Sign in with
Sign up | Sign in
Your question

I can't push the Bclk to more than 190. help?

Last response: in Overclocking
Share
November 23, 2009 5:39:30 PM

Hey guys, I've asked you this question before but no suggestions worked. currently my i7 920 D0 is running at a bclk of 190 and with speed step its at 3.997 GHz. My temps are low (~24C idle, ~63C 100% load for a while). my system right now is very stable and I've had no problems, but when I up the bclk it freezes when it say starting windows. and its stuck for hours, but it never crashes. Ive tried boosting the Vcore and the QPI voltage, but nothing is succesful.

My current voltages:
dimm 1.65V
Vcore 1.325V
PLL 1.875V( I tried 1.8 too)
ICH and IOH 1.2V (I tried the defalts as well)
QPI 1.4V (i tried everything from 1.1)

Thanks guys! please help me! I know I can go higher but it won't let me.

More about : push bclk 190

a b K Overclocking
November 23, 2009 6:15:58 PM

Has your memory reached its limit?

Have your tried a pci-e bump up?
November 23, 2009 7:09:18 PM

the mem is below stock[by only about 88mhz)(corsair dominator 1600mhz C8) and timings are at default.
Related resources
Can't find your answer ? Ask !
November 23, 2009 7:09:59 PM

When I have the bclk at 200 my mem runs at spec.
November 23, 2009 7:11:23 PM

And how much of a pci bump? I have a 5870 and I'm worried what it might do to it.
a b K Overclocking
November 23, 2009 7:33:13 PM

My current I7 920 D0 setup on an Asus P6T deluxe V2, Corsair 7-7-7-20 1.65 volt ram 3x2GB, 1100 watt PS.

Ai Overclock Tuner Manual
CPU Ratio Setting 20
Intel SpeedStep Tech Disabled
BCLK Frequency 200
DRAM Frequency DDR3-1603MHz
UCLK Frequency 3208 MHz
QPI Link Data Rate 7218 MT/s
CPU Voltage 1.31250
CPU PLL Voltage 1.90
QPI/DRAM Core Voltage 1.35
DRAM Bus Voltage 1.64
CPU Differential Amplitude 1000mV
________________________
DRAM CAS# Latency 7
DRAM RAS# to CAS# Delay 7
DRAM RAS# PRE Time 7
DRAM RAS# ACT Time 20
________________________

This system is rock stable for my setup. It runs Prime 95 for 24hrs with no errors and has run Seti for the last 2 months with no crashes.
a b K Overclocking
November 23, 2009 7:39:21 PM

snowdude1492 said:
And how much of a pci bump? I have a 5870 and I'm worried what it might do to it.


You usually only need a pci-e bump up when pushing 220b-clock, but you could try it. Just try setting it to 101mhz.
November 23, 2009 8:37:51 PM

thanks, if that doesn't work should I try 102 or is that too high?
November 23, 2009 9:06:08 PM

ahnilated said:
My current I7 920 D0 setup on an Asus P6T deluxe V2, Corsair 7-7-7-20 1.65 volt ram 3x2GB, 1100 watt PS.

Ai Overclock Tuner Manual
CPU Ratio Setting 20
Intel SpeedStep Tech Disabled
BCLK Frequency 200
DRAM Frequency DDR3-1603MHz
UCLK Frequency 3208 MHz
QPI Link Data Rate 7218 MT/s
CPU Voltage 1.31250
CPU PLL Voltage 1.90
QPI/DRAM Core Voltage 1.35
DRAM Bus Voltage 1.64
CPU Differential Amplitude 1000mV
________________________
DRAM CAS# Latency 7
DRAM RAS# to CAS# Delay 7
DRAM RAS# PRE Time 7
DRAM RAS# ACT Time 20
________________________



This system is rock stable for my setup. It runs Prime 95 for 24hrs with no errors and has run Seti for the last 2 months with no crashes.



"UCLK Frequency 3208 MHz "

Wow, isn't that super low?
November 23, 2009 9:09:52 PM

PS, how do you copy all of your bio's settings? so that I can paste mine :)  to try and help
November 23, 2009 9:30:08 PM

idk. But i've tried other's settings and they don't help. But its worth another shot.
November 23, 2009 9:47:37 PM

I tried bumping the PCI to 101 and it didn't work. It gave me a blue screen while booting until I bumped the QPI to 1.425V and the Vcore to 1.3375. and then it did the same as before.
:( 
does this mean my i7 is faulty or the one of a milion that can't make 4 ghz?
November 24, 2009 12:15:28 AM

You have a lot of room to up your CPU voltage...keep it under 1.4, but try upping that some. If your board has auto setting you might raise the bclk to 200, leave everything else on auto and see if she boots from there...then start working your way down from the auto settings (auto settings always apply too much juice)
November 24, 2009 4:26:17 AM

okay I'll try that.
Also when I run stress tests It freezes and goes grey in a matter of seconds. I've been boosting the vcore but I've never got it to work. I boosted the QPI to 1.425 and its still the same.
Help there too?

