B-Die Binning

[sabishiihito]

Flare X 3200C14 seems really strong.

 

4GHz 12-11-11-28-1T tight subs without a huge amount of voltage.

 

 

Soooooo close to 4133 12-11-11

 

[VenomousDesigns]

Tested 4266 rgb,failed to train 4000 12 12 12 1.85-2V. Also latest 4266 were worser on tight timings compared to 3600c16. Latest 3733 c17 and 3600c15 again were worser,1707 serial week,1701 on pcb,644 week on IC.

 

On the average from 5 sets of 3600C16 i had way better average than on 7 sets of 4266.Not to mention that 4266 i had 3 kits with one module DOA.

 

It is what it is , i binned over 100 kits overall and i have 3 modules super good,found about 6 GOOD(4133 12 11 11 Spi 32M wazza) and about 10 4000-4040 which is in my book average.

 

My advice would be to get a few kits of 3600 17-18-18 or 16 16 16 NON-RGB and keep the best,return the rest.

 

Sorry to be the deal-breaker for the RGB but the different PCB makes handling a bunny, for instance lots of 55 codes at reboot or cold-startup at the same time when they were memtest-stable,this is PCB to blaim,might be additional noise caused by the leds or motherboards that are not perfect tuned for this.

 

I stated my own experience , you can believe me or not,this is opinion based as i said 100 kits of b-die G.skill retail tested,completely waste of cash ,time and resources . While you might see other "retail" kits 4200+ superpi stable, you know unicorns do not exist and real-life experience is so harsh

 

Truly appreciate all the work you've done (and others!) with providing us novices the insight we would not otherwise have (or could afford haha!)!

 

I come here with a quick question regarding the purchasing of new RAM for my 8700k/Asrock K6 PC. I haven't bought a new PC in 6yrs and ALOT has changed - I've spent the last week or so trying to catch up with regards to single vs dual, Samsung Die-B and latency/MHz. I won't pretend I know anything as vast as the users her, but I'm hoping I've done some of the hardwork so you can help make the final decision.

 

I've narrowed it down to the following - all are Samsung IC:

 

1. Single-Rank, 4x8GB (32GB) @ 3200 CL14 - F4-3200C14Q-32GTZ

2. Single-Rank, 4x8GB (32GB) @ 3600 CL16 - F4-3600C16Q-32GTZ

3. Dual-Ranked, 2x16GB (32GB) @ 3200 CL14 - F4-3200C14D-32GTZ

4. Dual-Ranked, 2x16GB (32GB) @ 3400 CL16 - F4-3400C16D-32GTZ

 

I learnt from here to stay AWAY from RGB (I cancelled my original order). I understand (or have read) that Single-Rank will be quicker effectively, but at 4x8GB with strain on IMC the option to go 2x16 Dual-Ranked maybe be better?

 

Also the cost of 32GB of Single-Ranked is about $120 USD dearer as opposed to 2x16GB, so if the difference is negligible (I'm no extreme overclocker, I just play Games and the ocassional VM for work) I'll go with Dual-Ranked - unless you advise GREATLY against.

 

Thanks for your help and looking forward to your answers!

[DStealth]

I have this combo 8700k/Asrock K6 PC

For the moment Single-Rank, 4x8GB (32GB) are not booting with 1T over 3Ghz did't tried lower. Seems coffee lake IMC or Asrock BIOS related tried 1.0/1.1/and 1.11beta all acting the same. As for 2 dimms 4Ghz was also not bootable with 1T 3900 was max i achieved despite running these 4 modules on x299 same board Gaming K6 with 7800x up to 4200 15-15-15-28-280 1T in quad channel. So for the moment get Asus board Maximus if you want good memory OC on Z370 or wait for some fix from Asrock...tested very deeply RTLs IOLs etc and all(actually a couple compared to x299) voltages...this is the situation...Not good for memory overclocking especially B-dies...

