![[ch]](/mwf/v12/flags/ch.png "Switzerland")
Hi folks
On my computer mentioned higher, I have a Rybka mark of aroun 320 kn/s which seems to me too low.
By the way, launched a Fritz mark test and optained the following result : 4899
This seems very low as I think I should normally be around 8000.
Any idea whats happening ?
Thanks for your support
Best regards
On my computer mentioned higher, I have a Rybka mark of aroun 320 kn/s which seems to me too low.
By the way, launched a Fritz mark test and optained the following result : 4899
This seems very low as I think I should normally be around 8000.
Any idea whats happening ?
Thanks for your support
Best regards
![[gb]](/mwf/v12/flags/gb.png "United Kingdom")
If you can open windows task manager and check the performance tab, while Rybka or Fritz benchmark is running, you should see all 4 processors running at maximum.
You should also run CPU-Z at the same time, which will tell you if there is any speedstep interference going on. You should check the core speed and multiplier stay at the settings you picked in the bios.
Also on CPU-Z, check the memory tab, and see what speed your ram is at.
This might point to where the problem is.
On Fritz benchmark I get 9350 at a slightly faster CPU speed than yours, so 8000 would be about what I would be expecting
You should also run CPU-Z at the same time, which will tell you if there is any speedstep interference going on. You should check the core speed and multiplier stay at the settings you picked in the bios.
Also on CPU-Z, check the memory tab, and see what speed your ram is at.
This might point to where the problem is.
On Fritz benchmark I get 9350 at a slightly faster CPU speed than yours, so 8000 would be about what I would be expecting
![[ee]](/mwf/v12/flags/ee.png "Estonia")
There might be many reasons for performance loss, but as recent days were pretty hot I would recommend you check temperature of your CPU. Maybe it is too high and thermal throttling decreases your CPU performance. Failing cooler combined with hot days is quite often enemy of the computer enthusiast.
Hi all,
thanks for your answers.
I lunched CPU-Z and things are not clear for me :
normally its a multiplier overclocked QX6700 of 12x
When looking at CPU-Z, it was jumping sometimes from 12x to 6x and then back to 12x etc...
and also normally speed is 3.2 GHz and it was also jumping to 1.6 GHz and then back to 3.2 GHz, etc...
I tried to come back at factory settings with 10x multiplier and 2.66 GHz and had a better result with a Fritz mark of 6695.
But this number is still under what should be expected for this CPU and when looking at CPU-Z it seemed that the CPU is still switching to lower multiplier and speed but less than when I am at 3.2 GHz with 12x multiplier.
Can I do to something to avoid this problem ?
Thanks for your help and support.
Best regards
thanks for your answers.
I lunched CPU-Z and things are not clear for me :
normally its a multiplier overclocked QX6700 of 12x
When looking at CPU-Z, it was jumping sometimes from 12x to 6x and then back to 12x etc...
and also normally speed is 3.2 GHz and it was also jumping to 1.6 GHz and then back to 3.2 GHz, etc...
I tried to come back at factory settings with 10x multiplier and 2.66 GHz and had a better result with a Fritz mark of 6695.
But this number is still under what should be expected for this CPU and when looking at CPU-Z it seemed that the CPU is still switching to lower multiplier and speed but less than when I am at 3.2 GHz with 12x multiplier.
Can I do to something to avoid this problem ?
Thanks for your help and support.
Best regards
![[de]](/mwf/v12/flags/de.png "Germany")
Under normal conditions this is no problem. Thats called speedstep - a function of saving energy when beeing idle.
But you can try to switch that off - you'll find it in the BIOS - it is called EIST.
But you can try to switch that off - you'll find it in the BIOS - it is called EIST.
![[kw]](/mwf/v12/flags/kw.png "Kuwait")
Yes your speedstep is kicking in ... that is usually only useful in laptops. Get rid of it ... also check for thermal throttling ... some motherboards will throttle your cpu when it reaches a certain temperature ... make sure that is off or very high (like 85C).
What you are seeing is what Intel calls Thermal Monitor 2 (TM2). It is described in Intel document 31559205 in section 5.2.2 on page 77. TM2 is an internal mechanism that "causes the processor to adjust its operating frequency (via the bus multiplier) and input voltage (via the VID signals)." The best way for you to deal with this would be to improve the system that cools your processor.
