cancel
Showing results for 
Search instead for 
Did you mean: 

BCLK (Bus Clock) Spikes from 100.3 (default) to 102mhz

PlaneName
Level 10
Hello Everyone!

I've been so busy with college, already missed the forums and the people who I had helped and that helped me 🙂

I am now having a major issue with the BCLK to spike causing instability issues on browsing the net or leave my PC on idle causing a BSOD (hal.dll) which obviously means my overclock is not stable. It has been doing this for the past year (I've used ThrottleStop to downclock it so it wont happen) and finally to my conclusion I realized that the cause of the issue was that my BUS CLOCK jumps high when on idle, or browsing the net and mainly working on programs in windows.

I have not notice this until I downloaded a program called HWiNFO64 and open CPU-Z to check what is going on and then I saw this: http://i.imgur.com/cdtdo40.jpg?1- 4.502ghz to 4685.ghz with the BCLK at 102.3mhz - This is very odd, I've been searching around and there are only a couple people who have notice this on "GIGABYTES" boards. Though my brand of motherboard is not efficient for overclocking especially when I need my VCORE to run at 1.80-1.92v for the i5 2500k to run @ 4.5ghz. (GA-Z68A B3H rev 1.3).

Yes, I have tried to clear CMOS and so on and still the BCLK will increase without even overcloking as well, though no BSOD. My temps are fine also, I have upgraded from the EVO to the NOCTUA NH-D14 recently and on prime95 the temps reaches to 68-70 C (core temps) and main CPU at 58 C which is fine and I don't see the BCKL to increase when benchmarking or playing games it only happens when it's on IDLE or in windows browsing the net and so on.

I keep digging and digging wondering why the BCLK jumps like that when I also tried to set it manually at 100.0 but it still go above that number and this is a very serious concern as the PCI-E frequency can cause instability issues to the GPU as well. I've also recently upgraded to the EVGA GTX 980 SC and I have noticed that my temps can be at 35 C then jumps to 45 C in less then a sec in the start of Heaven Benchmark 4.0 when gpu usage jumps to 99% GPU Usage. The temps does not spike very high like that where it should go slowly to the 40 C range, though this is what the "PCI-E" frequency can cause damage to the components.

- This had the same side effects on my GTX 560 ti (sold) and the GTX 760 (handed to my brother) that had followed and now really worried for my GTX 980 and the rest of my system.

It also had the "Power Cycling" issue that "GIGAYBTE" is well known for where I must NOT touch or move my RAM before it does this - https://www.youtube.com/watch?v=6J2Qd3RevPQ - This person has fixed their issue by replacing the CPU but mine was a very different issue as it's cause by the ram slots which I also have another pair and the same issue I MUST wrestle onto the slots. It was happening when I was changing coolers from the EVO to the Noctua where obviously I need to clean out my system by removing the ram and so on, though I think I am sure that it was a malfunctioning board due to the "PCI-E" frequency. Though I am not worried when I do this only if my RAM need to be replaced or adding new paste to the CPU cooler.

I also realized that when clearing the CMOS by removing the battery and so on the BCLK will be at default but its .3mhz more and will still have that spike to 102mhz which is still dangerous. So when overclocking to 4.5ghz it will be at 4.513mhz instead of around 4.502mhz which I have to set the BCKL MANUALLY to 100mhz or 99.8 but again without overclocking the BCKL will jump to 102.3! when it should stay at 100.0 or less or what ever I set to manually...

Conclusion:

I really really should go for the ASUS ROG mobo sometime but I really cannot afford it at the moment as I already saved up for the i7 4970k but not enough for a really efficient board I am tired from EVGA (old 680i) and now the GIGABYTE mobo. I really do need your help and I know there is a gigaybye forums but they also cannot help me and also explained it through support but again they do not provide helpful solutions or "think it's normal" and now lead me to the best knowledge of these issues are right here!

This is frustrating me I have tried everything from the BIOS settings and still don't know why it jumps like that and also I have found another youtuber with the same gigaybe motherboard that he returned to replace a new one and it was working BUT mine was a used board without warranty. If this issue cannot be fixed, please if there is anyone willing or have a spare old rog board for the 2nd/3rd gen intel CPU it will help me so much and willing to pay for the postage, so I can actually have a decent mobo without any issues. It's difficult for me to save money as I am still in college trying to pay for my subjects and equipment for IS (Information System).

I really do hate to ask this but now I am very concern that these issues were the ones that cause the GPU's to behave like that especially during idle due to the "PCI-E" abnormal frequency... This is now happening to my brand new EVGA GTX 980 SC ACX 2.0 which I still can RMA it but the temps does not spike on my other system so it still good just worried about long term effects on this mobo....

