AMD Ryzen Fabric Clock 1467Mhz (DDR4 2933Mhz) vs Fabric Clock 1933Mhz (DDR4 3866Mhz)in couple games

EDIT: Jan 22 2020 Added 1920x1080 for those who thought first test was useless at 2560x1440

BF5 and GTAV






Just want to say all results are within margin of error and are the same for both resolutions 1920x1080 and 2560x1440 and will be the same for 98% of the PC games people play.

Want the best out of your system ram,fine tune Ram .I do not play at 1280x720 or 1920x1080. If you have normal gear tune your ram and put cash toward better GPU.

Tested a old game and new game at 1920x1080 and 2560 x1440 on AMD Ryzen 3800X.

The test is to show that tuned timings for your DDR4 Ram is more important than actual fabric clock on AMD Ryzen.

So for the people building a Ryzen system and can google or have a little knowledge, it may be advisable to pick a good value ram and tune up that ram or for the people with DDR4 3000Mhz ram there is no need to waste money buying DDR4 3800Mhz ram,thinking your gaming on AMD Ryzen system will be better.

I just could not be bothered testing a bunch of games ,as I have done this many times and results are always the same.

So I set CL12 DDR4 2933 Mhz vs CL16 DDR4 3866Mhz both have tight ram timings .
The memory clock (mclk), the memory controller clock (uclk), and the infinity fabric clock (fclk) are all synced for best results 1:1:1.
You can also run Ram at 1:1:2 and still have great results.


Of course I do not care what you do with your ram,I am just trying to help out someone who needs the help by showing actual video and how it does work when PC gaming.

Since I am not using a capture card and just Nvidia shadowplay with all task running,outcome may be different than with a capture card.


Computer and video editor used

Spoiler

♦ CPU - AMD 3800X With MasterLiquid Lite ML240L RGB AIO
♦ GPU - Nvidia RTX 2080
♦ RAM - G.Skill Trident Z 16GB DDR4(F4-4000C18D-16GTZ) (2x8)
♦ Mobo - MSI X470 - Gaming Plus(E7B79AMS.AA0-AMD ComboPI1.0.0.3)
♦ SSD - NVME SSD Intel 512GB
♦ DSP - LG 27" 4K UHD 5ms GTG IPS LED FreeSync Gaming Monitor (27UD59P-B.AUS) - Black
♦ PSU - Antec High Current Pro 1200W

VIDEO INFORMATION
► FPS Monitoring : MSI Afterburner/RTSS
► Gameplay Recorder : Nvidia Shadowplay
► Edit Videos : VSDC Free Video Editor http://www.videosoftdev.com/

2560x1440 video only

Spoiler

RDR2
Sleeping dogs 4:20
Timings/Voltages 6:05

 

The results agree with everything I have read and experienced myself; tuning the ram improves cpu performance. The more the cpu is taken out of the equation, the less impact everything affecting cpu performance has. Maybe stating the obvious but feel a lot of people loose site of this. I went through testing cpu voltage offsets with everything else being equal. Outside of synthetic benchmarks, there was really no discernible difference in gaming. Why went back to leaving the cpu voltage on auto.

 

Well I can only give my opinions on the gaming and ram testing,what blows my mind is how people think how the Ryzen CPU/Memory works.
I agree with you and have read you went back to CPU auto settings, let the CPU do it's thing.AMD nailed it,yeah you can get a higher score in Cinebench but for someone like me,I only game.I just use enhance mode and normal Ram settings at 3600Mhz ,it is good as it gets.

Also the video,I posted on a couple forums to see what people say,it is the internet and some of the stuff said well.... Anyway I also did the same stupid benchmarks at 1920x1080 .I will post video for those who want to see it.

 

I'm yet to experience the fun of tweaking RAM with Ryzen 3x00 since I only got
Patriot Viper Steel DDR4-4400MHz 16GB (2x8GB) for now. Didn't got a new motherboard and CPU yet, can't decide what to get :-/

 

All depends on what you do on your computer,a nice 9900KS would go good with that RAM for gaming .
AMD Ryzen 3900X looks like a good CPU and is fine for gaming.Great to have choices these days.

 

Well, my budget is limited and I'm not sure if I want better CPU (8c) or faster GPU :-D and I'm stuck with this for over two weeks.

 

Ahhh you tricked me with the DDR4 4400Mhz.
Budget build 3600 and B450 ,buy used ,less than $200USD on buy and sell forums.

