Micron Rev.E vs Rev.J & Hynix DJR - Overclock.net - An Overclocking Community

Forum Jump: 

Micron Rev.E vs Rev.J & Hynix DJR

Reply
 
Thread Tools
post #1 of 22 (permalink) Old 06-06-2020, 03:47 AM - Thread Starter
New to Overclock.net
 
Join Date: Feb 2017
Location: Sweden
Posts: 1,303
Rep: 57 (Unique: 43)
Micron Rev.E vs Rev.J & Hynix DJR

Want to make a short little write up on the experience I've gained from comparing these different Memory IC's I've attained and used in the same system.

Ryzen 3800X
Gigabyte X570 Aorus Xtreme

I've just recently attained 2x kits of Kingston HyperX Fury 2x8GB 3600Mhz CL17 HX436C17FB3K2/16
The kits came with different Memory IC's so I received 2x8GB Micron Rev.J, [week 15, 2020] 17nm & 2x8GB Hynix DJR, [week 4, 2020] 17nm according to [taiphoon burner].

From prior I've had 2x kits of Kingston HyperX Fury 2x8GB 3466Mhz CL19 HX434C19FB2K2/16
These came with Micron Rev.E 19nm [D9VPP], [week 14, 2018] I've been using these kits for 2 years.

All have A2 raw cards that are practically identical.

My Rev.E & Rev.J behave very much the same. In the end you could say the Rev.J is just a slightly better binned Rev.E. Voltage & timings match at every level but the Rev.J do a little better in regard to a few cases.
Rev.J do better tRCDRD, tRC & tRFC than the Rev.E I had.
Micron Rev.J do tRCDRD [21] tRC [60] tRFC [530] @ 3800Mhz vs tRCDRD [23] tRC [67] tRFC [600] for Rev.E
Hynix DJR do tRCDRD [21] tRC [58] tRFC [440] @ 3800Mhz
They then do everything the same really at this speed 3800Mhz & 1900fclk with similar voltage.

3800Mhz:
tCL 16 these ALL need ~1.450V with tight timings.
tCL 15 Rev.E & DJR needs ~1.500V with tight timings.
tCL 14 Rev.E & Rev.J need ~1.580V with tight timings.
tCL 13 Rev.E needs 1.700V with tight timings.

4800Mhz:
tCL 18 Rev.J needs ~1.600V
4866Mhz:
tCL 18 Rev.J needs ~1.650V

I used GeardownMode:disabled for tCL 13 & 15 on Rev.E & Hynix DJR only tCL 15.
GeardownMode:disabled can be a real hassle to manage on Micron IC's. They refuse to boot without everything being right.
For example Hynix DJR just works to GDM:disabled @ stock but will error on you until you raise clkDrvStr to at least 30Ohm, then it's a breeze to use.
The Micron Rev.E refused to work at all with GDM:disabled until I found these kits I had absolutely need clkDrvStr @ 60Ohm, otherwise it would not work.
The Micron Rev.J I've yet to found what they will need to do GDM:disabled. Have so far refused to do so on what I've tried. Will get stuck [07] on post.
[EDIT: Rev.J got GDM:disabled to work with CAD_BUS [20-20-24-24] on F20A AGESA ComboAM4v2PI 1.0.0.2]

So for those tCL15 results Hynix DJR was easy while it took me a long time to even manage to do GDM:disabled at all on Rev.E. Rev.J needed new AGESA version to do it.

Hynix DJR aren't as good as the Rev.J & Rev.E in regard to tRP. The Micron IC's could do tRP 17 with ease but the Hynix DJR struggled to do tRP 19 as I've yet explored.
But on the other hand they are more easy with tRCDRD even if it didn't get a better result than the Rev.J kit with tRCDRD 21 @ 3800Mhz it could boot lower and wasn't giving as much trouble at higher speed. The Rev.J & Rev.E just refuse to boot if you have it too low.

tRDRDSCL & tWRWRSCL these do 4/4 without trouble stock. For 2/2 results you need to go above 1.500V to be stable. 1.550V is usually enough if you go for tCL 14 on the Micron at the same time.
Hynix DJR can do SCL 2/2 @ 1.500V.

