Skylake hard locks seeminly when IDLE / CStates ON

[GGlyn]

Myself and others have been struggling with a different hard lock problem with Skylake and we have ran out of options but to assume it is a CPU fault/issue. Does anyone have any comments or feedback?

The long thread is over here, but I will summarise. http://www.tomshardware.co.uk/forum/id-2830772/skylake-build-randomly-freezing-crashing/page-7.html New Skylake Build Randomly Freezing/Crashing - Page 7 - windows 10 - Windows 10

Our systems hard lock randomly but appears on the most part to be when IDLE or next to IDLE. The hardlock can come when playing video, browsing (reading) or mostly when not interacting at all. Some (like myself) noted that they seem to occur more often if you are doing a large amount of disk IO (partition move for example or copying 40+ GB of data disk to disk).

Hard locks happen in either Windows (7,8,10) or Linux.

With CStates on, the crashes are often and usually within 10 minutes of IDLE. With CStates disabled completely most find that the system remains "mainly" stable for days. Generally, no one has reported locks when gaming or under high CPU load (so I do not think this is the Prime95 example but then, since compilers / drivers use the latest instructions (according to CPUID) I cant be sure).

The symptoms happen on many different manufacturers motherboards, RAM types and setups. We have successfully removed or replaced all components and the only change that seems to make a difference is CPU for a few. CStates off in the BIOS is absolutely a guarenteed way to stabilize significantly but not fix.

I changed CPU (after trying everything else) and find it now mostly stable with CStates ON. I must be clear that my system is 100% identical, CPU # 1 crashes all the time and must have CStates off, CPU # 2 does not crash so often even with CStates ON. One may argue this is a problem with my PC, a fault, but with so many people having the exact same issues I cant believe it is a defect. And with Cstates/SpeedStep so specifically (a known new part of Skylakes architecture), I remain convinced that this is something real.

Others were not so lucky, a replacement CPU did not make the situation better. Batches from different factories as well etc etc

My new CPU has hard locked but not when IDLE like the last one (not yet anyway). Mine can lock when under full load, video transcoding, but not more than 1 in 10 sessions. Cstates ON seems equally stable on this CPU as OFF.

Thanks for any suggestions you can give.

[LMonk]

Z170-K BIOS 1803

Current microcode is 0x84, why you are asking this?) WIth you help i installed this firmware

[CCris2]

It seems you have the same problem like eyeless45 : "Setting the XMP profile 1 (2133MHz, CL13, 1.2V, etc.) and setting VCore to Adaptive already helped quite a lot, but the system wasn't 100% stable. So after that I set VCCSA and VCCIO both to 1.15V. Since then I had no crash."

Your memory looks like it has same timings as his and on asus mainboard also.

So please set memory on XMP increase VCCSA and VCCIO both to 1.15V and processor on adaptive but with no voltage modification.

[AKota1]

Hi there!

I have the same Problem as described by many in this thread, i will be more specific about my problem:

I bought the following components:

Intel i7 6700k

Asus Z170 Pro Gaming

Crucial Ballistix Sport LT grey DIMM Kit 16GB, DDR4-2400, CL16-16-16 (BLS2C8G4D240FSB)

EVGA 980Ti SC+

Corsair RMi 650 with Cablemod Sleeved Cables (approved for RMi Series)

Phanteks Evolv ATX

Note: I use a Multimonitor Setup with the Intel Onboard Graphics: Main Monitor @ Displayport @ 980ti and 2nd Monitor @ Displayport @ Onboard GFX

Note2: CPU-Z lists 4 Profiles for my RAM, 3 JEDEC (# 8, # 9, # 10) and one XMP - but the XMP Profile with 2400 CL16-16-16-39-55 does not differ from the JEDEC # 8 Profile with 2400 CL16-16-16-39-55

so when i write "XMP off" i just disable the XMP settings in the bios, the frequency still stays at 2400 with CL16-16-16-39-55

At first i had a MSI GTX780 but it died (VRM Chip vaporated) and i decided to buy a new 980 ti.

