SimForums.com Homepage
Forum Home Forum Home > General Discussion Forums > Hardware, Software, and Computer Technology
  New Posts New Posts RSS Feed - My memory sticks wont run at listed frequency
  FAQ FAQ  Forum Search   Register Register  Login Login

My memory sticks wont run at listed frequency

 Post Reply Post Reply Page  <1 234
Author
Message
jfri View Drop Down
Senior Member
Senior Member


Joined: May-10-2004
Points: 350
Post Options Post Options   Thanks (0) Thanks(0)   Quote jfri Quote  Post ReplyReply Direct Link To This Post Posted: Yesterday at 8:56am
Originally posted by NickN NickN wrote:

You have a record of all your settings for running 4.5 @ 2133 memory speed which should include the ones I outlined in setting up before making any clocking/voltage changes so you can default back to that whenever necessary. There is the fact that I do not recall rebooting and retesting again after a successful run like we are doing here with the 2400 memory.
 
 
In that, if there is a BIOS bug or some kind of BIOS update issue in play then repeating that with the known successful setup may very well shed some light here and save time by not spinning our wheels trying to overcome something that can not be overcome.
 
 
Given these illogical results and the pattern that they display are pointing to, I am thinking at this point you need to go back to the 2133 sticks and that setup, confirm the settings are all restored to that successful stable clock and then repeat the tests, but this time reboot and run them again.
 
If the same result is seen as it is with the 2400 memory then we have been spinning our wheels trying to overcome an abnormal condition. There have been several tests setups so far where the system passes both tests but once rebooted it fails. I really believe at this point it would be worth exploring this by reverting back to 2133 and your established stable clock to find out what happens.
 


So I put back my 2133 memory and loaded a saved profile of the last stable clock. The first LINPACK test pased (no AVX) the following CPU test failed with bluescreen WATCHDOG_TIME_OUT
So it seem the new memory is not the only thing to blame.
Question now what is my best option ?
Back to Top
NickN View Drop Down
Certified Professional
Certified Professional


Joined: November-21-2007
Points: 17712
Post Options Post Options   Thanks (0) Thanks(0)   Quote NickN Quote  Post ReplyReply Direct Link To This Post Posted: 23 hours 28 minutes ago at 12:17pm
I have never heard of or seen any issues reported with OCCT v4.4.2 or above in CPU/Linpack tests.
 
I have heard of PRIME95 issues that had to be addressed some years back and updates to the P95 software were released when it was found.
 
 
I think the next step from here after reinstalling the 2133 sticks and seeing the strange off the wall result is to confirm all the BIOS settings and confirm the BIOS flash. Although saving profiles is a nice convenience I have never relied on them blindly to restore my BIOS to a previous state without having a defined written list, or, a set of BIOS screen images I can use (either way) to refer back to and confirm the saved profile restored everything, not just what I edited, when something strange like this crops up.
 
This is part of the work with overclocking. I would make that same list if I was not overclocking for future reference. 
 
 
Also, I do recall from years back that Raja specifically stated that when a BIOS update with BIOS management engine is performed, manually resetting the CMOS was a requirement. Going back before UEFI BIOSs It was a requirement to perform that manual CMOS reset task after any/all BIOS updates. It is not a common practice today however failure to do so when it is needed *(you would think the motherboard company would inform you of that, but may not)* the system could end up unstable or erratic.
 
 
There is also the possibility of a weak motherboard battery in the mix here. After years of use it should be replaced as a precaution and to eliminate that as a possible issue. Although once powered up the battery is out of the loop, if the battery is weak but not dead it could cause problems between points where the system is shut down and restarted again.
 
 
So from here and if it were me, I would proceed as follows and do ALL 8 outlined steps below in the order they appear:
 
You can skip #1 for now until items 2 through 7 are accomplished.
 
1. Manually verify every setting in the BIOS is as it was when I passed my last set of clocking tests with flying colors. In this case those settings should be as follows and take very careful note of the SPECIAL NOTICE: PLEASE READ CAREFULLY under the: INITIAL DEFAULT BIOS SETUP LIST Z87 CHIPSET ASUS MOTHERBOARDS.
 
 
This list will be used over and over again and includes the default + the overclock setup with my listed changes for overclocking.
 
 
 
             =====INITIAL DEFAULT BIOS SETUP LIST Z87 CHIPSET ASUS MOTHERBOARDS =====
 

BOOT LOGO DISPLAY: DISABLED
SETUP MODE: ADVANCED MODE
 
 
Please Note: Wireless keyboards and mice which are unified though a single USB port such as Logitech wireless keyboard/mouse work fine in the BIOS but may malfunction and not work correctly during operations such as Windows installs from the DVD if the following setting in the BIOS under the Advanced Menu > USB Configuration is not set to:
 
EHCI HANDOFF: ENABLED This can be set to disabled by Asus by default. If you are affected by this issue, it must be enabled. If you are not affected by the issue leave the setting at its default when installing Windows from the DVD or running a Windows repair. THIS IS ONLY NEEDED FOR LOGITECH WIRELESS UNIFIED DEVICES. Individual USB wireless devices typically do not display any such issues however if you see them you can change the setting above.
 
