Sign in with
Sign up | Sign in
Your question
Solved

OC'ed i7-860 @3.60GHz stable in stress tests but randomly crashing after some time

Last response: in Overclocking
Share
a b K Overclocking
March 25, 2013 7:37:38 AM

As I mention in the title I got my i7-860 overclocked to 3.60GHz (180x20) with 1.2V. With CM212Evo cooler, system is stable as temps under load 70-80 C. Stress tests are fine (tried with both intel burn test and prime95), system works good and responsive but i got BSOD or direct crashes after some operating time (averagely after couple hours) anybody has any idea what may be the cause? you can see the my rig as my signature. thx...
a b K Overclocking
March 25, 2013 7:45:52 AM

could be alot of things..how old is your system? it may be possible that capacitor wear has set in on your PSU giving you unstable power as it heats up...or your capacitors on your mother board could be going again making you unstable after a set amount of time...

what are all your voltages...not just CPU

also 80C is a bit hot for a cpu, while i have done this myself and currently do on another rig (for me its an experiment on how long it will last)...it can make your system unstable.
m
0
l
a b K Overclocking
March 25, 2013 7:55:01 AM

atomicWAR said:
could be alot of things..how old is your system? it may be possible that capacitor wear has set in on your PSU giving you unstable power as it heats up...or your capacitors on your mother board could be going again making you unstable after a set amount of time...

what are all your voltages...not just CPU

also 80C is a bit hot for a cpu, while i have done this myself and currently do on another rig (for me its an experiment on how long it will last)...it can make your system unstable.


well its a 3-year old system with gigabyte ultra durable motherboard. and those temps are under heavy load in stress testing. i could get those crashes while even in idle mode or net surfing. so i dont really think it can be a capacitor problem.
i keep track of temps and even when playing crysis 3 max temps are like 60-65 C...
my voltages are almost standard level. i only raised DRAM to 1.6V and QPI to 1.21V and cpu is 1.2V. by the way i tried raising or dropping the voltages (thinking they are not enough for 3.6G frequency...) didnt help either...
m
0
l
Related resources
a b K Overclocking
March 25, 2013 7:59:23 AM

n1ghtr4v3n said:
As I mention in the title I got my i7-860 overclocked to 3.60GHz (180x20) with 1.2V. With CM212Evo cooler, system is stable as temps under load 70-80 C. Stress tests are fine (tried with both intel burn test and prime95), system works good and responsive but i got BSOD or direct crashes after some operating time (averagely after couple hours) anybody has any idea what may be the cause? you can see the my rig as my signature. thx...


thats a bit hot... 70-80c is about the max you should ever be running your intel cpu... ever.

i get the feeling you didn't run prime very long... generally you should let it burn for 10-24 hours to see how stable your system is. that said your temps are already too high for me to suggest you give your system a 24hour stress test.

Get those temps down a bit.

Sidenote, all that heat getting put out by your cpu will affect parts around it pretty drastically, i'd be willing to bet your problem is mostly the ram is overheating. You might want to look into the airflow inside your case and make sure it's good enough. Pop the side of the case off and stick a fan in the side of it. if your temps drop by more then 5C you have an airflow issue in your case. If that's the case you might be able to stabilize your system if you can fix that airflow issue.
m
0
l
a b K Overclocking
March 25, 2013 8:00:11 AM

n1ghtr4v3n said:
atomicWAR said:
could be alot of things..how old is your system? it may be possible that capacitor wear has set in on your PSU giving you unstable power as it heats up...or your capacitors on your mother board could be going again making you unstable after a set amount of time...

what are all your voltages...not just CPU

also 80C is a bit hot for a cpu, while i have done this myself and currently do on another rig (for me its an experiment on how long it will last)...it can make your system unstable.


well its a 3-year old system with gigabyte ultra durable motherboard. and those temps are under heavy load in stress testing. i could get those crashes while even in idle mode or net surfing. so i dont really think it can be a capacitor problem.
i keep track of temps and even when playing crysis 3 max temps are like 60-65 C...
my voltages are almost standard level. i only raised DRAM to 1.6V and QPI to 1.21V and cpu is 1.2V. by the way i tried raising or dropping the voltages (thinking they are not enough for 3.6G frequency...) didnt help either...


if your rig is that old and been overclocked the whole time or even a lot of it...it most likely is a capacitor problem. i just had to swap a board out 3 weeks ago for that very reason and it was a very high end EVGA board with 10 phase power. i think your over confident in your boards durability. it could be your ram as well....might try clocking it down and see if that helps.
m
0
l
a b K Overclocking
March 25, 2013 8:02:55 AM

atomicWAR said:
n1ghtr4v3n said:
atomicWAR said:
could be alot of things..how old is your system? it may be possible that capacitor wear has set in on your PSU giving you unstable power as it heats up...or your capacitors on your mother board could be going again making you unstable after a set amount of time...

