Overclock.net › Forums › AMD › AMD Motherboards › ROG Crosshair VI overclocking thread
New Posts  All Forums:Forum Nav:

ROG Crosshair VI overclocking thread - Page 1283

post #12821 of 17471
Quote:
Originally Posted by 1TM1 View Post

Two interesting observations made during memory overclocking (to its rated speed of 3200).

First: I think motherboard checks only the module in A2 or if populated in A1 socket, then assumes the other one is identical.
This means that if you have two modules, one can be "better" than the other and by swapping your modules you get second chance at a faster speed.
If you have four modules then you have four chances. Rotate each module to the A1 socket and record the best memory speed you can get.

Second: It is important to do a ProcODT "sweep" to find the sweet spot for your particular memory.

[good stuff deleted for space]


cold boot (remove power plug) - CMOS loses the ProcODT setting. Default boot to 2133 memory speed. All other settings were retained in CMOS. After a reset to 43 Ohms BIOS ok
4000@1.368V; 3200@1.395V, SB 1.087V, 20(20)40 CB 1766 CPUZ 4973 - with four modules 32 GB Hynix M-die running at 3200.

Next is optimization of timings, voltages, etc. Hope this helps someone who is stuck at 2133 or 2666.

I think the bolded text may be a clue for the cold boot issue. Unless the BIOS default ProcODT resistance is close to what is needed, then losing the value upon booting certainly won't help the boot success.
Edited by kaseki - 4/30/17 at 5:03pm
     
CPUMotherboardGraphicsRAM
AMD Ryzen 1800X ASUS ROG CROSSHAIR VI HERO X370 Asus ROG STRIX nVidia GeForce GTX1080Ti-O11G G. Skill TridentZ F4-3200C14D-32GTZ 
Hard DriveOptical DriveCoolingOS
Samsung SSD 850 EVO 256GB Asus UltraDrive 8X Noctua NH-D15 SE-AM4  Linux Mint 18.1 MATE 64-bit  
MonitorKeyboardPowerCase
LG 4K TV OLED55B6P Logitech MK235 RF Wireless Keyboard & Mouse PC Power & Cooling Turbo-Cool 860i Nanoxia Project S 
Mouse
See Keyboard 
CPUGraphicsRAMHard Drive
147,456 cores all registers exposed at the console 4096 36-bit words rotary drum 
Optical DriveMonitor
6 magnetic tape drives neon lites for every register; oscilloscope for... 
  hide details  
Reply
     
CPUMotherboardGraphicsRAM
AMD Ryzen 1800X ASUS ROG CROSSHAIR VI HERO X370 Asus ROG STRIX nVidia GeForce GTX1080Ti-O11G G. Skill TridentZ F4-3200C14D-32GTZ 
Hard DriveOptical DriveCoolingOS
Samsung SSD 850 EVO 256GB Asus UltraDrive 8X Noctua NH-D15 SE-AM4  Linux Mint 18.1 MATE 64-bit  
MonitorKeyboardPowerCase
LG 4K TV OLED55B6P Logitech MK235 RF Wireless Keyboard & Mouse PC Power & Cooling Turbo-Cool 860i Nanoxia Project S 
Mouse
See Keyboard 
CPUGraphicsRAMHard Drive
147,456 cores all registers exposed at the console 4096 36-bit words rotary drum 
Optical DriveMonitor
6 magnetic tape drives neon lites for every register; oscilloscope for... 
  hide details  
Reply
post #12822 of 17471
Quote:
Originally Posted by hughjazz44 View Post

That doesn't require BIOS Flashback. Just use the EZ utility in the BIOS.

Im already on 1107 so flashback doesnt work on it from what ive seen on the thread, its flashback only.
    