Further, under the same menu if you use PCIe storage controllers that use a Linux based boot disk for toolbox access, this setting may need be disabled before use if the disk fails to boot to the utilities on it:
 
Intel xHCI MODE: ENBLED
 
That must be re-enabled after using such boot utility disks for proper operation in Windows. It only needs to be temporarily changed if you have issues with such tool disk boots. Otherwise leave it on the default setting of: SMART AUTO
 
 
Set up your memory correctly. Under the AI Tweaker Menu set:
AI OVERCLOCK TUNER: XMP
There may be other settings associated with this in the list where you must set the memory speed yourself. Be sure to set those manually based on the manufacture specification for the memory as well if required.
 
Advanced Menu - SATA Configuration. If you use any SSD drives make sure the SATA MODE SELECTION is set to: AHCI
If you use nothing but mechanical drives and know for a fact you will never run SSD NOR do you want or need 'HOT PLUG' support, then set the SATA MODE to: IDE
 
 
Advanced Menu - System Agent Configuration, set the following:
CPU AUDIO DEVICE: DISABLED    
NB PCIe CONFIGURATION - Verify the video card is running x16.
If your video card is PCIe 3.0 then set: PCIEx16_x LINK SPEED: GEN 3   
DMI LINK ASPM CONTROL: DISABLED
PEG ASPM: DISABLED
 
Advanced Menu - Memory Configuration - MEMORY REMAP and (if there) MEMORY SCRAMBLER (both): ENABELD
 
Advanced Menu - Graphics Configuration -
PRIMARY DISPLAY: PCIE
IGPU MULTI-MONITOR: DISABLED
Advanced Menu - CPU Configuration
Hyper Thread - DISABLED
 
Advanced Menu - Onboard Device Configuration - Set these as you need them. I can not define your required settings for this section. If you don't know, leave it at default. 
 
 

This completes the INITIAL BIOS SETUP without any clocking. It is what will be used to normally run a system using the manufacture CPU speed and memory speed/timing
 
 
************************************************************
SPECIAL NOTICE: PLEASE READ CAREFULLY
 
UNDER THE CONDITIONS WHEREBY THE SYSTEM IS EXPERIENCING STRANGE ISSUES OR GLITCHES... THE SYSTEM SHOULD BE FULLY TESTED WITH THE INSTALLED HARDWARE JUST LIKE ANY CLOCKED SYSTEM BEFORE PROCEEDING. FAILURE TO DO SO MEANS THAT A SYSTEM COULD HAVE DEFECTIVE HARDWARE, UNSTABLE OR FLAKEY BIOS INSTALLED, OR, A CORRUPT WINDOWS INSTALLATION.
 
ONCE THE SYSTEM PASSES ALL TESTS WHILE RUNNING THE DEFAULT HARDWARE SPEEDS WITH A DEFAULT STABLE SETUP THEN OVERCLOCKING CAN CONTINUE. ASSUMING THE HARDWARE AND WINDOWS IS FINE UNDER DEFAULT CONDITIONS AND SKIPPING TESTING IS NOT AN OPTION.
 
TESTS ARE AS FOLLOWS:
MEMTEST IN DOS: MINIMUM OF 3 TEST RUNS (9-10 TESTS FOR EACH RUN)
BOOT ONE "SINGLE" MEMORY STICK IN THE CORRECT SLOT FOR "SINGLE" STICK OPERATION. SWAP STICKS AND REPEAT TILL ALL CONFIRM PASSED
IN WINDOWS:
OCCT:LINPACK
OCCT:CPU
FINALLY:
MEMTEST IN DOS: WITH ALL STICKS INSTALLED INTO THE CORRECT SLOTS FOR MULTISTICK OPERATION.
 
IF A FAILURE IS SEEN, IT MUST BE ADDRESSED FIRST BEFORE CONTINUING BELOW.
 
HARDWARE FAILURES, BIOS SETTINGS OR FLAKEY BIOS ITSELF MUST BE CORRECTED.
 
IF ALL TESTS PASS, THEN CONTNUE BELOW
 
************************************************************
 
If this is a BIOS RESET and manual RESTORE then proceed to the next part to begin reset of a clocked system THAT TESTED SUCCESFULLY running at default.
 