Not any of these kits have trouble with XMP. It just works, set it and forget it. You can even do the XMP profile @ 4000Mhz on all kits tested the little run I did. They need 1.350V though. My Rev.E is 1.200V stock while the newer kits are 1.350V. But at same voltage they all do 4000Mhz with XMP profile.

I was targeting 3800Mhz tight timings first and a little disappointed these kits are practically the same in the end. The Micron are so similar it's a laugh, Hynix DJR might be a little worse overall [edit: actually better] I was having trouble for it doing as tight tertiary timings as the Micron kits on Hynix. [EDIT: needed proper settings]

I'm still in the middle of testing and not found 1000% stable tight settings for all speeds with the new kits but got a feel on the ballpark on what they are doing now.

Hynix DJR behave better @ 3800Mhz in general but struggle to be stable ~4600Mhz in my testing thus far, can't seem to click the proper settings for them.

Micron Rev.J really like to go for frequency, have 4866Mhz basically stable now. Had more trouble with the board/IMC wanting to run it than the memory kits.
Attached Thumbnails
Click image for larger version

Name:	Rev.J taiphoon.png
Views:	26
Size:	40.5 KB
ID:	352364  

Click image for larger version

Name:	DJR taiphoon.png
Views:	10
Size:	43.2 KB
ID:	352862  

Click image for larger version

Name:	Micron Rev.J CL4-17-21-17-30-60-530 SCL2-2.png
Views:	19
Size:	92.7 KB
ID:	352868  

Click image for larger version

Name:	Hynix DJR CL15-8-21-19-30-58-440 SCL2-2.png
Views:	21
Size:	151.8 KB
ID:	352870  

Attached Images
 

Gigabyte X570 AORUS XTREME, Ryzen 7 3800X PBO boost, 2x8GB 4866/1900 18.27.24.48.90.tRFC 680 1T 1.650V [Micron Rev.J]
4x8GB 3800/1900 13.23.17.30.66.tRFC 600 1T 1.700V Kingston HX434C19FB2K2/16 3466C19 1.2V (Micron Rev.E 19nm [D9VPP]),
2x8GB Kingston HX436C17FB3K2/16 3600C17 1.35V (Micron Rev.J 17nm),
2x8GB 3866/1933 15.21.20.32.60.tRFC 448 1T 1.590V Kingston HX436C17FB3K2/16 3600C17 1.35V (Hynix DJR 17nm)

Last edited by Nighthog; 06-27-2020 at 02:26 AM.
Nighthog is offline  
Sponsored Links
Advertisement
 
post #2 of 22 (permalink) Old 06-06-2020, 06:00 AM - Thread Starter
New to Overclock.net
 
Join Date: Feb 2017
Location: Sweden
Posts: 1,303
Rep: 57 (Unique: 43)
Images of the kits:

HX436C17FB3K2/16
99U5702-066.A00LF is Hynix DJR 17nm
9905702-194.A00LF is Micron Rev.J 17nm

HX434C19FBK2K2/16 removed the heat spreaders for water cooling
9905678-153.A00LF is Micron Rev.E 19nm [D9VPP]

I don't know which number really identifies the type but I guess some of the numbers above should and is the only numbers different.
Attached Thumbnails
Click image for larger version