With the arrival of the EVGA 980Ti my Problems began:

I have always had very randomly freezes of my system (Screen freezes and very loid white noise from Headphones @ onboard sound), but only in games (BF4, Rise of the Thomb Raider, WoW, League of Legends, Diablo 3).

In Windows the System is Rockstable, i never had any freezes in Windows (even while stresstesting)

I have been trying to figure it out since february 2016 and tried the following things:

- driver updates (Intel graphics drivers, nvidia graphics drivers, as media usb controller drivers, intel inf drivers, etc...)

- bios updates (currently @ 1805, microcode 0x84)

- stresstesting:

1) Memtest86+ for 12 Hours - No Errors (Bot RAM Sticks)

2) Prime95 v2.66 for several Hours (blend test, ram test, etc...) - No Errors

3) Unigine Valley 1.0 for several Hours - No Errors

4) Prime95 v2.66 + Unigine Valley 1.0 for several Hours - No Errors

- switched 980ti out for a gtx 970 - it took longer for a freeze, but the freezes still persisted and occured

- fiddled again with the bios - still freezed

As i mentioned, the freezes only happen in games!

I recently startet playing Diablo 3 again and in this game the Problem is very Present.

I do alt-tab very often to verify items if they apply for a certain build or something - so lots of c state switching

sometimes i play d3 a day without any freezes, sometimes 4 freezes occur in 2 hours or less.

then i came across your thread

at first i decided to disable all c-states and speedstep - i tested it for 2 weeks and it was a breeze, the system froze not once.

i decided to get my cpu to work also with c-states.

my first try was with the following settings:

- XMP on

- Core Sync to SYNC ALL CORES

- VCORE in Adaptive Mode, no offset, no turbo voltage

- V_DRAM 1.20V (DEFAULT)

- VCCIO 1.15V

- VCCSA 1.2V

It was stable some hours, but i got a freeze eventually -> back to c-states off and everything back to auto -> no freeze for several hours

after some research i read about the "Multicore Enhancement option" and that it would sometimes misbehave in combination with skylake

i decided to give the following settings a try:

- XMP on

- Core Sync to SYNC ALL CORES

- Multicore Enhancement off

- VCORE Auto

- V_DRAM 1.20V (DEFAULT)

- VCCIO 1.15V

- VCCSA 1.2V

Result: still got a freeze.

3rd try:

- XMP off

- CORE SYNC to Auto

- Multicore Enhancement off

- VCORE Auto

- V_DRAM Auto (DEFAULT, AUTO setting will be 1.2V too)

- VCCIO 1.15V

- VCCSA 1.2V

Result: got no freezes anymore, but a BSOD with PAGE_FAULT_IN_NONPAGED_AREA (ntoskrnl.exe+16f5c1, nvlddmkm.sys+d0793)

the my final settings:

- XMP Off

- Multicore Enhancement off

- Core Sync Auto

- V_DRAM 1.25V

- VCCIO 1.15V

- VCCSA 1.2V

with the last settings i had no freezes since 3 days. i hope i found my settings - but i still think one or more components are failing to perform at default settings.

so i have trouble to pinpoint which part is failing, is it the CPU which cannot perform with VCCIO and VCCSA at Auto?

is it the Motherboard which cannot provide the right voltages levels?

is it the RAM which cannot perform at default speed?

i hope i did not write too much.

cheers,

thehackman

[LMonk]

Why i don't need voltage modification when i use such settings? You mensioned this few times already i think.

[CCris2]

lks70 because those voltages are for memory controller and some other circuits on CPU including igpu and I/O controller( that includes PCIE lanes on CPU ). If you have a lower timing or higher frequency memory VCCIO and VCCSA need to be higher, and it seems that some ASUS mainboards have problem with setting those voltages correctly when activating XMP. Those voltages should be automatically set when you activate XMP ...

[CCris2]

CPU voltage is only for CPU (only those 4 cores are powerd by the VCore). And freezing and BSOD errors are caused also by insufficient power to VCCSA and VCCIO .

[CCris2]

Hi thehackman ,

Can you post a ss with HWinfo voltages, and what are they set in bios right now ? I want to see if they are correct or there are problems.