 

                       ----------INITIAL CLOCKING BIOS SETUP LIST Z87 CHIPSET ASUS MOTHERBOARDS ----------
 

BIOS in ADVANCED mode and confirm or make the following changes:
 
 
MONITOR MENU - bottom of the list: ANTI SURGE SUPPORT: DISABLED
 
 
ADVANCED MENU -
CPU CONFIGURATION: HYPER THREADING: DISABLED
THERMAL MONITOR: ENABLED
ACTIVE PROCESSOR CORES: ALL
LIMIT CPUID MAXIMUM: DISABLED
EXECUTE DISABLE BIT: ENABLED
INTEL VIRTUALIZATION TECH: DISABLED (Unless running a virtual version of Windows)
HARDWARE PREFETCH: ENABLED
BOOT PERFORMANCE MODE: MAXIMUM NON TURBO PERFORMANCE
DYNAMIC STORAGE ACCELERATOR: DISABLED
 
Scroll down to the bottom and enter the menu CPU POWER MANAGEMENT:
EIST: DISABLED
TURBO MODE: ENABLED
CPU C STATES: ENABLED   - This is set to AUTO by default, change it to ENABLED and under CPU C STATES the following will appear, change as outlined:

ENHANCED C1: DISABLED
CPU C3 REPORT: ENABLED
CPU C6 REPORT: DISABLED
C6 LATENCY: SHORT
CPU C7 REPORT: DISABLED
C7 LATENCY: LONG
PACKAGE C3 STATE REPORT: C3
 

ADVANCED MENU - PCH CONFIGURATION:
PCI EXPRESS CONFIGURATION -
DMI LINK ASPM CONTROL: DISABLED
PEG ASPM: DISABLED     NOTE: YOUR BIOS MAY READ THIS WAY INSTEAD -  ASPM SUPPORT: DISABLED
PCIe SPEED: Leave this on BIOS default   AUTO is fine
 

ADVANCED MENU - SYSTEM AGENT CONFIGURATION:
CPU AUDIO DEVICE: DISABLED
NB PCIE CONFIGURATION:
Verify PCIe slot the card is installed into (typically the first for BEST performance) displays: X16 and the LINK SPEED dropdown is set to GEN3
NOTE: All 600 and 700 series video cards are GEN3. Any video card lower than a 600 series is GEN2.
 
DMI LINK ASPM CONTROL: DISABLED
PEG ASPM: DISABLED
GRAPHICS CONFIGURATION:
PRIMARY DISPLAY: PCIE
IGPU MULTI-MONITOR: DISABLED
ADVANCED MENU - MEMORY CONFIGURATION:
MEMORY REMAP: ENABLED
MEMORY SCRAMBLER: ENABLED  (may not appear in all systems)
AI TWEAKER MENU -
AI OVERCLOCK TUNER: XMP 
Should already be set as well as speed however DRAM Voltage for overclocking will be different. DRAM Voltage - This depends on the memory modules. In most cases with 1.55v memory sticks this will be set manually to1.65v (1.66v if you know it is needed). 1.65/66 is fine for 1.55v modules.
 
CPU STRAP: 100
BCLK FREQUENCY: 100
Enter the DIGI+POWER CONTROL MENU
LOAD LINE CALIBRATION: 8   
CPU POWER PHASE CONTROL: OPTIMIZED
CPU THERMAL CONTROL: 130
 
All other settings should be OK at default however verify:
CPU VOLTAGE FREQUENCY: AUTO
VRM SPRAD SPECTRUM: DISABLED
ACTIVE FREQUENCY MODE: DISABLED
CPU POWER DUTY CYCLE: T.PROBE
CPU CURENT CAPABILITY: AUTO - NOTE: This will increase automatically as CPU clocking voltages go up. In some cases it may need to be set manually to 130% or 140% (RED) for any clock.

DRAM CURRENT CAPABILITY: 100% - Sometimes this may need to be increased depending on the memory and system stability.

DRAM VOLTAGE FREQUENCY: AUTO
DRAM POWER PHASE CONTROL: AUTO
Exit the DIGI+POWER CONTROL MENU
 

Enter the CPU POWER MANAGEMENT MENU - Confirm or change:
ENHANCED INTEL SPEEDSTEP: DISABLED
TURBO MODE: ENABLED
Exit the CPU POWER MANAGEMENT MENU. Scroll down to the bottom of the list:
SVID SUPPORT: AUTO 
All boards other than ROG
NOTE: This setting could also be listed as: SVID CONTROL instead of SUPPORT
SVID SUPPORT: ENABLED ROG BOARDS ONLY
 
Scroll down to: CPU SPREAD SPECTRUM: DISABLED
 

The base Asus BIOS setup list for manually clocking all systems is now complete. Please hit F-10 save and exit and then go back into your BIOS and RECONFIRM every setting posted above. It has been noted that some boards may not hold a setting. We want to confirm the above list has indeed been saved and the system is booting this list.
Once confirmed, then continue...

 

                --------- MY CUSTOM VOLTAGE/BIOS CHANGES FOR CLOCK/MEMORY SPEED X Z87 CHIPSET ASUS MOTHERBOARDS
------

This list is the recorded BIOS changes for my specific clocking level and hardware after I have tested the system to be stable. This list may (SHOULD) be MANUALLY edited over time as well as other settings posted above this list. When ever a CONFRIMED stable change is made that is to remain, the list BELOW and ABOVE should be edited with those changes for future reference...
 
I went back to page 2 of this thread and noted your last reported successful 2133 memory speed voltage settings. They are noted below in (parentheses) next t the settings.
If what you have recorded manually (not a saved BIOS profile) below is different then those should be edited (corrected) below...
 
