[ManniX-ITA]

That looks exactly like my first experience with DRAM Calculator (about 5 hours ago.) In fact, your RAM is probably the same kit I have (the model number is the same at least.) Don't import the XML from Thaiphoon. It results in completely unrealistic numbers. Your kit won't do CL14 at 3600 (or probably any other speed.)

Run DRAM calculator, and set the following:

processor: ZEN 2 AM4/xTRX4.
Memory type: Hynix CJR/DJR // taken from the "DJR" in the DRAM Components string "H5AN8G8NDJR-TFC"
DRAM PCB rev: A0
Memory Rank: 2
Freq: 3600
BCLK: 100
DIMM Modules: 2
Motherboard: (depends on which m/b, but if a new build, it's probably a X570)

Then calculate safe. Take a picture and put those values into your BIOS. It should boot fine. Then do the same thing, but calculate FAST and try those settings. Still working? Then try to change the DRAM PCB to A3/A2/B2, calculate FAST, and try those new settings. (That's where I am now myself.) From there... I have no idea. (Trying to find out next steps was the reason I was reading this thread and found your post.)

I'd recommend to check first the actual PCB version.
You can take a picture and post it here.

 

[rdr09]

I just grabbed another Flare-X kit. It seems to be on A2 pcb, unlike my older flare-x which was A0. But thaiphoon still says A1 weird. Anyway i'm stuck at 3000CL14 4*8GB with 2700X.

current settings

anyhelp for 3400CL14 or lower but CL12 settings would be appreciated.

Here is info on my Flares. I think i used 1.7.2 Calc. DRAMVolt is set at 1.4v and VSOC at 1.08v.

 

[man from atlantis]

Here is info on my Flares. I think i used 1.7.2 Calc. Voltage is set at 1.4v.

Thanks rdr, but i'm looking for 4*Single Rank or 2*Dual Rank on Ryzen 2000s tight timings, preferably 4 sticks.

Quick update I improved the timings as it's stabelish (half an hour TestMem5 and half an hour Prime95 blender) 3000MHz CL12 at 1.42V vDIMM and 1.012V vSoC atm. If i somehow can manage to get 3200 CL12 with 4 sticks on a daisy chain mobo, I think I can beat my 3600CL14 2*SR settings. 3000MHz CL12 4*SR is already faster than 3200MHz CL12 2*SR according to my prelimary gaming benchmarks, which is around 5% slower than my 3600CL14 2*SR setting.

World War Z
https://www.reddit.com/r/Amd/comments/gpc2v3/2700x_memory_scaling_world_war_z/

 

[rdr09]

Thanks rdr, but i'm looking for 4*Single Rank or 2*Dual Rank on Ryzen 2000s tight timings, preferably 4 sticks.

Quick update I improved the timings as it's stabelish (half an hour TestMem5 and half an hour Prime95 blender) 3000MHz CL12 at 1.42V vDIMM and 1.012V vSoC atm. If i somehow can manage to get 3200 CL12 with 4 sticks on a daisy chain mobo, I think I can beat my 3600CL14 2*SR settings. 3000MHz CL12 4*SR is already faster than 3200MHz CL12 2*SR according to my prelimary gaming benchmarks, which is at most 5% slower than my 3600CL14 2*SR setting.

My bad. I may have to try that 3000 CL12, though. A test was made on lower oc but tight timings.

+Rep on the doing these tests.

 

[garyd9]

I'd recommend to check first the actual PCB version.
You can take a picture and post it here.

I'm not challenging your advice, but I'm curious: Why bother checking the PCB version physically? If the A2 PCB timings work with the RAM, does it matter what the reality of the PCB is?

When I was playing with it, I found that the suggested settings differences between the A0 fast and A2 fast for the hynix JFR are very similar, so I just decided to plug in the tighter numbers just to see what would happen. They worked, passed an overnight memtest86, etc. I still haven't bothered to pull the RAM back out of the machine to physically examine it.

 

[ManniX-ITA]

I'm not challenging your advice, but I'm curious: Why bother checking the PCB version physically? If the A2 PCB timings work with the RAM, does it matter what the reality of the PCB is?

When I was playing with it, I found that the suggested settings differences between the A0 fast and A2 fast for the hynix JFR are very similar, so I just decided to plug in the tighter numbers just to see what would happen. They worked, passed an overnight memtest86, etc. I still haven't bothered to pull the RAM back out of the machine to physically examine it.

No problem
Cause it's better to know for sure; it's not only about the timings but also the BUS/ProcODT settings.
The layout will determine what you need to set and if you don't know (like me) and ask then a different PCB will require a different answer.
Timings too; going down it's not necessarily better. You need the right timings and the PCB layout is a key factor.
Same module, same ICs but with a different layout could need or perform better with lower/higher timings.

If you are interested in the topic read the last 30-40 pages of this thread (thought this was another ).

 

[yrelbirb]

my friends 1usmus test gets stuck like this;

any reason why would it do it?

 

[Zektbach]

Hey guys is this any good?

Also I was playing warzone and streaming and I got blue screen error: faulty corrupted hardware page
G.Skill Ripjaws V 3200 MHz CL16-4-3200C16D-16GVKB

There's a page that says its Hynix-B die but Thaiphoon burner says Samsung B die.

 

[rdr09]

my friends 1usmus test gets stuck like this;

any reason why would it do it?