Again, I do apologize I'm very worried that my expensive components in my system will not last long... I really wish I had a mobo without issues I'm having really bad luck with "used" mobo's at the moment... should have saved more money on a brand new motherboard.

While I was typing this and checking HWiNFO again I've also found out why the temp spikes high in a sec - http://i.imgur.com/Fjwr7hB.jpg?1- PCI-E 5.0 does not exist still and my PCI-E frequency should be 2.0 due to the i5 2500k. Now I am unsure what is going on if its sensor or the actual motherboard.

Thanks for reading this and taking your time, it appreciates me allot.
Case: Antec 300 Mini Tower Motherboard: GA-Z68A-D3H-B3 (rev1.3) CPU: i5 2500k @ 4.5ghz Cooler: Hyper 212 EVO GPU: GIGABYTE GTX 560ti (950mhz) RAM: G.Skill 8 GB DDR3 @ 1600mhz PSU: Corsair CX 750 Driver: 320.00
Been building PC's for 8 Years! 😉
17,307 Views
14 REPLIES 14

elesde
Level 10
Just to let you know the same thing happens on my ASUS board. Granted it is not ROG and an AMD model but the base clock behaves in the same manner no matter what i do in BIOS.
At this point I am not even sure if it is the BCLK that is unstable or just the readout in Windows.

As for the PCIe frequency, afaik that is a separate clock not coupled to the BCLK. The bandwidth displayed in HWiNFIO is fine btw., it idles at PCIe x16@1.1 speeds and clocks up to x16@2.0 under load.

elesde wrote:
Just to let you know the same thing happens on my ASUS board. Granted it is not ROG and an AMD model but the base clock behaves in the same manner no matter what i do in BIOS.
At this point I am not even sure if it is the BCLK that is unstable or just the readout in Windows.

As for the PCIe frequency, afaik that is a separate clock not coupled to the BCLK. The bandwidth displayed in HWiNFIO is fine btw., it idles at PCIe x16@1.1 speeds and clocks up to x16@2.0 under load.


Thanks for your reply, I am also trying to find out if the sensors were wrong too but it cant be due to BSOD when system is on idle or browsing the net that is when I have opened up the two programs CPUZ and HWiNFO and waited for about 2 minute or so, that is when both clocks speed went up.

The BCLK is also connected to the PCI-E Frequency and the memory speed as last time I have checked to my findings. That is why most do not touch it and leave as it be as it can cause instability to other components. Also thanks for letting me know that you had also have the same issue then it must be the mobo as one have got his replaced and no issue with the BCLK to spike. So I am now in a spot that I should not worry about long term effects or go a safer reliable board without the BCLK to spike?

Also thanks for clarifying the bandwidth as I have read it wrong. That is the right speed for my GPU.
Case: Antec 300 Mini Tower Motherboard: GA-Z68A-D3H-B3 (rev1.3) CPU: i5 2500k @ 4.5ghz Cooler: Hyper 212 EVO GPU: GIGABYTE GTX 560ti (950mhz) RAM: G.Skill 8 GB DDR3 @ 1600mhz PSU: Corsair CX 750 Driver: 320.00
Been building PC's for 8 Years! 😉

elesde
Level 10
Oh yea looks like the BCLK & PCIe frequencies are linked on socket 1155 😞 , last Intel CPU I overclocked was an i7 920^^

By looking at the specs of your motherboard I would seriously consider replacing it with something a little more "high end" especially in the VRM department when overclocking.
The ROG Maximus might be a bit too high priced but there surely are other nice boards out there in the 100-150$ range.

elesde wrote:
Oh yea looks like the BCLK & PCIe frequencies are linked on socket 1155 😞 , last Intel CPU I overclocked was an i7 920^^

By looking at the specs of your motherboard I would seriously consider replacing it with something a little more "high end" especially in the VRM department when overclocking.
The ROG Maximus might be a bit too high priced but there surely are other nice boards out there in the 100-150$ range.


Again, thanks for your reply.

Yeah I am trying to find some used ROG motherboards on ebay but I think I am tired to find issues that the seller will not explain or have not notice. Australia have the most expensive hardware and seems I might need to use my savings from the i7 4970k to find at least another mobo brand that I have not tried with and stick to the second generation for this year then :(, I am really still worried about the GTX 980 will may take a huge hit in a long term.

I saved about $320 USD on Amazon for the i7 4970k but still waiting for a reasonable price drop when it was at $299 and my savings on by bank account for a reliable used z97 mobo in AU, I was really happy saving it until I realize that my mobo did really had issues from the start when my first GPU (GTX 560 Ti) was the first one got hit by this but again the spike on temp will not happen when in other system.