 
AI OVERCLOCK TUNER: XMP
BCLOCK: 100
CPU RATIO: PER CORE
CPU MULTIPLIER: <MY DEFINED MULTIPLIER> ALL 4 SETTINGS EQUAL 
(45)
(45)
(45)
(45)
 
OVERCLOCK TUNER: AS IS
DRAM FREQUENCY: MY DEFINED DRAM SPEED IN MHZ (2133)
DRAM TIMING CONTROL MENU - VERIFY INSTALLED MEMORY TIMING IN THE LIST
(9)
(11)
(11)
(31)  I believe this is correct, it was never listed. XMP should set it right automatically.
(1T or 2T) depends on the memory manufacture or tested clock change. 1T is faster. Do not change it to 1T id it is 2T until everything else is tested and then 1T can be tested (all tests including Memtest)
 
CPU CACHE MULTIPLIER: <MY DEFINED MULTIPLIER> BOTH SETTINGS EQUAL
(43)
(43)
 
CPU CORE VOLTAGE: MANUAL MODE
CPU CORE VOLTAGE OVERRIDE: MY DEFINED CLOCKING VOLTAGE (1.28v)
 
CPU CACHE VOLTAGE: MANUAL MODE
CPU CACHE VOLTAGE OVERRIDE: MY DEFINED CLOCKING VOLTAGE (1.26v) (1.28v)
 
CPU SYSTEM AGENT: OFFSET MODE SIGN: +
CPU SYSTEM AGENT VOLTAGE: MY DEFINED CLOCKING VOLTAGE (0.280v)
 
CPU ANALOG I/O VOLTAGE OFFSET MODE SIGN: +
CPU ANALOG I/O VOLTAGE OFFSET: MY DEFINED CLOCKING VOLTAGE (0.050v)
 
CPU ANALOG I/O DIGITAL OFFSET MODE SIGN: +
CPU DIGITAL I/O VOLTAGE OFFSET: MY DEFINED CLOCKING VOLTAGE (.100v)
 
CPU INPUT VOLTAGE: MY DEFINED CLOCKING VOLTAGE (1.88v)
 
DRAM VOLTAGE: MY DEFINED CLOCKING VOLTAGE (1.66v)
 
CPU SPEAD SPECTRUM: DISABLED
 
 
THE FOLLOWING SETTINGS SHOULD BE SET AS SHOWN BY DEFAULT IN ALL BIOS’s. IF THEY ARE DIFFERENT CHANGE THEM TO THE FOLLOWING:
 
CPU STRAP: 100MHz
PLL SELECTION: AUTO
FILTER PLL: AUTO
INITIAL BCLOCK FREQUENCY: AUTO
ASUS MULTICORE ENVIRONMENT: AUTO
BCLOCK FREQUENCY RATIO: AUTO
INTERNAL PLL OVERVOLTAGE: AUTO
EPU POWER SAVING MODE: DISABLED
DRAM TIMING CONTROL MENU: Scroll down to: MEMEORY SCRAMBLER SETTING: OPTIMIZED
EXTRMEME OVERVOLTAGE: DISABLED
ALL OTHER SETTINGS OTHER THAN YOUR CUSTOM INPUTS NOTED ABOVE: AUTO

 
 
==============================================
 
 
 
 
 
 
Getting back to what I would do next....
 
 
 
2. Knowing I will probably lose my saved profile in the next step, make sure I indeed have a complete written or printed record of the <entire BIOS, not just what I changed> settings that passed.
 
 
3. Reflash the same BIOS using the same files/tools from Windows that I used from the support site to update the BIOS. Reboot as instructed. Note that after such a reset the system may appear to hang.. do not touch it and let it go through its cycles until the initial BIOS screen or enter BIOS screen appears.
 
THIS TIME, after the system goes through its initial reboots and you end up at the BIOS screen, not before, then shut down and POWER OFF (unplug).
 
 
4. Open the tower. On the board there is a pancake battery, it is typically a number CR2032. If I need assistance locating it on the board the motherboard manual will show me where it is but its very easy to see as it is usually a bright nickel circle and pops out of the socket like any other pancake battery would. NOTE: Typically POSITIVE on the battery is facing me in the socket, confirm then pull the battery and proceed to step 5.
 
 
5. In the motherboard manual it will show me where to manually reset the CMOS (also known as clear CMOS or clear CLRTC). Most of the time this is a 3 pin jumper with a cap over 2 pins. Sometimes it is a button. The manual will explain or display which and where it is. Sometimes this is referred to as CLRTC in the manual. Locate that in the manual and follow the procedure. If its 3 pins I would move the jumper for 10 seconds and put it back the way it was, if it is a button I would press and hold for 10 seconds.
 
 
6. Install a NEW CR2032 battery in the motherboard noting the +/- sides when originally removed.
 
 
7. Power up and boot. Note that after such a reset the system may appear to hang.. do not touch it and let it go through its cycles until the initial BIOS screen or enter BIOS screen appears. Then enter the BIOS, go directly to F-10 save and exit, reboot and enter the BIOS again.
 