Name:	DSC_0058.JPG
Views:	13
Size:	2.40 MB
ID:	352386  

Click image for larger version

Name:	DSC_0065.JPG
Views:	37
Size:	2.35 MB
ID:	352392  

Click image for larger version

Name:	DSC_0096.JPG
Views:	13
Size:	2.13 MB
ID:	352396  

Click image for larger version

Name:	DSC_0097.JPG
Views:	13
Size:	2.25 MB
ID:	352398  


Gigabyte X570 AORUS XTREME, Ryzen 7 3800X PBO boost, 2x8GB 4866/1900 18.27.24.48.90.tRFC 680 1T 1.650V [Micron Rev.J]
4x8GB 3800/1900 13.23.17.30.66.tRFC 600 1T 1.700V Kingston HX434C19FB2K2/16 3466C19 1.2V (Micron Rev.E 19nm [D9VPP]),
2x8GB Kingston HX436C17FB3K2/16 3600C17 1.35V (Micron Rev.J 17nm),
2x8GB 3866/1933 15.21.20.32.60.tRFC 448 1T 1.590V Kingston HX436C17FB3K2/16 3600C17 1.35V (Hynix DJR 17nm)
Nighthog is offline  
post #3 of 22 (permalink) Old 06-06-2020, 02:21 PM
Overclock the World
 
Veii's Avatar
 
Join Date: Aug 2018
Location: Austria / Bulgaria
Posts: 1,686
Rep: 135 (Unique: 70)
Error 7 on TM5
"Simple Test , size 2"
Is a burst test
Meaning:
- Voltage is too high
- tRP is too slow to recharge in time
- transition on the recharge part is messed up

Try if you +1/-1 step change vDIMM on that micron kit, if the error dissapears
Please keep posting TM5 error screenshots, so we can know where to help

tRP on your side was always a bit too low ~ technically it's correct but its not optimal
If you go split tRCD, use the AVG tRCD delay between RD & WR as tRP delay
same also as tWRRD delay to look for

SiSandra Inter-Core Bandwith 14nm Ryzen global #1 | Twitter
Personal Records
RAM > Viper Steel @ 3466C14 (64.2ns) & Alt | Hynix-MFR 3333CL14 (72.5ns) & Alt

GPU > Gigabyte 970 G1 @ 1595|1900 (Hynix) | RX 580 Red Devil GS @ 1620|2200 (Micron)
CPU > TR 1950X [email protected]1.2v & [email protected]1.365v[105Mhz FSB](3.23Beta)
#Main: 1700X [email protected]1.365v1.73Ghz IMC CL14 (64.2ns)」+ Dark Rock 3 | X370 Taichi 1.0.0.3 ABBA by The Stilt | Viper Steel [PVS416G400C9K]| Gigabyte 970 G1 | Samsung 850 Evo
Veii is offline  
Sponsored Links
Advertisement
 
post #4 of 22 (permalink) Old 06-06-2020, 03:27 PM - Thread Starter
New to Overclock.net
 
Join Date: Feb 2017
Location: Sweden
Posts: 1,303
Rep: 57 (Unique: 43)
Quote: Originally Posted by Veii View Post
Error 7 on TM5
"Simple Test , size 2"
Is a burst test
Meaning:
- Voltage is too high
- tRP is too slow to recharge in time
- transition on the recharge part is messed up

Try if you +1/-1 step change vDIMM on that micron kit, if the error dissapears
Please keep posting TM5 error screenshots, so we can know where to help

tRP on your side was always a bit too low ~ technically it's correct but its not optimal
If you go split tRCD, use the AVG tRCD delay between RD & WR as tRP delay
same also as tWRRD delay to look for
I had just started to consider that might be it but I really want it to be able to do 17 honestly as my Rev.E could do it.
Been trying all kinds of tricks and things with the procODT, CAD_BUS & RZQ terminations to make it work. Not gotten a 20cycle pass to go yet but really close with 17cycles as best yet.

Need to get back on to test the Hynix DJR again. Didn't get good enough results on them the little time I put on them, neither tried all settings.
My 3800Mhz with tight timings is getting a little long to get stable on the Rev.J kit.