[VenomousDesigns]

I have this combo 8700k/Asrock K6 PC

For the moment Single-Rank, 4x8GB (32GB) are not booting with 1T over 3Ghz did't tried lower. Seems coffee lake IMC or Asrock BIOS related tried 1.0/1.1/and 1.11beta all acting the same. As for 2 dimms 4Ghz was also not bootable with 1T 3900 was max i achieved despite running these 4 modules on x299 same board Gaming K6 with 7800x up to 4200 15-15-15-28-280 1T in quad channel. So for the moment get Asus board Maximus if you want good memory OC on Z370 or wait for some fix from Asrock...tested very deeply RTLs IOLs etc and all(actually a couple compared to x299) voltages...this is the situation...Not good for memory overclocking especially B-dies...

 

Thank you for taking to respond - some worrying news there . I ended up ordering the 2x16GB (32GB) @ 3200 CL14 - F4-3200C14D-32GTZKW kit but now I'm worried even more since its dual-ranked!

 

Will I at least be able to run it at the stock speeds or will we run into issues on the Asrock boards?

 

This is my first time with a new build in 6yrs so its all very new to me - unfortunately in my country the Asus Hero very expensive compared to the Asrock K6 ($150USD MORE). Will Asrock come through eventually with a 'fix' - this is my first Asrock product so not sure about their track record.

 

With the kit I bought above (F4-3200C14D-32GTZKW) will this be OK for 'lite' overclocking or should I cancel the order and choose another item from the list above?

 

Thxs!

[speed.fastest]

Using 4 Stick Memory better use Asus Maximus with 4 DIMM, because it using T-Topology all slot can clocking almost same, 4 DIMM board that not using that topology will have hard time to boot.

[DStealth]

Can some somebody give me directions...What I'm doing wrong on the Maximus X Hero (a great board anyway)

Moved from Asrock k6 z370 where my 4dimms kit was hitting 3800-3900 14-14-14-28-280 2T for benchmarks.

On the Asus right now can run 1T and 2T with these speeds but nothing less than cl16-16-16-xx cannot boot performance is fine as RTLs are lower but...why is this difference as on x299 this kit was able to bench same voltages(sub 1.5v) up to 4100 14-14-14-28-290 1T in quad channel ...

8700k IMC is different but the CPU is better in general cannot blame it for such difference. VCCIO and VCCSA tried even more than 1.3v the same. Cannot boot and 55 or 4F error codes are staying for very long time.

 

Edit: MRC is disabled, tried skew control ... on the latest and previous BIOS original and modified from Raja all the same

Edited December 8, 2017 by DStealth

[Splave]

^ Think you should just pop out two sticks when you bench 32m

[yosarianilives]

Can some somebody give me directions...What I'm doing wrong on the Maximus X Hero (a great board anyway)

Moved from Asrock k6 z370 where my 4dimms kit was hitting 3800-3900 14-14-14-28-280 2T for benchmarks.

On the Asus right now can run 1T and 2T with these speeds but nothing less than cl16-16-16-xx cannot boot performance is fine as RTLs are lower but...why is this difference as on x299 this kit was able to bench same voltages(sub 1.5v) up to 4100 14-14-14-28-290 1T in quad channel ...

8700k IMC is different but the CPU is better in general cannot blame it for such difference. VCCIO and VCCSA tried even more than 1.3v the same. Cannot boot and 55 or 4F error codes are staying for very long time.

 

Edit: MRC is disabled, tried skew control ... on the latest and previous BIOS original and modified from Raja all the same

one single rank dimm per channel is best. Good bdie is IMC limited not ram limited

[websmile]

Vary twcl and also check the four trdwr values, they are sometimes set weird since skylake days and kill certain cas values especially on asus at full slots

[DStealth]

Seems in order to me will try some additional values for them thanks websmile

 

Edit: Looks like only CL is affected as other valuse can be reduced w/o issues but lowering to cl15 the boot hangs on code 49

Get to 16-15-15-30-300 1T settings w/o problems

Edited December 9, 2017 by DStealth

[websmile]

You can try cas15 twcl 16/11/9 and maybe trdwr 15, but this is only based on Z170, on Z370 I only have a two dimm board which makes 4x tests obsolete^^

[DStealth]

Vielen Dank ...