Also if you have igpu and discrete card enabled your system may need more voltage on VCCSA, but first I want to see whats the voltage right now in HWiNFO.

[AKota1]

VCCSA is set to 1.2V

VCCIO is set to 1.15V (= CPU VCCIO)

V_DRAM is set to 1.25V

[LMonk]

i mean why i don't need to add offset to cpu after this changes? i know that it is stable with 1.1 and 0.125

[CCris2]

lks70 ,

You add offset only and only if it does't work.

[CCris2]

thehackman ,

Set Vcore to Adaptive, set memory on XMP, VCCIO 1.15V .

For VCCSA you need to add step by step voltage till you find the stable point, because I/O controller is forced more when you use a high performance video card + igpu active + memory on non standard timings. You should check in HWiNFO for each step you make because it doesn't correspond with what you set in bios. VCCSA can go safe till 1.28V in bios, in your case, after that you should be carefull (usually you get red numbers when it becomes dangerous, but those red numbers are for voltages you see in real in HWiNFO. There is +0.025 Voltage difference for VCCSA from what you set in bios ...

So if you set in bios white 1.25V in real is 1.275V ... and 1.275 V can be red. I hope you got the point.

But it maybe possible that you obtain stability with VCCSA at 1.25v or even lower.

I know it will be time taking but ...

[AKota1]

thank you ronin for your reply. but why do i need to up my dram voltage to 1.25 (default or auto would be 1.2V) for XMP off (so im using the JEDEC Profile) to get no BSODs?

[ssote]

thehackman wrote:

Hi there!

I have the same Problem as described by many in this thread, i will be more specific about my problem:

I bought the following components:

Intel i7 6700k

Asus Z170 Pro Gaming

Crucial Ballistix Sport LT grey DIMM Kit 16GB, DDR4-2400, CL16-16-16 (BLS2C8G4D240FSB)

EVGA 980Ti SC+

Corsair RMi 650 with Cablemod Sleeved Cables (approved for RMi Series)

Phanteks Evolv ATX

Note: I use a Multimonitor Setup with the Intel Onboard Graphics: Main Monitor @ Displayport @ 980ti and 2nd Monitor @ Displayport @ Onboard GFX

Note2: CPU-Z lists 4 Profiles for my RAM, 3 JEDEC (# 8, # 9, # 10) and one XMP - but the XMP Profile with 2400 CL16-16-16-39-55 does not differ from the JEDEC # 8 Profile with 2400 CL16-16-16-39-55

so when i write "XMP off" i just disable the XMP settings in the bios, the frequency still stays at 2400 with CL16-16-16-39-55

At first i had a MSI GTX780 but it died (VRM Chip vaporated) and i decided to buy a new 980 ti.

With the arrival of the EVGA 980Ti my Problems began:

I have always had very randomly freezes of my system (Screen freezes and very loid white noise from Headphones @ onboard sound), but only in games (BF4, Rise of the Thomb Raider, WoW, League of Legends, Diablo 3).

In Windows the System is Rockstable, i never had any freezes in Windows (even while stresstesting)

I have been trying to figure it out since february 2016 and tried the following things:

- driver updates (Intel graphics drivers, nvidia graphics drivers, as media usb controller drivers, intel inf drivers, etc...)

- bios updates (currently @ 1805, microcode 0x84)

- stresstesting:

1) Memtest86+ for 12 Hours - No Errors (Bot RAM Sticks)

2) Prime95 v2.66 for several Hours (blend test, ram test, etc...) - No Errors

3) Unigine Valley 1.0 for several Hours - No Errors

4) Prime95 v2.66 + Unigine Valley 1.0 for several Hours - No Errors

- switched 980ti out for a gtx 970 - it took longer for a freeze, but the freezes still persisted and occured

- fiddled again with the bios - still freezed

As i mentioned, the freezes only happen in games!

I recently startet playing Diablo 3 again and in this game the Problem is very Present.