Currently trying to get this to pass: EDIT: still failed 10 cycles.
Attached Thumbnails
Click image for larger version

Name:	Micron Rev.J CL4-17-21-17-30-60-530 SCL2-2.png
Views:	22
Size:	92.7 KB
ID:	352532  


Gigabyte X570 AORUS XTREME, Ryzen 7 3800X PBO boost, 2x8GB 4866/1900 18.27.24.48.90.tRFC 680 1T 1.650V [Micron Rev.J]
4x8GB 3800/1900 13.23.17.30.66.tRFC 600 1T 1.700V Kingston HX434C19FB2K2/16 3466C19 1.2V (Micron Rev.E 19nm [D9VPP]),
2x8GB Kingston HX436C17FB3K2/16 3600C17 1.35V (Micron Rev.J 17nm),
2x8GB 3866/1933 15.21.20.32.60.tRFC 448 1T 1.590V Kingston HX436C17FB3K2/16 3600C17 1.35V (Hynix DJR 17nm)

Last edited by Nighthog; 06-06-2020 at 04:07 PM.
Nighthog is offline  
post #5 of 22 (permalink) Old 06-06-2020, 07:04 PM
Overclock the World
 
Veii's Avatar
 
Join Date: Aug 2018
Location: Austria / Bulgaria
Posts: 1,686
Rep: 135 (Unique: 70)
Quote: Originally Posted by Nighthog View Post
Currently trying to get this to pass: EDIT: still failed 10 cycles.
But how, why ?
With what error under which test

you'd likely need to increase tRP

SiSandra Inter-Core Bandwith 14nm Ryzen global #1 | Twitter
Personal Records
RAM > Viper Steel @ 3466C14 (64.2ns) & Alt | Hynix-MFR 3333CL14 (72.5ns) & Alt

GPU > Gigabyte 970 G1 @ 1595|1900 (Hynix) | RX 580 Red Devil GS @ 1620|2200 (Micron)
CPU > TR 1950X [email protected]1.2v & [email protected]1.365v[105Mhz FSB](3.23Beta)
#Main: 1700X [email protected]1.365v1.73Ghz IMC CL14 (64.2ns)」+ Dark Rock 3 | X370 Taichi 1.0.0.3 ABBA by The Stilt | Viper Steel [PVS416G400C9K]| Gigabyte 970 G1 | Samsung 850 Evo
Veii is offline  
post #6 of 22 (permalink) Old 06-07-2020, 03:09 AM - Thread Starter
New to Overclock.net
 
Join Date: Feb 2017
Location: Sweden
Posts: 1,303
Rep: 57 (Unique: 43)
Quote: Originally Posted by Veii View Post
But how, why ?
With what error under which test

you'd likely need to increase tRP
Was a voltage issue for my tCL 14 try. Needed a little more than I had wanted. I tried tRP 18 but still got issues so put a little more.

Finally a 20cycle pass!

Needed to set 1.580V in bios for ~1.600V HWiNFO reading.

procODT 32Ohm
CAD BUS: 30-20-20-20
Termination: [RZQ/7][RZQ/3][disabled] *might change these around
CL14-8-21-17-30-60 4-4-32-2-7-10 530 2-2 14-8 8-1 1-4-4-1-6-6-1 1T 3800Mhz 1900Fclk

VDDP: 900mv
VDDG_CCD: 950mv
VDDG_IOD: 1000mv
vSOC: ~1.100V AUTO LLC +0.0125V offset
Attached Thumbnails
Click image for larger version

Name:	Rev.J TestMem5 20cycle stable.png
Views:	50
Size:	1.75 MB
ID:	352648  

Click image for larger version

Name:	AIDA Rev.J 3800Mhz 14-21-17-30-60 530 1T.png
Views:	10
Size:	98.4 KB
ID:	352650  


Gigabyte X570 AORUS XTREME, Ryzen 7 3800X PBO boost, 2x8GB 4866/1900 18.27.24.48.90.tRFC 680 1T 1.650V [Micron Rev.J]
4x8GB 3800/1900 13.23.17.30.66.tRFC 600 1T 1.700V Kingston HX434C19FB2K2/16 3466C19 1.2V (Micron Rev.E 19nm [D9VPP]),
2x8GB Kingston HX436C17FB3K2/16 3600C17 1.35V (Micron Rev.J 17nm),
2x8GB 3866/1933 15.21.20.32.60.tRFC 448 1T 1.590V Kingston HX436C17FB3K2/16 3600C17 1.35V (Hynix DJR 17nm)