 
8. I would then proceed to manually restore my BIOS settings including the time/day/year if required and the settings that I had manually recorded, both the initial setup settings and the clock settings from the successful test runs. The compete list is provided above in item #1
 
DO NOT USE A SAVED PROFILE. Set everything manually, F-10 save and exit, then go back into the BIOS and confirm all settings held before retesting.
 
 
From there I would boot into Windows give the system time to fully boot. Verify the Windows clock, CPU clock, memory speed and timing are all correct, then proceed to retest.
 
I would not save any BIOS profiles in the BIOS until I have achieved successful test runs of everything associated with the 2133 memory. Then proceed to the 2400 memory again, but this time I would not start at 4.5/4.4GHz, I would start at 4.2 and work up.
 
 
The above would eliminate the possibility of a botched BIOS flash, a weak or flakey motherboard battery and a corrupt saved BIOS profile all at the same time.
 
 
 
This is how I would proceed, and I would not skip a single step or do anything out of the order I posted (2 through 8).  Wink 
 
 
 
 
Back to Top
jfri View Drop Down
Senior Member
Senior Member


Joined: May-10-2004
Points: 350
Post Options Post Options   Thanks (0) Thanks(0)   Quote jfri Quote  Post ReplyReply Direct Link To This Post Posted: 20 hours 47 minutes ago at 2:58pm
Originally posted by jfri jfri wrote:

Originally posted by NickN NickN wrote:

You have a record of all your settings for running 4.5 @ 2133 memory speed which should include the ones I outlined in setting up before making any clocking/voltage changes so you can default back to that whenever necessary. There is the fact that I do not recall rebooting and retesting again after a successful run like we are doing here with the 2400 memory.
 
 
In that, if there is a BIOS bug or some kind of BIOS update issue in play then repeating that with the known successful setup may very well shed some light here and save time by not spinning our wheels trying to overcome something that can not be overcome.
 
 
Given these illogical results and the pattern that they display are pointing to, I am thinking at this point you need to go back to the 2133 sticks and that setup, confirm the settings are all restored to that successful stable clock and then repeat the tests, but this time reboot and run them again.
 
If the same result is seen as it is with the 2400 memory then we have been spinning our wheels trying to overcome an abnormal condition. There have been several tests setups so far where the system passes both tests but once rebooted it fails. I really believe at this point it would be worth exploring this by reverting back to 2133 and your established stable clock to find out what happens.
 


So I put back my 2133 memory and loaded a saved profile of the last stable clock. The first LINPACK test pased (no AVX) the following CPU test failed with bluescreen WATCHDOG_TIME_OUT
So it seem the new memory is not the only thing to blame.
Question now what is my best option ?


There seem to be a mistake here. I read earlier in this thread that I got bluescreen before. It tested OK after setting 4.4 GHz instead of 4.5 cache x43. I think the loaded profile was before the memory upgrade

I'm referring to this earlier posting

OCCT:CPU failed twice with bluescreens something about watchdog  - This indicates voltage too low, most likely the CPU CORE VOLTAGE
 
Lets review the numbers that are passing 4.4GHz tests...
 
CPU core 1.26 V  - This is probably too low and needs to come up. I would raise this to 1.28v and if it fails go to 1.30v and stop there. Remember to verify heat and stability.
 
CPU cache 1.28 V  - I would think this would be fine for now but it could come up to 1.30v ONLY IF you failed at CPU Vcore 1.28v and had raise CPU CORE VOLTAGE to 1.3v0 --and-- added in VCCSA to .300 (system agent below) and it still failed with ERRORS (not temp)
 
CPU agent 0.28 V  - This one is tricky. You could raise this to .300v for the next run but I would leave it as is for now and see if raising CPU CORE VOLTAGE alone will stabilize 4.5GHz. If not then bump this up to .330 if you had to raise CPU core voltage to 1.30v and it still failed. Remember to verify heat and stability.
 
CPU input 1.88 V  -This is fine the way it is. No need to change it.
 
 
Question: What is your current DRAM VOLTAGE?   1.65?  If it is, raise this to 1.66v
Back to Top
NickN View Drop Down
Certified Professional
Certified Professional


Joined: November-21-2007
Points: 17712
Post Options Post Options   Thanks (0) Thanks(0)   Quote NickN Quote  Post ReplyReply Direct Link To This Post Posted: 20 hours 8 minutes ago at 3:37pm
Ok, this is getting convoluted..  Lets make it easy
 
 
 
 
Originally posted by jfri jfri wrote:

Originally posted by NickN NickN wrote:

Ok.. then continue as I outlined with the CPU voltage bumps first and do not skip 1.28v
If you pass on the CPU core voltage bumps, then we move onto CPU cache bumps on the same voltages and retest (heat and stability). You can take the first bump to 4.1GHz from 3900 and see what happens. If you pass, go to 4200, and so on until you have duplicated your original clock speeds.. 