I do alt-tab very often to verify items if they apply for a certain build or something - so lots of c state switching

sometimes i play d3 a day without any freezes, sometimes 4 freezes occur in 2 hours or less.

then i came across your thread

at first i decided to disable all c-states and speedstep - i tested it for 2 weeks and it was a breeze, the system froze not once.

i decided to get my cpu to work also with c-states.

my first try was with the following settings:

- XMP on

- Core Sync to SYNC ALL CORES

- VCORE in Adaptive Mode, no offset, no turbo voltage

- V_DRAM 1.20V (DEFAULT)

- VCCIO 1.15V

- VCCSA 1.2V

It was stable some hours, but i got a freeze eventually -> back to c-states off and everything back to auto -> no freeze for several hours

after some research i read about the "Multicore Enhancement option" and that it would sometimes misbehave in combination with skylake

i decided to give the following settings a try:

- XMP on

- Core Sync to SYNC ALL CORES

- Multicore Enhancement off

- VCORE Auto

- V_DRAM 1.20V (DEFAULT)

- VCCIO 1.15V

- VCCSA 1.2V

Result: still got a freeze.

3rd try:

- XMP off

- CORE SYNC to Auto

- Multicore Enhancement off

- VCORE Auto

- V_DRAM Auto (DEFAULT, AUTO setting will be 1.2V too)

- VCCIO 1.15V

- VCCSA 1.2V

Result: got no freezes anymore, but a BSOD with PAGE_FAULT_IN_NONPAGED_AREA (ntoskrnl.exe+16f5c1, nvlddmkm.sys+d0793)

the my final settings:

- XMP Off

- Multicore Enhancement off

- Core Sync Auto

- V_DRAM 1.25V

- VCCIO 1.15V

- VCCSA 1.2V

with the last settings i had no freezes since 3 days. i hope i found my settings - but i still think one or more components are failing to perform at default settings.

so i have trouble to pinpoint which part is failing, is it the CPU which cannot perform with VCCIO and VCCSA at Auto?

is it the Motherboard which cannot provide the right voltages levels?

is it the RAM which cannot perform at default speed?

i hope i did not write too much.

cheers,

thehackman

Have you tried down clocking your video card to stock clocks not the SC+ clocks? Why did you use the old version of prime95?

[CCris2]

If you set memory to XMP it will set automatically the voltage for DRAM to what was programmed. You don't need to set voltage.

I have an Asrock mainboard Z170 extreme 7 +, DDR4 3200. When I activate XMP DDR voltage goes to 1.35, VCCIO goes to 1.2V (Which is to high, on red, and I put it lower manually, because it works the same), and VCCSA to 1.25V.

[CCris2]

Also normally the memory on auto should go on JDEC coresponding to 2100.

I also checked and that memory is not supported by your mainboard. I think thats why you have those problems because mainboard doesn't have setting in bios for that memory at all.

Here is the list with supported memory models by your mainboard: http://dlcdnet.asus.com/pub/ASUS/mb/LGA1151/Z170-K/Z170_DDR4_4DIMM_memory_QVL_report160216.pdf?_ga=1.79646544.59267568.1456783981 http://dlcdnet.asus.com/pub/ASUS/mb/LGA1151/Z170-K/Z170_DDR4_4DIMM_memory_QVL_report160216.pdf?_ga=1.79646544.59267568.1…

[ssote]

Ronin wrote:

If you set memory to XMP it will set automatically the voltage for DRAM to what was programmed. You don't need to set voltage.

I have an Asrock mainboard Z170 extreme 7 +, DDR4 3200. When I activate XMP DDR voltage goes to 1.35, VCCIO goes to 1.2V (Which is to high, on red, and I put it lower manually, because it works the same), and VCCSA to 1.25V.

I talked to G.Skill and they said for 3200MHz All you have to set is vCCSA (System agent) to 1.150v that is for the memory controller voltage and other things.http://www.hardwaresecrets.com/understanding-all-voltage-configurations-from-the-motherboard/4/ Understanding All Voltage Configurations from the Motherboard - Intel Processors of 7 - Hardware Secrets

VCCIO: Available starting with the second-generation Core i CPUs ("Sandy Bridge"), this voltage is used for feeding all input/output (I/O) pins of the CPU, except memory-related pins. On CPUs that have this voltage, it is also used to feed the thermal control bus (PECI, Platform Environmental Control Interface).