what are all your voltages...not just CPU

also 80C is a bit hot for a cpu, while i have done this myself and currently do on another rig (for me its an experiment on how long it will last)...it can make your system unstable.


well its a 3-year old system with gigabyte ultra durable motherboard. and those temps are under heavy load in stress testing. i could get those crashes while even in idle mode or net surfing. so i dont really think it can be a capacitor problem.
i keep track of temps and even when playing crysis 3 max temps are like 60-65 C...
my voltages are almost standard level. i only raised DRAM to 1.6V and QPI to 1.21V and cpu is 1.2V. by the way i tried raising or dropping the voltages (thinking they are not enough for 3.6G frequency...) didnt help either...


if your rig is that old and been overclocked the whole time or even a lot of it...it most likely is a capacitor problem. i just had to swap a board out 3 weeks ago for that very reason and it was a very high end EVGA board with 10 phase power. i think your over confident in your boards durability. it could be your ram as well....might try clocking it down and see if that helps.


+1
and this guy knows what he's talking about too... that said i like to troubleshoot the easy stuff first. Lets make sure its not the airflow in your case and temps on the "parts" on your board before we go spending cash on a new MB.
m
0
l
March 25, 2013 8:05:01 AM

Indeed, atomicWar has many points that may cause unstability, still, you should try to lower clocks and test, also RAM timings may affect stability under load. As many people know, burn tests like prime and intel burn test, stress out CPU and little to nothing RAM.

You may try lower RAM frecuency, higher timings or even take away one module to see of any of them is working wrong.

Also a good test may be just playing games, or benchies as Ungine Heaven or something like that... When you overclock, try with those programs at 1st before actually get into a game.

You should note that unstable voltages is a real problem when overclocking, Vdrops will affect stability as energy saving features. Be sure to disable all of them and have a fixed voltage.

Many things will play important rolls in overclocking, just be sure that all you parts are working properly and can be stressed out before overclock. The best way is stress your system at stock and then, little by little increase the frequencies.

Hope it helps to give you a better image of what may be... Best of lucks.
m
0
l

Best solution

a b K Overclocking
March 25, 2013 8:06:13 AM

so kind ingtar thank you....and i agree lets rule everything else out first. but i just wanted you prepared for the potential bad news.


if hyper threading is enabled..disable it...make sure speed step is off is off too. set your ram timings to default and then try a burn in for a couple hours. you could also enable Vdroop for more stable power but be warned your temps will go up a few degrees C at least.
Share
a b K Overclocking
March 25, 2013 8:11:52 AM

well... last week i totally stripped the case... and rearranged all the cables and stuff... believe me its empty inside like a highway lol... and yea added the new cooler with push pull since i wanted to go for OCing.

and also I've been doing OC for the 1st time (started OCing it like 2 weeks ago) since i purchased the system. so i am skeptical that the mobo could be damaged...

the funny thing is stock voltages (when i load optimized defaults) are even higher than my overclocked setup. so i am confused.

maybe i should downclock ? (but i saw people on forums clocking a i7-860 even more than 4.0GHz with hyper evo cooler)
m
0
l
a b K Overclocking
March 25, 2013 8:14:55 AM

uhm.. yea i should have mentioned. ive already set those up beforehand.

LLC: enabled

turbo: disabled
c1: disabled
eist: disabled

i am thinking to loosen timings of ram i guess... what do you think? they were originally CL9 but i run them CL8-1T
m
0
l
a b K Overclocking
March 25, 2013 8:21:40 AM

yeah loosen the ram first thing
m
0
l
a b K Overclocking
March 25, 2013 8:39:20 AM

thanks for the help guys... i'll try those tips tonight... cheers... let you know if it gets more interesting lol
m
0
l
a b K Overclocking
March 26, 2013 8:23:48 AM

n1ghtr4v3n said:
thanks for the help guys... i'll try those tips tonight... cheers... let you know if it gets more interesting lol


I had a bear of a time getting my PhIIx4 965BE stable on any OC over 3.7ghz... and i knew from the crashes and behavior it was a ram instability issue.

In the end i had to push more voltage into the ram then i wanted to believe it needed, even bump up the voltage to the nb and a few other parts, downclock (this step also fried my mind, as it was really counter entuitive) my HT/CPU-NB and a number of other things to work it out.

The long and short of this is, sometimes when you're overclocking, certain parts will give you a headache... it doesn't mean they're broken persay... it just means you've not hit on whats needed to make it work together yet. Granted nothing i did to stabalize my oc will help you, but the moral of the story is you'll probably have to experiment a bit, and find the sweet spot.
m
0
l
a b K Overclocking
March 27, 2013 5:00:21 AM

ingtar33 said:
n1ghtr4v3n said:
thanks for the help guys... i'll try those tips tonight... cheers... let you know if it gets more interesting lol


I had a bear of a time getting my PhIIx4 965BE stable on any OC over 3.7ghz... and i knew from the crashes and behavior it was a ram instability issue.