So I have set
CPU core 1.28 V
DRAM voltage 1.66 V (which turned the numbers red colored)
Both CPU:OCCT and CPU:LINPACK passed the test and the highest registered core temperature was 91 C.
Note that I think that 3900 was my original setting.
 
 
Originally posted by jfri jfri wrote:

Originally posted by NickN NickN wrote:

So now that there is a CONFIRMED stable overclocked CPU tested..  go ahead and raise CPU Cache to 4100 (41/41) with no voltage change and see what the tests do.

Once the CPU cache is defined, then we move to checking the memory again, but this time leaving both sticks in the system and run MEMTEST while overclocked on the 2133 sticks.


When raising Cache to 41 42 43 both OCCT tests passed
When I tried 44 things didn't work at all. After starting the CPU test the computer hanged up after less than 10 minutes so I had to do a hardware reset.
I returned Cache to 43 and started the Memtest86+ program which are running right now. So far it display three passes without any errors detected
 
 
Originally posted by NickN NickN wrote:

So at this point, congratulations!  Clap Beer    Its time to move to the 2400 memory and see where it stands.
 
 
 
 
 
1. Did you restore your original list of voltages and settings that passed on the 2133 memory and did it pass this time as it did in the past?
Y/N
 
If NO, or not sure, then do so and test them again with a reboot between. Those passed AVX temp load according to your report on page 2 of the thread so they should pass again the same.
 
            If it doesn't pass then print out what I post above (the 8 steps to take) and do
                                        them. You should test the 2133 memory in the system
                                        like you just bought the system and it is running default. INITIAL
                                         DEFAULT BIOS SETUP LIST Z87 CHIPSET ASUS
                                         MOTHERBOARDS... and then proceed to test the 2133 memory
                                        with a default BIOS setup exactly as you would if you had just
                                        bought the system with 2133 memory installed and was not
                                        clocked. Report the outcome here. If it fails, that has to
                                        be resolved first.                                 
                                      
 
 
            If it now passes because of some confusion on your part in retesting then at this
                                      point the 2400 will be reinstalled and this time you will not start
                                      out running overclocked. Restore your BIOS to default settings
                                      (load setup defaults) reboot and then using the list I provided in #1
                                       of the 8 step outline above INITIAL DEFAULT BIOS SETUP LIST
                                       Z87 CHIPSET ASUS MOTHERBOARDS... and then proceed to test
                                      the 2400 memory with a default BIOS setup exactly as you would if
                                      you had just bought the system with 2400 memory installed and it
                                      was not clocked. Report the outcome here. If it fails, that has to
                                      be resolved first.
 
 
 
If YES, and you saw fails that make no sense since they did pass successfully,...  then print out what I post above (the 8 steps to take) and do them all just like I outlined in:  If it doesn't pass above.
 
 
I am trying to eliminate where the flake is in this system because what is happening is not normal or expected at this point.
 
 
 
 
Back to Top
NickN View Drop Down
Certified Professional
Certified Professional


Joined: November-21-2007
Points: 17712
Post Options Post Options   Thanks (0) Thanks(0)   Quote NickN Quote  Post ReplyReply Direct Link To This Post Posted: 20 hours 4 minutes ago at 3:41pm
Please go back up and reread my last post.. I just made some edits 3:35pm EST
Back to Top
jfri View Drop Down
Senior Member
Senior Member


Joined: May-10-2004
Points: 350
Post Options Post Options   Thanks (0) Thanks(0)   Quote jfri Quote  Post ReplyReply Direct Link To This Post Posted: 18 hours 37 minutes ago at 5:08pm
Originally posted by NickN NickN wrote:

1. Manually verify every setting in the BIOS is as it was when I passed my last set of clocking tests with flying colors.

I have printed out that list of BIOS settings and checked in my current BIOS which was loaded
from a file called stable2133oc.
I found out that the only thing differing was

PCIEx16 x LINK SPEED: was not GEN3 so I changed it to this

CPU cache was not 1.26 V but 1.28 V

I have also looked back at what I wrote earlier in the thread and found

I have now reinstalled the old memory sticks. Made the base BIOS setup and then started with the 4.3 GHz start point.
CPU test failed.
I then set my original settings which I made a note of before installing the new memory
and that is
CPU core 1.26 V
CPU cache 1.28 V
CPU agent 0.28 V
CPU input 1.88 V
other according to base setup
Both OCCT:CPU and LINPACK passed the test
Then I increased the CPU core multiplier to 45 i.e should be the same settings I have used before new memory
OCCT:CPU failed twice with bluescreen something about watchdog
Then I have changed the multipliers to 44 instead and
Both OCCT:CPU and LINPACK passed the test
and max registered core temperature was 86 C

So it seem I got problem getting stable at 4.5 GHz right after putting back my old memory
but the tests passed at 4.4 GHz cache 43. And I am quite sure that the loaded BIOS file was from our current session and not before the new memory was acquired and here it was 4.5 GHz. Later you told me to rerun the test at 4.5 GHz. If so and the tests passed then the result is different now. Could it have been an isolated freak event ? I made the CPU test at 4.4 GHz and it passed.