Sorry for all the problems :( 
a b K Overclocking
November 24, 2009 5:11:23 AM

maybe you just hit the limit of this cpu.

What are the temps?
November 24, 2009 5:50:16 AM

low when on ~35-40C idle it crashes before i can see the load temp.
November 24, 2009 5:50:54 AM

Do you think its overheating, or its the literal limit.
a b K Overclocking
November 25, 2009 6:46:44 AM

snowdude1492 said:
low when on ~35-40C idle it crashes before i can see the load temp.


well the crashing is from instability....

so i would say this is the limit.
November 25, 2009 7:15:16 PM

why can other i7s do this at much lower vcores, and mine requires a much higher vcore and its not stable?
November 25, 2009 7:38:54 PM

Because of a bad batch.
November 25, 2009 7:40:29 PM

I have an ASUS P6T board with my i7 920. These are the settings for 3.6GHz

1.30V CPU
1.35V QPI
1.92V PLL
1.64V IMC
RAM Timings: 7-7-7-24
BCLK: 180
Multi: X20
Speedstep and turbo mode were turned off
Everything else was on auto

I stress tested for 22 hours using Prime 95 with small FFTs. I'm in a warm room now, so I need better cooling to be able to use HT/SMT without going over the max temp.
a b K Overclocking
November 25, 2009 11:05:09 PM

snowdude1492 said:
why can other i7s do this at much lower vcores, and mine requires a much higher vcore and its not stable?


They just have a really good batch.
November 25, 2009 11:53:59 PM

For reference, Toms uses 1.45v to overclock intels, and has for a long time. So expecting a huge overclock at 1.3 is not very realistic. Plus, a lot of people claim huge overclocks on low voltages without ever really stress testing it with prime 95. Sure, they might boot at that, but it won't pass a 100% cpu stress test.
November 26, 2009 2:08:38 AM

I agree more juice is needed.
November 26, 2009 3:17:34 AM

belial2k said:
For reference, Toms uses 1.45v to overclock intels, and has for a long time. So expecting a huge overclock at 1.3 is not very realistic. Plus, a lot of people claim huge overclocks on low voltages without ever really stress testing it with prime 95. Sure, they might boot at that, but it won't pass a 100% cpu stress test.


for the vcore, or the QPI?
November 26, 2009 4:08:08 AM

yes, on the cpu. Of course they don't have time to sit down and tweak for days like an end user can, so they go with high number at the edge of the safe range to save time....this was all explained in the i5 article where all the foxconn sockets fried at that voltage.
November 26, 2009 5:50:06 PM

belial2k said:
yes, on the cpu. Of course they don't have time to sit down and tweak for days like an end user can, so they go with high number at the edge of the safe range to save time....this was all explained in the i5 article where all the foxconn sockets fried at that voltage.


On Intel i7(Bloomfield), the max CPU Voltage is 1.375V on the 45nm process. I wouldn't call 1.45V safe if Intel states 1.375V as the highest. Sure, it can run higher, but it's obviously not recommended. I can hit 3.6GHz stable for the 22hr stress test, as I have posted above with only 1.3V on the CPU, but 3.8 BSODs with even 1.325 and 1.35V doesn't help much either. I can run 3.7GHz @ 195 BCLK X 19. This was stable for the 17 hour test.

http://processorfinder.intel.com/details.aspx?sSpec=SLB...
a b K Overclocking
November 26, 2009 7:44:36 PM

Well not to mention that at 1.375 it will be getting damn hot. i would call 1.45v relatively safe, but i dont want to get into an arguement about this again.

Read the data sheet on it and they say you can go up to 1.5 before you dramatically increase electron migration. But agan, i dont want to hijack this thread and get into this arguement again.
November 26, 2009 10:06:20 PM

As Overshocked says, I don't want to turn this into a "what is safe voltage" argument. I was simply pointing out his cpu voltage is considered fairly low in overclocking circles where 4.0ghz is the goal. I personally don't like to go over 1.4 for a 24/7 overclock as I stated in a previous post in this thread. That is because for the most parts it takes more and more voltage to get smaller and smaller gains past that point....But for the majority of OCers, 1.45v is still considered in the safe zone as long as you have good cooling and are running safe temps...Of course, intels specs are for the terrible stock cooler, which I would not recommending making any voltage increases on.
a b K Overclocking
November 27, 2009 12:15:14 AM

Agreed, if you keep the temps in check, the voltage will matter less.
November 27, 2009 4:21:05 AM

Thanks all of you guys! I have succsesfuly pushed my i7 920 to 4.1Ghz stably @ 1.3875 vcore (my temps on full load are 69~71)! I really could not have done this without all of you.