Going to twcl 16 let me use cl15 instead of Auto twcl 15 and not getting cl 15

Gonna test further

[DStealth]

Just pushed to 4g 15-14-14-30-300 1T with 4 dimms TWCL was the main issue it seems 35ns on this platform seems fine...Just need time to bench this combo

Thanks once again .

[sabishiihito]

There has to be some logical reason why Asus boards hate B-Die with this particular PCB layout type:

 

 

I have three kits with that layout, one 2x8GB Patriot Viper 4 3600C16, one 4x8GB TridentZ RGB 3600C17 and one 4x8GB TridentZ 3200C14. On any Asus board, 3866C12 is the max I can boot 1T and even that is a struggle. On the other hand, Asrock X299 OC Formula or MSI X299 Xpower with KBL-X can boot 4000C12 1T tight subs with all of them except some particularly weak sticks in the bunch.

[websmile]

To kill your pattern, my 3866 TrodentZ RGB had same layout and could do 4133 12-11 gb/xtu for me on mocf and on a very fast test bootet and ran 4000 12-11 waza on an outdated impact8 with kbl

Edited December 11, 2017 by websmile

[websmile]

[sabishiihito]

Not surprised the MOCF handled them well, but the Impact is a surprise. You didn't have to do any crazy tweaks? Maximus IX Apex and Rampage V Apex with KBL-X hate all my sticks with that layout.

[speed.fastest]

I can boot 4000 12-12-12 and do geekbench on Asus Maximus VIII Gene with Skylake with that layout, there is some pros & cons with that lay out, that lay out seems the best at memory clocker.

[sabishiihito]

 

What were your subs on that run?

[websmile]

I do not remember rtls, but I used a normal test pattern for subs like at this screen

[sabishiihito]

I do not remember rtls, but I used a normal test pattern for subs like at this screen

 

I pulled out my Impact and tried testing these problematic sticks and I can't get it to boot higher than 3866. Plopped them in the X299 OCF+7740x and boot 4GHz with a bit of fuss. *shrugs* I don't know what to say.

[sabishiihito]

Can we talk more about the different PCB layouts for B-die kits?

 

The top as we all know works on everything for stuff like 4KC12, but the bottom is very picky.  I've found that Maximus VIII Impact, MSI Z270i Gaming Pro Carbon, MSI X299 Xpower and Asrock X299 OC Formula can handle them well.  It seems a lot of kits these days are getting the bottom layout even for non-RGB modules and it's a shame we can't clock them on stuff like the Apex boards higher than 3733-3866C12.  I wonder if a UEFI update for those boards could fix the compatibility or if it's simply an issue with the physical DRAM topology.

 

Original link for photo:  https://clockemup.com/samsung-bdie_4800mhz/

Edited August 26, 2018 by sabishiihito

[yosarianilives]

I've always understood that the "rgb layout"  was optimal for max mhz but couldn't do good timings. I noticed on both my apex and z270 gaming that with my 4266 rgb kit only certain timings would stick, so I could run 4000 19-12-12-28 2t with tcl and tcr sticking high but I could bench this all day even with 4 dimms. At least that was my findings.

[sabishiihito]

I managed 4K 12-11-11 1T tight subs with a Trident Z 3200C14 kit that uses that newer layout on MSI Z270i board, so it doesn't seem to be chipset/platform limitation but rather mobo-specific.

Edited August 27, 2018 by sabishiihito

[sabishiihito]

3600C15 on paper sounds like a super bin but I got two kits and they fail 32M instantly 4K 12-11-11 no matter how I tweak volts.