget :P

BTW some motherboards seems to have 8pin+6pin whats that more power for better VRM stability? Those 6pins are also needed to start computer?

bios reset parameter- Which motherboards support backup bios which has no backup bios which has semi backup bios etc. Also which motherboard has feature to reset bios more easily and stuff.

I/O stuff- AMD used to have lots of problems with I/O speed so connect like hundreds of devices into the pc then check I/O speed.