Back to Top
NickN View Drop Down
Certified Professional
Certified Professional


Joined: November-21-2007
Points: 17712
Post Options Post Options   Thanks (0) Thanks(0)   Quote NickN Quote  Post ReplyReply Direct Link To This Post Posted: 16 hours 60 minutes ago at 6:45pm
Ok then there are some crossed wires. The PCIe GEN3 wont have any effect, and the CPU cache voltage was 1.28v and 1.26 was a typo on my part.
 
None the less going back and looking again I see no reference to your final successfully tested clock being 45x (4.5GHz) I assumed it was because that was the goal and in that discussion I said:
 
 
Originally posted by NickN NickN wrote:

Lets review the numbers that are passing 4.4GHz tests...
 
CPU core 1.26 V  - This is probably too low and needs to come up. I would raise this to 1.28v and if it fails go to 1.30v and stop there. Remember to verify heat and stability.
<snipped>
---------------------------------------
 
With the numbers and what to do with them outlined above, make the changes and rerun the tests @ 4.5GHz CPU CORE RATIO: PER CORE and 45 for each core.
 
Report the results and let me know what the current CPU cache multiplier is set to when it passes/fails
 
 
 
Going further I said and you quoted me, and reported back with:
 
Originally posted by jfri jfri wrote:

Originally posted by NickN NickN wrote:

THIS MEANT CONTINUE TESTING CPU FOR 4.5GHZ ON HIGHER VOLTAGES... let me know if it doesn't pass 4.5 when you hit 1.30v
 
Ok.. then continue as I outlined with the CPU voltage bumps first and do not skip 1.28v
AND THEN.....
If you pass on the CPU core voltage bumps, (Meaning 4.5GHz) then we move onto CPU cache bumps on the same voltages and retest (heat and stability). You can take the first bump to 4.1GHz from 3900 and see what happens. If you pass, go to 4200, and so on until you have duplicated your original clock speeds.. 


So I have set
CPU core 1.28 V
DRAM voltage 1.66 V (which turned the numbers red colored)
Both CPU:OCCT and CPU:LINPACK passed the test and the highest registered core temperature was 91 C.
Note that I think that 3900 was my original setting.
 
 
And after that I said in confirmation but apparently was not told different:
 
 
Originally posted by NickN NickN wrote:

Note that I think that 3900 was my original setting.
 
Ok well that makes sense because you are getting close to your max AVX temp limit and to bring it up might require more voltage.
 
So good, we are now getting along at a good clip, back to 4.5GHz and if you are correct about the CPU cache, back to where you started.
 
So now that there is a CONFIRMED stable overclocked CPU tested..  go ahead and raise CPU Cache to 4100 (41/41) with no voltage change and see what the tests do.
 
<snipped>
 
 
 
..and the return was..
 
Originally posted by jfri jfri wrote:

Originally posted by NickN NickN wrote:

So now that there is a CONFIRMED stable overclocked CPU tested..  go ahead and raise CPU Cache to 4100 (41/41) with no voltage change and see what the tests do.

Once the CPU cache is defined, then we move to checking the memory again, but this time leaving both sticks in the system and run MEMTEST while overclocked on the 2133 sticks.


When raising Cache to 41 42 43 both OCCT tests passed
When I tried 44 things didn't work at all. After starting the CPU test the computer hanged up after less than 10 minutes so I had to do a hardware reset.
I returned Cache to 43 and started the Memtest86+ program which are running right now. So far it display three passes without any errors detected
 
 
 
 
So this entire time I thought you had reached 4.5GHz @ 2133 and now you are saying you didn't and instead it was 4.4GHz?
 
So it appears there are some crossed wires if that is true.
 
I never would have had you restart testing 4.5 @ 2400 if that was the case.
 
 
 
So at this point you have one of 3 choices..
 
CHOICE 1.
Go back up to the 8 step list and do that before doing anything else (I would if it were me) except when it comes to the list for your settings after <Reflash BIOS, Clear CMOS and replace the motherboard battery>.., use 45x (4.5GHz) for CPU multipliers and run the tests again
 
 
IF IT PASSES 4.5GHz @ 2133 - Run Memtest in DOS again and let me know it all passed.
                                                 We go from there.
                                            
                                            
 
IF IT FAILS 4.5GHz @ 2133 - Then something is very wacky because you did pass that in
                                            the thread on page 2 with 4.3Ghz (43x) CPU CACHE speed. At
                                            that point we know the BIOS flash, BIOS reset and the battery
                                            are all out of the mix and the only thing that leaves is a BIOS
                                            that does not clock well OR a hardware issue with the
                                            CPU/motherboard itself.
 
                                                          I need to know if this fails as it should not
 
 
 
 
CHOICE 2.
Skip the 8 step BIOS/reset and battery change and go back to 4.4GHz with the correct previous voltages and retest.
 
IF IT PASSES 4.5 GHz @ 2133 - Follow the directions in CHOICE 1 above for "IF
                                                   IT PASSES"
 
 
IF IT FAILS 4.5GHz @ 2133 - Then you need to perform the entire 8 step process with
                                               reported outcome results before going further.
 