This is not speedstep, a technology Intel provides on mobile processors to reduce power usage while the processor is idle. TM2 is designed to prevent long-term damage to the CPU and is being tripped by excessive core temperatures.
Regards,
Alan
This is not speedstep, a technology Intel provides on mobile processors to reduce power usage while the processor is idle. TM2 is designed to prevent long-term damage to the CPU and is being tripped by excessive core temperatures.
Regards,
Alan
![[fi]](/mwf/v12/flags/fi.png "Finland")
Does other processors except extreme editions support Thermal Monitor 2 ? I am expecially interested in Q6600
Is there any way to turn off TM2 ? - It is not wise thing to do but: is it possible to accidentially burn the processor, which has TM2?
Is there any way to turn off TM2 ? - It is not wise thing to do but: is it possible to accidentially burn the processor, which has TM2?
Seems like thermal throttling to me. Check your CPU cooler!
It is practical to clean the cooler from dust and debris before hot season each year. Use compressed air or soft brush for this. Stock cooler usually is not enough for quattros at high load. Maybe it is wise to replace it with some much efective one.
It is practical to clean the cooler from dust and debris before hot season each year. Use compressed air or soft brush for this. Stock cooler usually is not enough for quattros at high load. Maybe it is wise to replace it with some much efective one.
Hi,
If you read http://www.hiarcs.net/forums/viewtopic.php?t=68 you will see I was having a prob with HIARCS loseing on time on my backup machine, for me all the problems were related to Intels speedstep stuff which I had forgotten to turn off .... On my 4Ghz+ quad with it all turned off these problems never happened
Regards
If you read http://www.hiarcs.net/forums/viewtopic.php?t=68 you will see I was having a prob with HIARCS loseing on time on my backup machine, for me all the problems were related to Intels speedstep stuff which I had forgotten to turn off .... On my 4Ghz+ quad with it all turned off these problems never happened
Regards
As much as I know, the idea of speedstep is to decrease CPU power consumption and reduce noise of the cooler when CPU is idle. Under heavy load the speedstep should not reduce CPU cycles.
I think, that the speedstep might get involved under heavy load only when you use some fancy super silent BIOS settings. If BIOS settings are optimised for performance, the speedstep should not significantly influence the performance.
I think, that the speedstep might get involved under heavy load only when you use some fancy super silent BIOS settings. If BIOS settings are optimised for performance, the speedstep should not significantly influence the performance.
This is called Enhanced Halt State (C1E) and is enabled by default. Disable it in the BIOS.
Enhanced Intel SpeedStep (EIST) is disabled by default in desktop CPUs. If its enabled then Disable it.
Do not disable Thermal Monitor 2 (TM2) as it can protect your CPU from overheating
Enhanced Intel SpeedStep (EIST) is disabled by default in desktop CPUs. If its enabled then Disable it.
Do not disable Thermal Monitor 2 (TM2) as it can protect your CPU from overheating
Extended Halt State is used to implement power down modes for low power processing and is not invoked during normal mode to reduce excessive temperature readings. See page 84 and Figure 21 from Intel document 31559205. There is little doubt that the initial post referred to TM2 and the only safe solution is better cooling to prevent this.
Regards,
Alan
Regards,
Alan
I have identical processor as Bouddha. Enhanced Halt State (C1E) is Enabled by default in the BIOS (Gigabyte 965P-DQ6)
When it is enabled then I get Identical effects as Bouddha gets. When Disabled then the CPU is at 100% speed.
When it is enabled then it will reduce CPU speed Instantly Regardless of the CPU temp. This is 100% tested.
You can use RMClock to verify that.
QX6700 does support Enhanced Halt State (C1E): http://processorfinder.intel.com/details.aspx?sspec=sl9ul
When it is enabled then I get Identical effects as Bouddha gets. When Disabled then the CPU is at 100% speed.
When it is enabled then it will reduce CPU speed Instantly Regardless of the CPU temp. This is 100% tested.
You can use RMClock to verify that.