The temps are still in safe range as the temp at max is at 62 C on my GTX 980 but its that huge spike (35 to 45 C or more in less then a sec depends on the game-benchmark %99 GPU USAGE) that happens which it never had done before until a week ago. This was the same symptoms I had with both the GTX 560 ti and the GTX 760 when used about two or three months in this system. It only comes to the BCLK which I have not noticed it had done that spike 😞
Case: Antec 300 Mini Tower Motherboard: GA-Z68A-D3H-B3 (rev1.3) CPU: i5 2500k @ 4.5ghz Cooler: Hyper 212 EVO GPU: GIGABYTE GTX 560ti (950mhz) RAM: G.Skill 8 GB DDR3 @ 1600mhz PSU: Corsair CX 750 Driver: 320.00
Been building PC's for 8 Years! 😉

unclewebb
Level 10
HWiNFO is not accurate when reporting the BCLK. The variation it shows is usually not happening. The BCLK on most motherboards is very stable these days and if you were monitoring with an oscilloscope, you wouldn't see even +/- 0.001 MHz variation. Varying BCLK is mostly monitoring error that you are seeing. Try using CPU-Z 1.72 and RealTemp 3.70.

The monitoring timers within Intel CPUs are a shared resource. This means that if 2 different programs are running at the same time and trying to use the same timers, they can interfere with each other. For best results, only use 1 monitoring program at a time.

If your computer has a BSOD when lightly loaded or idle, that usually means the CPU is not getting enough voltage. Are you using offset voltage? Are you using the core C States or the package C States or both.

If you need to know what C States your CPU is using then try using this.

RealTemp T|I Edition
https://www.sendspace.com/file/55yvry

PlaneName
Level 10
unclewebb wrote:
HWiNFO is not accurate when reporting the BCLK. The variation it shows is usually not happening. The BCLK on most motherboards is very stable these days and if you were monitoring with an oscilloscope, you wouldn't see even +/- 0.001 MHz variation. Varying BCLK is mostly monitoring error that you are seeing. Try using CPU-Z 1.72 and RealTemp 3.70.

The monitoring timers within Intel CPUs are a shared resource. This means that if 2 different programs are running at the same time and trying to use the same timers, they can interfere with each other. For best results, only use 1 monitoring program at a time.

If your computer has a BSOD when lightly loaded or idle, that usually means the CPU is not getting enough voltage. Are you using offset voltage? Are you using the core C States or the package C States or both.

If you need to know what C States your CPU is using then try using this.

RealTemp T|I Edition
https://www.sendspace.com/file/55yvry


Hello unclewebb, thanks for your reply.

I've found out first by using CPU-Z and saw my core clock went higher 80-100mhz higher, that is when I have download HWiNFO and CPU-Z to run seeing if these were right together and both have the same results without each other, HWiNFO seems very accurate. I have already disabled C states, turbo and other related settings for the voltage and core clock to not decrease. Thanks for Realtemp as this is the one I was looking for ages, will also do my test on this one.

CPU is running at 1.368v-1.380v constant with the CORE VID at 1.361-1.371 and running prime 95 over 8 hours and it's fine - No sign of the BCKL to be increase. At idle the voltage does not lower as like I said those settings have been disabled. Last night I made a test that what ever voltages from 1.356v to 1.392v it will still cause a BSOD when left idle or browsing the net and using Microsoft word and will not BSOD during a gameplay or other related 3d programs (benchmarking and so on). So obviously my CPU does not need that much voltage on idle...

These are the only BIOS settings overclocking abilities I have: (The "offset" is the DVID)

- http://i40.tinypic.com/21m88s0.jpg - There are only two levels of LLC - The best one was Level 1 +.040v DVID (1.368v) for me as Level 2 will use higher voltages and still does not make a difference with the BCLK even when set to auto, intel spec or disabled it will still BSOD when using prime 95 for a minute or so due to using less voltages. Level 1 was much more efficient.

These are the CPU core features on BIOS:

http://i.imgur.com/gyy8t.jpg - I have disabled all settings and the CPU PLL at enable, auto or disabled still does not make a difference. I've also enabled EIST last night and the highest BCKL was at 101 which still is not the culprit. When enabling C1 or C3 it will BSOD on start up or in a few seconds on the window screen (for overclocking) - That is why they are disabled as this motherboard require allot of voltage for the CPU or the CPU itself.

This mobo is certainly not one of the most stable motherboards and there have been reports that some have these BCLK issues and they have it replaced AS it was a defective board... and I am still not sure what made the BCKL to behave like that during idle, browsing and other activities on windows such as watching a movie as well where it will BSOD most of the time there. Unclewebb that is what I have THOUGHT that it needed more voltage BUT it WASN'T and realize that the core and BCKL increases... that is why I cannot get a stable OC on those activities on windows or left idle.