CPUMotherboardGraphicsRAM
R7 1700X asus crosshair 6 Sapphire R9 390 Nitro CMK16GX4M2B3200C16 
Hard DriveHard DriveCoolingCooling
Kingston HyperX 3K Seagate ST2000DM006 EKWB Supremacy Evo EK-XRES 100 Revo D5 - Acetal 
CoolingCoolingCoolingCooling
Alphacool GPX r9 390 m01 alphacool xt45 240 Alphacool xt45 360 Alphacool D5 PWM 
OSMonitorKeyboardPower
Windows 10 LG 29UM68 Corsair Strafe Corsair RM1000 
CaseMouseMouse Pad
Thermaltake X5 Corsair M65 Pro RGB Razer Goliathus Speed 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
R7 1700X asus crosshair 6 Sapphire R9 390 Nitro CMK16GX4M2B3200C16 
Hard DriveHard DriveCoolingCooling
Kingston HyperX 3K Seagate ST2000DM006 EKWB Supremacy Evo EK-XRES 100 Revo D5 - Acetal 
CoolingCoolingCoolingCooling
Alphacool GPX r9 390 m01 alphacool xt45 240 Alphacool xt45 360 Alphacool D5 PWM 
OSMonitorKeyboardPower
Windows 10 LG 29UM68 Corsair Strafe Corsair RM1000 
CaseMouseMouse Pad
Thermaltake X5 Corsair M65 Pro RGB Razer Goliathus Speed 
  hide details  
Reply
post #12823 of 17471
Quote:
Originally Posted by bluej511 View Post

Im already on 1107 so flashback doesnt work on it from what ive seen on the thread, its flashback only.

Oh, I didn't know you already updated. I grabbed 0003 when it first came out, but then dropped back down to 0079, then went up to 1107.
post #12824 of 17471
Quote:
Originally Posted by gavinh87 View Post

Hi guys newbie here just been reading on these Gskill SPD issue and I think I may be affected although not as bad as some of you guys.
One stick shows its part number as "F4ä" The other reports it as F4-4266C19-8GTZR
SPD tables looks exactly the same. I would try to dump the "good" one onto the bad one. The issue here is that the "good" one has SPD Checksum (Bytes 80h-FDh): A01Ch (Error)

My sticks are not in the database yet so I cannot copy from a known good one. Am I looking at an RMA?

 

Gskill does not list the f4-4266c19-8gtzr number as a valid part number so that may be the reason the checksum fails. It is missing the character D (for dual stick kit) it should read F44266C19D-8GTZR

 

That seems to suggest that the corruption of the SPD is actually progressive or that bits are being flipped at random and one stick has had a more important bit flipped than the other. 

post #12825 of 17471
So... Everything has been pretty stable for the last few weeks, however I recently upgraded to a 1080 ti and for some reason I was having some screen flickering while posting. Visiting a few forums ended up leading me to a solution, turning off csm support. However, and I am not sure if this is a coincidence or not, but my system has crashed a few times with no error report ever since I disabled that setting. could disabling csm cause this kind of crash?
Crosshair VI
(14 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1800x Crosshair VI Hero Asus GTX 1080 ti OC Kingston HyperX Predator 
Hard DriveHard DriveHard DriveCooling
Samsung Evo 850 Samsung Evo 850 Seagate Barracuda Noctua NH-D15S 
OSMonitorKeyboardPower
Windows 10 Pro XB271HU Das Keyboard 4 Pro Evga Supernova 1000 P2 
CaseMouse
Phanteks Enthoo Pro  Logitech G900 
  hide details  
Reply
Crosshair VI
(14 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1800x Crosshair VI Hero Asus GTX 1080 ti OC Kingston HyperX Predator 
Hard DriveHard DriveHard DriveCooling
Samsung Evo 850 Samsung Evo 850 Seagate Barracuda Noctua NH-D15S 
OSMonitorKeyboardPower
Windows 10 Pro XB271HU Das Keyboard 4 Pro Evga Supernova 1000 P2 
CaseMouse
Phanteks Enthoo Pro  Logitech G900 
  hide details  
Reply
post #12826 of 17471
Quote:
Originally Posted by Demoniacstar View Post

that is what he recommends ....but to much SoC will cause instability issues as well as to low.....there is a way i use Aida64 cache mem test to judge whether or not i have enough SoC voltage....