But I have one more question:

When overshocked said:
Read the data sheet on it and they say you can go up to 1.5 before you dramatically increase electron migration.

Ummmm, whats that? (I'm planning on pushing higher when I get the money for a liquid cooling system)
November 27, 2009 4:37:33 AM

hey guys. sorry to hijack ur thread, but what is the intel specified vcore max for i5 750 (if the i7 920 is 1.375V)?
and what is the default for OCing it?
November 27, 2009 4:42:56 AM

I heard its much lower for the i5, all of the i5 from toms burt out when they set the vcore to what the i7 oc's at. Just type it in on Google search (stupid answer but there are usually many answers)

I forgive you for hijacking my thread :p 
a b K Overclocking
November 27, 2009 6:08:18 AM

micky_lund said:
hey guys. sorry to hijack ur thread, but what is the intel specified vcore max for i5 750 (if the i7 920 is 1.375V)?
and what is the default for OCing it?


Ive heard 1.4v and 1.5v, havnt had the chance to read intels 80 page( [:thegreatgrapeape:5] ) document on the i5 yet, so i cant tel you exactly.
a b K Overclocking
November 27, 2009 6:12:43 AM

snowdude1492 said:
Thanks all of you guys! I have succsesfuly pushed my i7 920 to 4.1Ghz stably @ 1.3875 vcore (my temps on full load are 69~71)! I really could not have done this without all of you.

But I have one more question:

When overshocked said:
Read the data sheet on it and they say you can go up to 1.5 before you dramatically increase electron migration.

Ummmm, whats that? (I'm planning on pushing higher when I get the money for a liquid cooling system)


Actually im sorry, i believe its 1.55v.

It pretty much goes like this....

intel puts on their specs for the cpu that it shouldnt exceed 1.375v. But if you read their long ass technical document they specifically say that at 1.55v you will greatly increase elecron migration.

Moral of the story:

You can go up to 1.375 without much risk of electron migration. if you go above 1.375 there is a signifigant increase in electron migration. If you go above 1.55 there is a GREAT increase in electron migration. [:lectrocrew:2]
a b K Overclocking
November 27, 2009 6:13:55 AM

snowdude1492 said:
all of the i5 from toms burt out when they set the vcore to what the i7 oc's at.

That is from the sucky foxconn socket.
November 27, 2009 6:39:50 AM

srry again guys but I've been looking around and people require MUCH less vcores that I do on D0, and I have a D0 and its sucking up just as much volts as a C0/C1!

Why is that?
November 27, 2009 6:56:25 AM

If you have speedstep and turbo on it will effect your results, many OCers turn these off.
Every chip is different.
Some people do not stress test at full load
Some people are not honest
Every board has a different vdroop
Overall, your results fall in line with what I've see with most of my overclocks on the D stepping processors. Generally it takes me 1.38 1.4 to get a rock stable 4.0ghz.
a b K Overclocking
November 27, 2009 7:05:26 AM

snowdude1492 said:
srry again guys but I've been looking around and people require MUCH less vcores that I do on D0, and I have a D0 and its sucking up just as much volts as a C0/C1!

Why is that?

Well, every cpu is not the same.... you just dont have the magic batch like they do. Yours is actually pretty good.
November 27, 2009 7:21:51 AM

so....what is the default setting for the i5 750? is it like 1.300V? or will that do just to set a mark to OC from, or would 1.35V be better?
a b K Overclocking
November 27, 2009 7:40:34 AM

I dont know what stock voltage is for it, but just set it somewhere low like 1.2 and start from there.
November 27, 2009 8:21:04 AM

kk...this site here (pg 67, table 7.4) that the max is indeed 1.55V for vcore. toms uses 1.208vcore for overclocking to start with, so ill probably do the same.

however, it talks of VTT: what is this the voltage to? and y does toms overclock with this above 1.21 ( 1.308 to be exact ) sometimes, where the max is 1.21V? is this just to get the slight increase in stability at 4.3ghz, when it wouldn't work otherwise?

and what do u think is the PCH they talk about? do u think that 1.85V should do it, like the use for the test for 3.8ghz?

sorry to hijack, but creating my own thread normally ends up completely empty
thanks for any help
MICK
!