 
 
 
 
CHOICE 3.
Skip the 8 step BIOS/reset and battery change, restore the BIOS to default for no clocking making sure the section of the list INITIAL DEFAULT BIOS SETUP LIST Z87 CHIPSET ASUS MOTHERBOARDS is in fact restored with 2400 memory installed and then starting from DEFAULT, run the tests with the 2400 memory in the system like you just bought the system with that memory in it and it has not been OCCT tested. Since voltages will be in a non-clocked setup you should use OCCT AVX enabled when testing with Linpack.
 
 
IF IT PASSES ON A DEFAULT SETUP - Then run through the: INITIAL CLOCKING BIOS SETUP LIST Z87 CHIPSET ASUS MOTHERBOARDS and make those changes. THEN start with the Haswell clocking guide for: Section 103
 
STARTING SETUP OUTLINE: 4.2GHz-4.3GHz 
 
https://www.simforums.com/forums/haswell-4-8ghz-on-air-building-a-haswell-system_topic46180_post286808.html#286808
 
NO HIGHER for now
 
 
 
IF IT FAIL ON A DEFAULT SETUP - Then I would do the entire 8 step BIOS/reset and battery change, restore the BIOS to default for no clocking and REPEAT the tests again.
 
                              IF IT PASSES - Then you had a flakey BIOS flash and/or battery issue.
 
                              IF IT FAILS - The his needs to be RESLOVED before going any further.
 
 
 
 
 
 
So there is you list..   should keep you busy for a spell  Beer Thumbs Up
 
 
 
 
Back to Top
jfri View Drop Down
Senior Member
Senior Member


Joined: May-10-2004
Points: 350
Post Options Post Options   Thanks (0) Thanks(0)   Quote jfri Quote  Post ReplyReply Direct Link To This Post Posted: 16 hours 48 minutes ago at 6:57pm
Originally posted by NickN NickN wrote:

Ok then there are some crossed wires. The PCIe GEN3 wont have any effect, and the CPU cache voltage was 1.28v and 1.26 was a typo on my part.
 
None the less going back and looking again I see no reference to your final successfully tested clock being 45x (4.5GHz) I assumed it was because that was the goal and in that discussion I said:


I think it was x45 since it was saved that way. And I have now made a new CPU test at 4.5 GHz and this time it passed. I will do the LINPACK test tomorrow. Should I have AVX enabled ?
This mean computer powered down between these tests. Is that any problem ?
Then I suppose the intent is to do another CPU and LINPACK test with a reboot between
Back to Top
NickN View Drop Down
Certified Professional
Certified Professional


Joined: November-21-2007
Points: 17712
Post Options Post Options   Thanks (0) Thanks(0)   Quote NickN Quote  Post ReplyReply Direct Link To This Post Posted: 16 hours 41 minutes ago at 7:04pm
Yes that is the intent... and yes since its back to the original clock you should have AVX enabled.
 
But you have also stated you had a RTC issue that should not be and you said you had failures @ 4.5 2133 here, at lot of them not just one in that test..  so unless you missed something somewhere I cant see how you would test fine before we moved to 2400 memory and then blow tests on the same settings when putting the 2133 memory back in and retesting it.
 
 
I wouldn't screw around and do CHOICE 1 if it were me to eliminate that completely and do not use any saved profile to restore settings set them all manually, but you can pick which one you wish to follow just let me know which one you did and the result..
 
AND IMPORTANT: IF you are sure you were running 45x, then everywhere I edited back to 45x above and may have missed a 44x somewhere,..you should use 45x instead.
Back to Top
jfri View Drop Down
Senior Member
Senior Member


Joined: May-10-2004
Points: 350
Post Options Post Options   Thanks (0) Thanks(0)   Quote jfri Quote  Post ReplyReply Direct Link To This Post Posted: 5 hours 28 minutes ago at 6:17am
Originally posted by NickN NickN wrote:


But you have also stated you had a RTC issue that should not be and you said you had failures @ 4.5 2133 here, at lot of them not just one in that test..  so unless you missed something somewhere I cant see how you would test fine before we moved to 2400 memory and then blow tests on the same settings when putting the 2133 memory back in and retesting it.


The RTC issue occured with the 2400 memory installed. After letting windows update the time the RTC clock has worked correctly.

After reinstalling the 2133 memory the last time I had only one failed test. I made
1) LINPACK test without AVX which passed
2) CPU test which failed @4.5 GHz with bluescreen and WATCHDOG_TIME_OUT
3) New CPU test yesterday that passed

I had other failures during the earlier 2133 reinstallation and then we made a change. And I think we got it stable at 4.5 Ghz. Then the new 2400 memory was installed and we got problems that didn't made sense so I put back the 2133 memory which are installed now
Back to Top
 Post Reply Post Reply Page  <1 234
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.01
Copyright ©2001-2018 Web Wiz Ltd.

This page was generated in 0.266 seconds.