this is bad soc voltage



this is good SoC voltage


these are old screenshots before Aida fixed the latencies and other stuff for this test...
and this is basicly when i first got my board back on 092 bios or whatever it wuz.....
but ....if you have higher then 3.0 latency on l2 cache now you're SoC voltage is not right
if you have higher then 10.0 on l3 cache your latencies are not right....
if you run this test a screen shot it ...then change you SoC voltage and compare you will see how things change ....also if you l3 cache is not above 400 GB/s you are running a bad SoC voltage and it needs to be tuned
im going to edit this with some better pics i think ....if i can find some ...lol...if i cant find some ill just redo this test with current settings and then purposely throw off SoC voltage so you guys can see more clearly now what im talking about...give me a few to find and reedit this

I tried out a test using AIDA64 - I have to say I hate the crap of hiding results because you are using a trial version, and the price of $40.00 seems excessive. Maybe it is worth it but not to me I think. Ends Rant. I used SIV64 5.9 to get the latencies that AIDA hides.

As a side note SIV64 is deep,lol. The interface is not as intuitive as some utilities, and I need to go through the guide to really learn it, but man, if you are a techie and want to delve into your system, there is nothing I know of that will get you more detailed information.

Anyway, if I am understanding correctly I should examine:

1. "if you have higher then 3.0 latency on l2 cache now you're SoC voltage is not right"

My L2 Cache using SIV64 reads 3.7 ns, so this may be 0.7 ns high?

2. "if you have higher then 10.0 on l3 cache your latencies are not right"

My L3 Cache using SIV64 reads 7.7 ns, so I think I am looking good there.

3. "also if you l3 cache is not above 400 GB/s you are running a bad SoC voltage and it needs to be tuned"

My L3 Cache in AIDA is 415.31 GB/s for Read, 410.84 GB/s for Write, and 480.16 for Copy, so that seems good.

So based on these results I should change my SOC volts to get L2 Cache latency lower?


Edited by CeltPC - 4/30/17 at 5:48pm
post #12827 of 17471
Quote:
Originally Posted by orlfman View Post

i'm not sure if this has been mentioned and this thread moves fast... but the bios 1107 on the front page isn't linking to the correct link... its linking to bios 0083.
both of your photos show higher than 3.0ns on l2 and higher than 10.0ns on the l3 so not sure which one is bad or good.

as i said in my post if you would care to read the entire POST......those pics are from old bio 092 or whatever it wuz when i first got my board.....i know my grammar sucks but the things i type are NOT that hard to follow ....lol...

bad soc voltage


good soc


and i purposely show lower ram clocks and lower cpu clock here to show that its not the clock speed messing with these numbers .......
l2 and l3 depend on a greatly tuned SoC voltage........
just slamming the SoC voltage up and thinking hey it is stable .....or even running a battery of tests might say it is stable when in fact it might be
but if you do not tune that SoC voltage you are not going to get your best performance out of your Ryzern build.......

o wait here you go

Edited by Demoniacstar - 4/30/17 at 6:22pm
post #12828 of 17471
Quote:
Originally Posted by Demoniacstar View Post

as i said in my post if you would care to read the entire POST......those pics are from old bio 092 or whatever it wuz when i first got my board.....i know my grammar sucks but the things i type are NOT that hard to follow ....lol...

bad soc voltage


good soc


and i purposely show lower ram clocks and lower cpu clock here to show that its not the clock speed messing with these numbers .......
l2 and l3 depend on a greatly tuned SoC voltage........
just slamming the SoC voltage up and thinking hey it is stable .....or even running a battery of tests might say it is stable when in fact it might be
but if you do not tune that SoC voltage you are not going to get your best performance out of your Ryzern build.......

o wait here you go
your own admission you stated the pics where prior to aida updating, so why would you post faulty pics to show good soc then in a thread to help people? mind you, i did read your entire post the first time but was irritated that you used pics that contradicted what you were stating. many people come here for help, and photos help a ton. correct photos that easily correspond accordingly.
strix
(14 items)
 