 

Well, I'm not planning to have so high frequency, but I read a lot about this Vipers and it should do 3733Mhz at 14-14-14 or 15-15-15 without much issue.
And since 3600CL14 TridentZ cost much more I got these ones
I will report my results once I got it all running.
Will probably get mobo > ASRock Fatal1ty B450 Gaming K4 with Ryzen 3600

 

Sounds great,I suggest looking closley at the BIOS release dates on the Motherboards if you are worried about that stuff in a timely manner.
https://www.hardwareluxx.de/communi...4-uefi-bios-agesa-Übersicht-17-01-19.1228903/

 

Thing is most of the motherboards Ive had seem to massively overvolt the CPU. Crosshair VI Hero X370 gives around 1.56v and Crosshair VIII Hero X570 gives about 1.58V.

My 3900X boosts to around 4.65GHz on a few cores/threads but its for literally nanoseconds and mostly hovers around the 4.35GHz mark whilst still vcore is super high, even in HWINFO and Ryzen Master which shouldn't be the case. We should have all programs show exact volts.

I have taken to setting a manual vcore of 1.35v and then letting the motherboard do the rest and still I am boosting to 4.575GHz (again for nanoseconds) and I can even do an all core 4.35GHz locked. But the good part is that temps are MUCH better, like 10C better!

With auto volts on a 360mm rad in push pull (6x noctua fans @ 1,350rpm locked for quietness) temps can reach 74C full load and around 69C gaming. Sure this isnt bad, but idle temps spike all over the place as voltages are consistently changed on these chips. Now with 1.35v manual volts, I see in multiple programs that voltages still go down when idle but never boost above 1.358v (LLC kicking in) but now full load temps are 69C (5C lower) and gaming never exceeds 58C!!! thats 16C difference and still getting basically the same performance. My Cinebench R20 score increased massively as all cores are locked to 4.35GHz whereas auto volts sees the CPU clock as low as 4.1GHz as the motherboard throws a ton of volts at the CPU thinking it can cope, the temps get too high and then it starts to downclock it self leading to lower scores. Single core testing with auto volts gets around 4.3-4.4GHz so I lost around 30points in single threaded. Which is not bad at all really.

Not to mention now my idle temps are more than under control as before they would spike up and down from 35C to 48C to 31C to 58C and so on. Remounting the cooler did nothing. But changing volts made a huge difference now volts stay around the 1.35v mark and still drop as low as 0.9v and idle temps are now in the 27C-36C range depending on other factors.

Planning on messing around with Game Mode, as this disables 1 CCX and basically turns a 3900X into a 3600X with high clockspeeds. Going to see if I can tweak it some more and get better game performance as I am into CEMU emulation and it hates multiple die CPU's like Ryzen so disabling a CCX and overclocking all cores could help massively. Might disable SMT too and just run 6 cores and see if I can hit 4.5GHz on all cores.

 

There was also an article on Tom's where using 4 ranks of memory (either 4 DIMMs or 2x 2 rank DIMMs) could increase performance from 5 to 10% in certain games.
In my case, I'm using 2 sets/kits of Trident Z RGB 3200MHZ memory (one has Hynix, the other Samsung B-die) and they're actually fine and stable together.

 

I was thinking about doing this yesterday but I may sell my 16GB 4000Mhz ram and pick up a cheap 32GB kit.I would like to test that few % in a couple games.

 

I do not let the temps jumping around like that get to me at all. I have seen more consistent behavior with the most current chipset driver and bios version 1004B. Vcore even appears to be less sensitive to negative offsets than before. So if temps are a concern look to setting a negative offset; it just becomes a balancing act between temps and what is considered an acceptable Vcore result. Before the last two chipset driver releases, I achieved the best results with a CPU negative offset of 0.06875. Have since reverted that setting back to auto. Heat is also why SOC has a negative offset of 0.03125 which HWiNFO shows as 1.050 to 1.056v. Through long and tedious testing, I achieved best results with CLDO VDDP at .875; anything higher or lower resulted in a performance reduction. Same type of thing with VDDG CCD at .850 and VDDG IO at .875. The DRAM Calculator suggested higher settings for all those voltages. Borderlands 3 is the only game I play that causes the cpu temp to peak at 73C. All other games, the peak is more like 64C.

 

Yea I might look into negative offsets, new chipset drivers do seem a bit better with temps tbh.

But I think setting a manual vcore of 1.35v is fine as vcore still drops when idle and I have seen ryzen master reporting 1.41v but once the CPU is fully loaded vcore settles at 1.33v and sticks to 4.35GHz.

 

Added BF5 1920x1080 for anyone interested

 

12/2933*2000 is 8.182 ns
16/3866*2000 is 8.277 ns so they are more or less the same. Secondary and tertiary timings also can play a role giving a slight edge to one or the other.
This is why speed alone does not mean much so a 6000 24 CL ram would be the same as a 3000 12 CL ram. Also try 1t command rate