Read more at http://www.hardwaresecrets.com/understanding-all-voltage-configurations-from-the-motherboard/4/# 2eL7s7kVPiSFQ2ey.99

How did you see your default voltages for vCCSA and vCCIO on XMP and what is the voltages with XMP off.

[ssote]

Ronin wrote:

thehackman ,

Set Vcore to Adaptive, set memory on XMP, VCCIO 1.15V .

For VCCSA you need to add step by step voltage till you find the stable point, because I/O controller is forced more when you use a high performance video card + igpu active + memory on non standard timings. You should check in HWiNFO for each step you make because it doesn't correspond with what you set in bios. VCCSA can go safe till 1.28V in bios, in your case, after that you should be carefull (usually you get red numbers when it becomes dangerous, but those red numbers are for voltages you see in real in HWiNFO. There is +0.025 Voltage difference for VCCSA from what you set in bios ...

So if you set in bios white 1.25V in real is 1.275V ... and 1.275 V can be red. I hope you got the point.

But it maybe possible that you obtain stability with VCCSA at 1.25v or even lower.

I know it will be time taking but ...

vCCIO is (I/O) Pins from the VRM. vCCSA Is (System Agent)memory controller 2D IGPU. The voltage It is not cut and dry like you say, the software can be off, the hardware voltage reading can be off. All manufactures are different in there voltage readings and quality control.

http://www.hardwaresecrets.com/understanding-all-voltage-configurations-from-the-motherboard/4/ Understanding All Voltage Configurations from the Motherboard - Intel Processors of 7 - Hardware Secrets

[CCris2]

wingman99

This is not the first mainboard that has these problems. On skylake VCCIO is important for C7 state of the CPU, and is the voltage for the memory controller and shared cache also.

This is another generation of CPU's, that document is from 2011. If you really want to read more, here you go : http://www.intel.com/content/www/us/en/processors/core/desktop-6th-gen-core-family-datasheet-vol-1.html 6th Generation Intel® Core™ Processor Family Datasheet, Vol. 1.

[AKota1]

thank you for your reply.

i did not test my evga 980ti @ stock clocks - but why would that be beneficial? i switched to a msi gtx 970 and still got freezes (less, but still freezes)

and since my rig is stable when i disable c states, i dont think its related to my gfx card.

i used the old prime version since it does not push the temps as high as the newer versions due to the changed instruction set. it would be a very synthetic test because no task would push the cpu as high as the newest version of prime95.

[CCris2]

It doesn't matter what clocks evga 980ti is, pcie buss should be 100 Mhz, if you don't overclock cpu by increasing bclk .

As for cstates, they should be left as the mainboard set them. I never touched cstates on my mainboard. I left them on auto. There are 2 cstates : CPU cstates and package cstates. Bios automatically disables package cstates in case of overclocking ( thats on my mainboard ).

I don't know what temps are, but normally the CPU should stay under 80 degress C no matter you do on that CPU. So if the CPU temps go higher then 80 lets say 90, at stock frequencies, even with prime, then you have a cooler problem.

[ssote]

thehackman wrote:

thank you for your reply.

i did not test my evga 980ti @ stock clocks - but why would that be beneficial? i switched to a msi gtx 970 and still got freezes (less, but still freezes)

and since my rig is stable when i disable c states, i dont think its related to my gfx card.

i used the old prime version since it does not push the temps as high as the newer versions due to the changed instruction set. it would be a very synthetic test because no task would push the cpu as high as the newest version of prime95.

Thanks, I heard that the new prime pushes the temps higher how much does it? Sorry I missed that you swapped cards then it would not be the graphics card causing the freeze. I had a freeze at Idle with skylake I had to disable c states and up my vcore 0.050v intel told me I should not need to do that and recommended RMA for the CPU.

When you run your memory on XMP do you still have BSOD, if you do have you run a PassMark Memtest86 to test for your memory?