  
CPUMotherboardGraphicsRAM
AMD Ryzen R7 1800x Asus X370 Crosshair Vi Hero EVGA 1080 FTW G.skill Flare X 3200 2x8 16gb 
Hard DriveHard DriveHard DriveCooling
2x Western Digital Black 2TB 2x 512GB Samsung 850 Pro 1x 512GB Samsung 950 Pro Noctua NH-D15S + 2x Corsair ML120's 
OSMonitorKeyboardPower
Windows 10 Education Dell S2417DG Ducky One White & Grey - Cherry Reds Evga Supernova G3 750 
CaseMouse
Corsair Air 740 Steelseries Rival 700 
  hide details  
Reply
strix
(14 items)
 
  
CPUMotherboardGraphicsRAM
AMD Ryzen R7 1800x Asus X370 Crosshair Vi Hero EVGA 1080 FTW G.skill Flare X 3200 2x8 16gb 
Hard DriveHard DriveHard DriveCooling
2x Western Digital Black 2TB 2x 512GB Samsung 850 Pro 1x 512GB Samsung 950 Pro Noctua NH-D15S + 2x Corsair ML120's 
OSMonitorKeyboardPower
Windows 10 Education Dell S2417DG Ducky One White & Grey - Cherry Reds Evga Supernova G3 750 
CaseMouse
Corsair Air 740 Steelseries Rival 700 
  hide details  
Reply
post #12829 of 17471
Quote:
Originally Posted by CeltPC View Post

I tried out a test using AIDA64 - I have to say I hate the crap of hiding results because you are using a trial version, and the price of $40.00 seems excessive. Maybe it is worth it but not to me I think. Ends Rant. I used SIV64 5.9 to get the latencies that AIDA hides.

As a side note SIV64 is deep,lol. The interface is not as intuitive as some utilities, and I need to go through the guide to really learn it, but man, if you are a techie and want to delve into your system, there is nothing I know of that will get you more detailed information.

Anyway, if I am understanding correctly I should examine:

1. "if you have higher then 3.0 latency on l2 cache now you're SoC voltage is not right"

My L2 Cache using SIV64 reads 3.7 ns, so this may be 0.7 ns high?

2. "if you have higher then 10.0 on l3 cache your latencies are not right"

My L3 Cache using SIV64 reads 7.7 ns, so I think I am looking good there.

3. "also if you l3 cache is not above 400 GB/s you are running a bad SoC voltage and it needs to be tuned"

My L3 Cache in AIDA is 415.31 GB/s for Read, 410.84 GB/s for Write, and 480.16 for Copy, so that seems good.

So based on these results I should change my SOC volts to get L2 Cache latency lower?


i have not used the test you are using ......and that test is showing different numbers then i get in aida for latencies on l3 so i know they are not going to mach up!...
you can experiment with how and what you test with ......but always use the same test in the same situation and environment...meaning same programs running in background or what not .....
if you have other things open i'd close them ...and always test like that .....and change SoC one click at a time .....just one click higher or lower will affect those scores ...
i have spent many many many days and nights testing everything on this set up ...as well as my 8350.....i dont mean to sound like im the king of overclocking or anything like that
but i do have some credibility with what i am saying here man .....
i am in the number 1 spot in geekbench 4 with in the windows environment as well in geekbench 3 i just placed second out of the 1800x's .....so i know my method of tuning systems is solid....and have taught this method of tuning to many others as well as rankdropper who also got his 8350 and 9590 to some of the uper spots in geekbench and has even beaten me due to his remarkable ability at understanding ram overclocking and choosing the right ram to get wink.gif
post #12830 of 17471
Quote:
Originally Posted by orlfman View Post

your own admission you stated the pics where prior to aida updating, so why would you post faulty pics to show good soc then in a thread to help people? mind you, i did read your entire post the first time but was irritated that you used pics that contradicted what you were stating. many people come here for help, and photos help a ton. correct photos that easily correspond accordingly.

you can go troll someone else here man ......
get off my ass already dude..........and it does show how l3 and l2 change with SoC.......regardless of the latencies .....
and as you can see in my other pics i just posted......and i did state that id update with new pics from this newer bios......
so go cry over something else already dude
you're starting to piss me off
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD Motherboards
Overclock.net › Forums › AMD › AMD Motherboards › ROG Crosshair VI overclocking thread