It has been doing this for more than a year with this motherboard that is why I use Throttlestop to use lower clock speed (x33) on windows and at gaming or benching I turned it off. Like I said I thought the issue was it needed more voltage but I was wrong for the last year.

It should not BSOD in the first place from the unstable BCLK, I am now gonna report later with RealTemp and thank you for your help.
Case: Antec 300 Mini Tower Motherboard: GA-Z68A-D3H-B3 (rev1.3) CPU: i5 2500k @ 4.5ghz Cooler: Hyper 212 EVO GPU: GIGABYTE GTX 560ti (950mhz) RAM: G.Skill 8 GB DDR3 @ 1600mhz PSU: Corsair CX 750 Driver: 320.00
Been building PC's for 8 Years! 😉

unclewebb
Level 10
Are you sure that the AUTO setting in the bios is disabling all of your C States? You can use ThrottleStop to look at your C States. Make sure ThrottleStop shows all 0.0 if you want these all disabled. Most monitoring programs ignore if C1E is enabled or disabled. If all of your C States are disabled, the multiplier reported by ThrottleStop should be very steady when the CPU is idle.

RealTemp and ThrottleStop use a different method and a different timer to determine the BCLK. Every time you click on the BCLK button in ThrottleStop, it will measure your BCLK. I am curious. Does ThrottleStop show these BCLK changes? You have to keep clicking on the BCLK button. It only updates when you manually click on this button.

On most motherboards, ThrottleStop rarely reports a change of more than +/- 0.001 MHz.

If you are still having problems, there is a way to change what timers Windows uses. Windows has a few bugs when trying to measure time accurately.

Here is a program to test your Windows timers.

WinTimerTester
https://www.sendspace.com/file/xadvhe

Let the program run for about 100 seconds. It compares two different Windows timers and over the long run, they should both be showing the same thing.



I switched my Windows timer to a fixed frequency of 14.31818 MHz. You might get better results with this timer.

unclewebb wrote:
Are you sure that the AUTO setting in the bios is disabling all of your C States? You can use ThrottleStop to look at your C States. Make sure ThrottleStop shows all 0.0 if you want these all disabled. Most monitoring programs ignore if C1E is enabled or disabled. If all of your C States are disabled, the multiplier reported by ThrottleStop should be very steady when the CPU is idle.

RealTemp and ThrottleStop use a different method and a different timer to determine the BCLK. Every time you click on the BCLK button in ThrottleStop, it will measure your BCLK. I am curious. Does ThrottleStop show these BCLK changes? You have to keep clicking on the BCLK button. It only updates when you manually click on this button.

On most motherboards, ThrottleStop rarely reports a change of more than +/- 0.001 MHz.

If you are still having problems, there is a way to change what timers Windows uses. Windows has a few bugs when trying to measure time accurately.

Here is a program to test your Windows timers.

WinTimerTester
https://www.sendspace.com/file/xadvhe

Let the program run for about 100 seconds. It compares two different Windows timers and over the long run, they should both be showing the same thing.



I switched my Windows timer to a fixed frequency of 14.31818 MHz. You might get better results with this timer.


Again, thanks for your reply.

The images I have add was an image from google not my actual settings, they are all disabled as I said. Yes, realtemp showed my C States at all 0 which are disabled. 48296
The Corevoltage on CPU-Z is defiantly wrong as in HW monitor and others it's at 1.368v

Here is my windows timer:
48297

____

I've also just did a 30 minutes of Prime 95 by monitoring with RealTemp and no changes to the BCLK, I remember there was a graph log to save real time of the voltages and so on? I need to leave my PC for IDLE or just watch a movie and see if it the BCLK has gone higher or if there is more information about the voltage settings when it went BSOD.

Again I appreciate the help.
Case: Antec 300 Mini Tower Motherboard: GA-Z68A-D3H-B3 (rev1.3) CPU: i5 2500k @ 4.5ghz Cooler: Hyper 212 EVO GPU: GIGABYTE GTX 560ti (950mhz) RAM: G.Skill 8 GB DDR3 @ 1600mhz PSU: Corsair CX 750 Driver: 320.00
Been building PC's for 8 Years! 😉

unclewebb
Level 10
The C States AUTO setting in the bios is not always consistent from board to board or with different bios versions. I just wanted to double check since sometimes the C States are enabled and people don't realize it.

99.773 MHz is a typical BCLK and this usually doesn't vary from idle to full load.

Have you tried older versions of CPU-Z to see if any of them report your vcore correctly? It gets most boards right. Does CPU-Z report the same voltage at idle or full load?

Your timers look OK.

RealTemp and ThrottleStop let you enable the C1E C State. Does this cause a BSOD?