Last edited by Nighthog; 06-07-2020 at 03:13 AM.
Nighthog is offline  
post #7 of 22 (permalink) Old 06-07-2020, 10:11 AM - Thread Starter
New to Overclock.net
 
Join Date: Feb 2017
Location: Sweden
Posts: 1,303
Rep: 57 (Unique: 43)
A little tease on Frequency progress. 4733Mhz 4800Mhz

Voltages & tRCDRD were only real hindrances to get to 4733, after a little more finicky to get 4800Mhz to boot or not blue screen right away.

EDIT2: 4800Mhz stable base achieved!
Attached Thumbnails
Click image for larger version

Name:	Micron Rev.J 4733Mhz CL18-23-27-23-50-90 1.600V start.png
Views:	15
Size:	67.3 KB
ID:	352694  

Click image for larger version

Name:	Micron Rev.J 4800Mhz stable v1.png
Views:	323
Size:	398.8 KB
ID:	352748  


Gigabyte X570 AORUS XTREME, Ryzen 7 3800X PBO boost, 2x8GB 4866/1900 18.27.24.48.90.tRFC 680 1T 1.650V [Micron Rev.J]
4x8GB 3800/1900 13.23.17.30.66.tRFC 600 1T 1.700V Kingston HX434C19FB2K2/16 3466C19 1.2V (Micron Rev.E 19nm [D9VPP]),
2x8GB Kingston HX436C17FB3K2/16 3600C17 1.35V (Micron Rev.J 17nm),
2x8GB 3866/1933 15.21.20.32.60.tRFC 448 1T 1.590V Kingston HX436C17FB3K2/16 3600C17 1.35V (Hynix DJR 17nm)

Last edited by Nighthog; 06-07-2020 at 01:47 PM.
Nighthog is offline  
post #8 of 22 (permalink) Old 06-07-2020, 05:25 PM
New to Overclock.net
 
hazium233's Avatar
 
Join Date: Jul 2017
Posts: 177
Rep: 4 (Unique: 3)
Quote: Originally Posted by Nighthog View Post

My Rev.E & Rev.J behave very much the same. In the end you could say the Rev.J is just a slightly better binned Rev.E. Voltage & timings match at every level but the Rev.J do a little better in regard to a few cases.
Wonder if this is just a difference in Kingston's bin between those kits, or a characteristic of J generally. The ~280ns tRFC is interesting. If part number is decoded correctly for the J, they are -62E for the speed grade which means 0.62ns native, aka 3200MT/s at 1.2V with 22-22-22 timings. D9VPP is -75, or 2666MT/s native. Granted, with binning part of that can go out the window.

Any voltage scaling to speak of on tRCD with the J? Just hard limit at 21t for 3800?

Did you remove the heatspreaders on the new kits at all? I know that thaiphoon seems to often have trouble differentiating different variants of E, at least in Crucial sticks. Mostly it is getting the wrong speed bin (eg D9VPP instead of C9BJZ or C9BKV).

COL1
(3 items)
CPU
Ryzen 7 2700X
Motherboard
Asus Strix X370-F
RAM
Ballistix 3200
▲ hide details ▲
hazium233 is offline  
post #9 of 22 (permalink) Old 06-08-2020, 04:22 AM - Thread Starter
New to Overclock.net
 
Join Date: Feb 2017
Location: Sweden
Posts: 1,303
Rep: 57 (Unique: 43)
Quote: Originally Posted by hazium233 View Post
Wonder if this is just a difference in Kingston's bin between those kits, or a characteristic of J generally. The ~280ns tRFC is interesting. If part number is decoded correctly for the J, they are -62E for the speed grade which means 0.62ns native, aka 3200MT/s at 1.2V with 22-22-22 timings. D9VPP is -75, or 2666MT/s native. Granted, with binning part of that can go out the window.

