[GUIDE] Skylake Memory Timings on Asus Motherboards !

[GtiJason]

Wow. I've run upwards of 2.0V so far lol. Any scaling with more than 1.92V?

 

Wow, I've never been higher than 1.88v and never benched higher than 1.85v. I'm either lucky or my kits are ones that hate high volts, Im gonna stick with lucky

[speed.fastest]

Wow, nice combination G.SKILL and Corsair DC Kit

Yep you should be locking the rtl.

[sabishiihito]

Fixed RTLs.

 

[rtsurfer]

Should be 49/50 6/6 or 49/50 4/4 if going via offsets.

[xxbassplayerxx]

Aren't symmetrical RTL's optimal?

[rtsurfer]

Aren't symmetrical RTL's optimal?

 

Not as far as I know.

 

If he can get 49/50 working, it will be faster.

[speed.fastest]

32m i3 6100 4g/4g with E-Die @ 3750 11-18-18-28 300 1T

7m 25s 500ms

Anyone have a i3 or Pentium G 32m 4ghz reference? Cause i still dont know my efficiency is good or not. Still comparing with 6700K.

 

[sabishiihito]

M8E seems pretty useless for RAM OC especially compared to the Gene. I can boot and bench B-Die 4000C12 2T and AFR 3866C13 1T on Gene but Extreme always gives Code 41. Same timings set manually on both.

 

Sent from my SM-G928T using Tapatalk

[speed.fastest]

Larger motherboard size had longer circuit trace from cpu to memory, thats why Impact is one of the best memory tweaking performance, because it's 2 DIMMS and shorter circuit trace from CPU to Memory than Gene or Extreme.

[sabishiihito]

I know Impact is the best, but what I don't get is why the Gene is so much better than the Extreme, given they're both 4-DIMM boards with the same BIOS options.

 

Sent from my SM-G928T using Tapatalk

[speed.fastest]

As far as i know, the bigger and many feature to motherboard, will have more bad pcb circuit trace, maybe thats why gene is better on memory tweaking (because have better pcb circuit trace).

[speed.fastest]

Anyone experience the same with Gene? I need 2T and loose RTL for such memory frequency?

 

[xxbassplayerxx]

I don't think there are very many people pushing E-Die right now but with B-Die, 4-Dimm boards wall around 1900 12-12-12 1T. Above that they require 2T.

[speed.fastest]

So 2 dimm board is a must, my wall at 1T is at about 1980 12-19-19 1T. Thanks for the information.

[xxbassplayerxx]

:celebration:

[Alex@ro]

So 2 dimm board is a must, my wall at 1T is at about 1980 12-19-19 1T. Thanks for the information.

 

I think your wall is caused by the mems or IMC,4133 is doable on Gene and Extreme 1T,did it myself and also Websmile did that on Hero...

[rtsurfer]

Yeah, E-die doesnt need 2T.

 

Sent from my Nexus 5X using Tapatalk

[xxbassplayerxx]

[Guest invasmani]

Larger motherboard size had longer circuit trace from cpu to memory, thats why Impact is one of the best memory tweaking performance, because it's 2 DIMMS and shorter circuit trace from CPU to Memory than Gene or Extreme.

Sounds great in theory, but considering I managed to get rank #2 in MaxxMem global/WR with "air" cooling mind you with a ASrock Z170 Pro4S which is full ATX motherboard and am running quad channel DIMM's. So with that said I can't say that statement completely holds true.

 

I think your right in that longer circuit traces from CPU to memory makes a difference longer wires tend to mean more jitter and latency with less bandwidth. I think the biggest issue is the layout of the DIMM's themselves the closer they are located to the CPU the better and in the case of quad channel it's better to wire them side by side as opposed to using more wiring to run dual channels on each side of the CPU.

 

As far why one motherboard is better despite a identical appearing bios that's likely just due to better quality components on one versus the other it's pretty easy for the bios software to be identical otherwise. It's quite natural that better components can impact performance and stability I mean that's why liquid nitrogen tends to allow for better benchmarks results than air cooling with the same hardware less heat more stability.

 

On a side note I really want to beat l0ud_sil3nc3`s liquid nitrogen i7-6700K 6,159.6MHz MaxxMem score with my meager air cooled i3-6100 4,625.89MHz it would just be great on multiple levels quite frankly between the CPU the OC and the cooling itself. The fact that I got within 24 marks itself is remarkable to me actually plus it benched and display more quickly which is odd and makes me somewhat question how the results are measured in the first place.

[zeropluszero]

Lol. maxxmem might be the worlds worst benchmark.

[GtiJason]

Sounds great in theory, but considering I managed to get rank #2 in MaxxMem global/WR with "air" cooling mind you with a ASrock Z170 Pro4S which is full ATX motherboard and am running quad channel DIMM's. So with that said I can't say that statement completely holds true.

 

I think your right in that longer circuit traces from CPU to memory makes a difference longer wires tend to mean more jitter and latency with less bandwidth. I think the biggest issue is the layout of the DIMM's themselves the closer they are located to the CPU the better and in the case of quad channel it's better to wire them side by side as opposed to using more wiring to run dual channels on each side of the CPU.

 

As far why one motherboard is better despite a identical appearing bios that's likely just due to better quality components on one versus the other it's pretty easy for the bios software to be identical otherwise. It's quite natural that better components can impact performance and stability I mean that's why liquid nitrogen tends to allow for better benchmarks results than air cooling with the same hardware less heat more stability.

 

On a side note I really want to beat l0ud_sil3nc3`s liquid nitrogen i7-6700K 6,159.6MHz MaxxMem score with my meager air cooled i3-6100 4,625.89MHz it would just be great on multiple levels quite frankly between the CPU the OC and the cooling itself. The fact that I got within 24 marks itself is remarkable to me actually plus it benched and display more quickly which is odd and makes me somewhat question how the results are measured in the first place.

 

This is not really even a benchmark, there is a reason you didn't receive any global points for it, and it's not because you didn't include a single cpu-z tab on your screenshot

[Guest invasmani]

Only because there isn't a lulcatz benchmark that fills your screens with animated cat meme gifs til your PC reaches 100% CPU usage and records how long it takes...yeah some developer is going to copy that idea just you wait!

[zeropluszero]

That sounds like a good benchmark, can't wait!

[xxbassplayerxx]

Only because there isn't a lulcatz benchmark that fills your screens with animated cat meme gifs til your PC reaches 100% CPU usage and records how long it takes...yeah some developer is going to copy that idea just you wait!

 

It's a bad benchmark because it's basically broken. It scales massively with CPU speed, it performs better when you run with less cores and HT turned off, etc. People don't really run it anymore.

[Guest invasmani]

It's a bad benchmark because it's basically broken. It scales massively with CPU speed, it performs better when you run with less cores and HT turned off, etc. People don't really run it anymore.

I just beat a 6700K with HT disabled running 2 cores 2 threads overclocked on liquid nitrogen to 6159.6MHz with a i3-6100 with HT enabled 2 cores 4 threads overclocked to 4,801.16MHz on air by 31.4 Marks. It certainly seems like that invalidates a bit of your blanket statements.

 

I don't know as if it runs better with cores turned off versus on it's plausible though. You'll have less throttling to worry about that could skew results up to a point though that's true with any benchmark that's not overly multi-threaded as well.

 

http://hwbot.org/submission/3202227_invasmani_maxxmem_ddr4_sdram_3925.9_marks/