In the end i had to push more voltage into the ram then i wanted to believe it needed, even bump up the voltage to the nb and a few other parts, downclock (this step also fried my mind, as it was really counter entuitive) my HT/CPU-NB and a number of other things to work it out.

The long and short of this is, sometimes when you're overclocking, certain parts will give you a headache... it doesn't mean they're broken persay... it just means you've not hit on whats needed to make it work together yet. Granted nothing i did to stabalize my oc will help you, but the moral of the story is you'll probably have to experiment a bit, and find the sweet spot.


yea you are right... since every build and and piece is unique not every solution works. its just i dont understand how people gets stable oc above 3.8GHz... because i cant even run more than 180bclk * 20 which is 3.60 and thats also barely stable. I guess thats how far i can push my old bud :( 
(note: i recently found out that more than 6.0 GHz qpi link speed is bad for a lynnfield and also memories doesnt really like tight timing when they are also matched with high frequency at the same time)

in the end;
i kinda get @3.60 stable(ish) with only pushing; (didnt test with a long stress run, but 75-80C stable temps and didnt crash with IBT, prime95, occt with 10 passes on standard or higher runs)

-bclk to 180 (above 180bclk nothing works for me, i could log on to windows8 with 4.0GHz (200*20, cpu 1.3V and qpi 1.31V) but it only runs until you start the stress test of any kind lol),
-cpu voltage at 1.2, qpi voltage 1.9, dram voltage 1.66,
-rams timing manually set to normal values,
-also other than stated above, all voltages manually set to normal (since when u leave them auto, mobo sets them much higher which i find i dont need it only generates extra heat),
-also LLC enabled (eventhough it shortens the usefull lifespan of cpu) makes things better since it dampens the voltage peaks.

thanks for your opinions again. i guess i will keep on experimenting until i burn stuff, literally :) 
m
0
l
a b K Overclocking
March 27, 2013 5:48:15 AM

i wonder if your mobo is just the weak link. how old is it. maybe with the ram switch it's finally showing it's lack of OC.
m
0
l
a b K Overclocking
March 27, 2013 6:06:42 AM

atomicWAR said:
i wonder if your mobo is just the weak link. how old is it. maybe with the ram switch it's finally showing it's lack of OC.


its like 3 years old...
well about the memories they actually dont have heat spreader on them maybe thats the problem ? but im not really pushing them hard (1440MHz, 1.66V, standard timings O_o) or maybe i shouldnt even bother DRAM volts since its close to default (considering... no heat spreader with higher volts is a NG ? )
m
0
l
a b K Overclocking
March 27, 2013 6:09:14 AM

i have many a mother board die at the 3 year mark. i think its a decent possibility. capacitor wear will get you eventually. do you have a buddies board you could borrow and try to see if there is a difference.
m
0
l
a b K Overclocking
March 27, 2013 6:18:39 AM

uhm... not really, i got my little brothers h55 board though... but i think they kind of different anyways...
m
0
l
a b K Overclocking
March 27, 2013 6:21:44 AM

not saying all boards die i have some make 5 or 6 years...but the 3 year mark is like right of passage for mother boards in my eyes. they live to four years you likely get it to make 6.
m
0
l
a b K Overclocking
March 28, 2013 1:36:21 AM

yeah so....

Finally i got a stable oc for my i7-860 with better temps, stress tested in IBT, prime95 and OCCT respectively not a single crash, also played games Crysis3, Skyrim, Path of Exile for about 2hours each, seemed no issues... i assume this will be final...

Hopefully anyone else with a Lynnfield wont have the same problem with me and these would help...
Cheers...

i7-860 @3.60GHz, 65-67C Temps Max (100% cpu load, 92% memory load, these are in stress test)
memory @1440MHz, CL9

c1/eist/turbo: off
c3/c6/prochot/all threads/thermal: on

cpu ratio: x20
qpi ratio: x32
bclk freq: 180
pci freq: 100

memory mult: x8
memory perf: turbo (turbo mode sets it automatically CL9)
timings: all auto

LLC: on
vcore: 1.15V
qpi/vtt: 1.19V
dram: 1.60V
all other voltages: normal (not auto)
smart fan control: off (so fans run at max rpm)
quick boot: on

GPU O/C:
gpu volt: 988mV
base freq: 800MHz
gpu freq: 1600MHz
memory freq: 2000MHz
smart user-set auto
m
0
l
a b K Overclocking
March 28, 2013 1:38:03 AM

happy to hear your all golden bro! i hope i was of help.
m
0
l
a b K Overclocking
March 28, 2013 1:40:27 AM

atomicWAR said:
happy to hear your all golden bro! i hope i was of help.


yea it took 2 weeks but finally it works :D  i never wanted push it above 4GHz anyways but desperately wanted to run at least 3.60 :) 
thanks...
m
0
l
a b K Overclocking
March 28, 2013 1:43:09 AM

anytime...let me know if i can help you with anything else man.
m
0
l
!