Any voltage scaling to speak of on tRCD with the J? Just hard limit at 21t for 3800?

Did you remove the heatspreaders on the new kits at all? I know that thaiphoon seems to often have trouble differentiating different variants of E, at least in Crucial sticks. Mostly it is getting the wrong speed bin (eg D9VPP instead of C9BJZ or C9BKV).
No a hard limit @ 21 rRCDRD @ 3800. Would not boot @ 20 tRCDRD for Rev.J. My older Rev.E would not be stable at all @ 22 tRCDRD and bluescreen more often than not if I got it to boot. 23 just worked with GDM:enabled but GDM:disabled needed 24 for stability in a longer test.

These are not really binned for tight tRCDRD on these kits, they are really loose overall on the timings.

tRCDRD doesn't scale with voltage on Micron. You need a good bin from the start if you want it low. So if you can buy the kits with lowest tRCDRD from the outset.
I really wanted the 3733Mhz kits but they where at least 50% more expensive than the 3600Mhz kit. And they had the same XMP 3600Mhz profile inside them. So I just made a leap with the cheaper kit.

And I haven't removed the heatsinks yet, unsure if I will do. I don't know what I will do with the extra kits after testing is done. I bought them to spend time just testing stuff for curiosity. I might give away a kit to my niece later so I can't destroy that kit at least.

Gigabyte X570 AORUS XTREME, Ryzen 7 3800X PBO boost, 2x8GB 4866/1900 18.27.24.48.90.tRFC 680 1T 1.650V [Micron Rev.J]
4x8GB 3800/1900 13.23.17.30.66.tRFC 600 1T 1.700V Kingston HX434C19FB2K2/16 3466C19 1.2V (Micron Rev.E 19nm [D9VPP]),
2x8GB Kingston HX436C17FB3K2/16 3600C17 1.35V (Micron Rev.J 17nm),
2x8GB 3866/1933 15.21.20.32.60.tRFC 448 1T 1.590V Kingston HX436C17FB3K2/16 3600C17 1.35V (Hynix DJR 17nm)
Nighthog is offline  
post #10 of 22 (permalink) Old 06-08-2020, 07:27 AM
Overclock the World
 
Veii's Avatar
 
Join Date: Aug 2018
Location: Austria / Bulgaria
Posts: 1,686
Rep: 135 (Unique: 70)
Quote: Originally Posted by Nighthog View Post
A little tease on Frequency progress. 4733Mhz 4800Mhz

Voltages & tRCDRD were only real hindrances to get to 4733, after a little more finicky to get 4800Mhz to boot or not blue screen right away.

EDIT2: 4800Mhz stable base achieved!
Great Job !
Do you think you can try to torture it a bit more ?
Can you change tWTRL to 16 please
and try to get tRC 72 to run together with tRFC 648-481-296, and later maybe even down to 576-428-264 for 4800MT/s

Another stepup with a bit higher is tRC 90, there then either 720-535-329 or 672-499-307
all of them should work with tWR 24 only the tWTRL of 14 looks like an eyesore
Still great job !
The closest on 4800 of a full cycle would be ~ 640, 672, 704, 736
With whatever tRC you want to hit that tRFC

SiSandra Inter-Core Bandwith 14nm Ryzen global #1 | Twitter
Personal Records
RAM > Viper Steel @ 3466C14 (64.2ns) & Alt | Hynix-MFR 3333CL14 (72.5ns) & Alt

GPU > Gigabyte 970 G1 @ 1595|1900 (Hynix) | RX 580 Red Devil GS @ 1620|2200 (Micron)
CPU > TR 1950X [email protected]1.2v & [email protected]1.365v[105Mhz FSB](3.23Beta)
#Main: 1700X [email protected]1.365v1.73Ghz IMC CL14 (64.2ns)」+ Dark Rock 3 | X370 Taichi 1.0.0.3 ABBA by The Stilt | Viper Steel [PVS416G400C9K]| Gigabyte 970 G1 | Samsung 850 Evo
Veii is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off