Same thing happened to mine and decided to just use HCI test.

 

[yrelbirb]

scams xd happened in turkey because of a retailer

hear me out: a retailer called teknolojipazari.com listed these rams for very good price;

and color us surprised, they were said to be samsung b-die...

well... 6 people in our local technology forum bough this kit hoping that it is B-die.

turns out it was a scam, the retailer send "JDC01" instead of "EDC01". which meant hynix afr and mixed hynix CJRs XD

had a great laugh, just wanted to share..

i hope they can refund it

im also having troubles refunding my teamforce rams myself...

 

[Tsk_Force]

Hello guys
I have these RAMs, I used DRAM calculator 1.7.3 to calculate the various timings. As motherboard I have an aorus master x570 and ryzen 7 3700x processor, I entered the parameters and some times the pc does not start, it remains in bootloop until the BIOS settings are automatically reset. now i am going with the xmp profile and it seems that everything is fine without problems. can you suggest me some changes? I don't really know where to start

I wanted to get a frequency around 3800 for RAM and infinity fabric at 1900 (if it holds) with good timings, or if you have anything better to suggest I listen to you

PS.
for I'm only focusing on RAM, the CPU is stock, I haven't even installed AMD overdrive

 

[mongoled]

my friends 1usmus test gets stuck like this;

any reason why would it do it?

For me this happens when you are not completely stable, instead of it erroring out it just stops using resources but the timer ticks away, so it is a bug that is instigated from non stable overclock

 

[Tobiman]

Hello guys
I have these RAMs, I used DRAM calculator 1.7.3 to calculate the various timings. As motherboard I have an aorus master x570 and ryzen 7 3700x processor, I entered the parameters and some times the pc does not start, it remains in bootloop until the BIOS settings are automatically reset. now i am going with the xmp profile and it seems that everything is fine without problems. can you suggest me some changes? I don't really know where to start

I wanted to get a frequency around 3800 for RAM and infinity fabric at 1900 (if it holds) with good timings, or if you have anything better to suggest I listen to you

PS.
for I'm only focusing on RAM, the CPU is stock, I haven't even installed AMD overdrive

Change DRAM PCB revision to A0/B0, try the values there and also pay attention to the voltage values given in the advanced tab.

 

[Martin778]

The voltages from the calculator tend to be incorrect. Not sure about the ODT and Clk settings as my TR rig didn't work with the values from the calculator but it works when left on auto (and it chooses differente values then).

 

[rdr09]

For me this happens when you are not completely stable, instead of it erroring out it just stops using resources but the timer ticks away, so it is a bug that is instigated from non stable overclock

My 3666 speed did that. I thought it was the app, so i switched to HCI and passed for 1.5 hrs run. Been using it for about a week now. Actually, using same settings i just raised it the speed to 3733MHz two days ago and been gaming on it. Went higher to 3800 but see no difference in performance just a high 1.45 voltage.

 

[Hale59]

This was my old GOC2017 @3733MHz (IMC couldn't do more), with imported XMP. Passed TM5 (20 cycles). No tweaking. Not bad at all.

 

[rdr09]

This was my old GOC2017 @3733MHz (IMC couldn't do more), with imported XMP. Passed TM5 (20 cicles). No tweaking. Not bad at all.

Hell yah that is nice.

 

[Dollar]

While pushing for a stable 1900 FCLK I ran into one corrected WHEA error.

"EventID 19

A corrected hardware error has occurred.

Reported by component: Processor Core
Error Source: Unknown Error Source
Error Type: Bus/Interconnect Error
Processor APIC ID: 0"

I noticed it in HWinfo after a 100 minute run of TM5 1usmus_v3 config that never showed any errors.


3700x
Asus Crosshair VI x370
4x8GB single rank samsung b-die a0 pcb
soc = 1.1v (no LLC so this droops to 1.081 during TM5)
cldo_vddg = 0.950
cldo_vddp = 0.900

Too low CLDO_VDDG? CLDO_VDDP? SOC? I can start the trial and error process but it would save a lot of time if someone has already gone through this.

 

[highdude702]

I have a question regarding the number of sticks and number of ranks.

I understand that if I have 2 sticks of single rank memory the settings need to be 2 sticks 1 rank in RTC.
However if I have 4 sticks of single rank memory I also know that if I'm on a dual channel setup I end up with a dual rank setup.
So in this instance would I set RTC as 4 sticks 2 ranks as I have been doing. Or would it be 4 sticks 1 rank? since theoretically if I had 4 sticks of 2 rank memory it would be quad rank. If someone that knows more than me about this could help me out that would be awesome! Thank you guys for the help and thank you 1usmus for making an awesome tool for the community!

 

[KedarWolf]

I have a question regarding the number of sticks and number of ranks.

I understand that if I have 2 sticks of single rank memory the settings need to be 2 sticks 1 rank in RTC.
However if I have 4 sticks of single rank memory I also know that if I'm on a dual channel setup I end up with a dual rank setup.
So in this instance would I set RTC as 4 sticks 2 ranks as I have been doing. Or would it be 4 sticks 1 rank? since theoretically if I had 4 sticks of 2 rank memory it would be quad rank. If someone that knows more than me about this could help me out that would be awesome! Thank you guys for the help and thank you 1usmus for making an awesome tool for the community!

I'm sure the DRAM Calculator you choose 4 sticks 1 rank as all four sticks are single rank.