QX6700 does support Enhanced Halt State (C1E): http://processorfinder.intel.com/details.aspx?sspec=sl9ul
The question was not whether the Q6700 supports Enhanced Halt State. It does, and this is clearly documented in the link I gave above. The question is whether Enhanced Halt State is responsible for modulating the multiplier under high load conditions. The Intel documentation and state diagram are totally unambiguous on this issue (it doesn't). So your results indicate that either:
1) The Intel documentation is totally wrong, or
2) The Intel documentation is correct and the BIOS is turning off both Enhanced Halt State and TM2.
I would go with the Intel documentation being correct on this issue. The implications are that the processor is being run at temperatures above established guidelines. This may or may not be OK. If it were my system, I would think about improving CPU cooling unless I considered the processor to be expendable.
Alan
1) The Intel documentation is totally wrong, or
2) The Intel documentation is correct and the BIOS is turning off both Enhanced Halt State and TM2.
I would go with the Intel documentation being correct on this issue. The implications are that the processor is being run at temperatures above established guidelines. This may or may not be OK. If it were my system, I would think about improving CPU cooling unless I considered the processor to be expendable.
Alan
This is not my point.
When Enhanced Halt State (C1E) is Enabled, which is the default in my BIOS, then Identical effects as Bouddha occurs.
Mr. Bouddha, please do the following:
1. Get into the BIOS
2. Disable Enhanced Halt State (C1E)
3. Disable Enhanced Intel SpeedStep (EIST)
4. Disable Thermal Monitor 2 (TM2) (but I recommend you enable it later)
5. Save, reboot, run CPUZ, run CoreTemp or your mobo CPU temp tool, then run Fritz mark
6. Monitor you CPU temp and speed
7. You can enable TM2 and retest
Hope things workout well for you
When Enhanced Halt State (C1E) is Enabled, which is the default in my BIOS, then Identical effects as Bouddha occurs.
Mr. Bouddha, please do the following:
1. Get into the BIOS
2. Disable Enhanced Halt State (C1E)
3. Disable Enhanced Intel SpeedStep (EIST)
4. Disable Thermal Monitor 2 (TM2) (but I recommend you enable it later)
5. Save, reboot, run CPUZ, run CoreTemp or your mobo CPU temp tool, then run Fritz mark
6. Monitor you CPU temp and speed
7. You can enable TM2 and retest
Hope things workout well for you
Thanks to all of you for the interesting posts and answers.
I learned a lot
thanks
I learned a lot
thanks
If your CPU is clocked down while beeing under full load, the problem shurely is thermal throtteling as described by Allan. EIST is something that is used when beeing idle (sorry- i didn't read all your posts at first). EIST and thermal throtteling look the same in CPU-Z, but under different circumstances.
You'll have to improove CPU cooling - is your cooler fixed allright? Did you use some themal compound? What kind of cooler are you using?
It is hard to overclock a quad core chip when cooling isn't good. If you only have got an Intel boxed cooler, you should not set VCore higher, you might even get better results when lowering it.
You'll have to improove CPU cooling - is your cooler fixed allright? Did you use some themal compound? What kind of cooler are you using?
It is hard to overclock a quad core chip when cooling isn't good. If you only have got an Intel boxed cooler, you should not set VCore higher, you might even get better results when lowering it.
![[it]](/mwf/v12/flags/it.png "Italy")
I attach to the present two small programs that can help to increase the speed
of your engines on your PC.
The first one, GameXp, tax Windows so that to perform more quickly the games.
It intervenes on the management of the cache, on some parameters of Windows and it reduces
to the least one the functions of Windows (what you can restore in expectancy. The program performs a backup of the register of system so that you can put everything as to the beginning when you will want).
The second, Clock Generator, serves to overcklocare your PC.
I recommend you not to change the clock of the video card. To do this in the window,
of the program "PLL setup", it selects the box "Ignore GSB/PCI." In this way
you won't change the speed of the bus AGP and PCI.
Know however that the increase of speed that you will have, will be modest.
The limits of today's PC are these.
of your engines on your PC.
The first one, GameXp, tax Windows so that to perform more quickly the games.
It intervenes on the management of the cache, on some parameters of Windows and it reduces
to the least one the functions of Windows (what you can restore in expectancy. The program performs a backup of the register of system so that you can put everything as to the beginning when you will want).
The second, Clock Generator, serves to overcklocare your PC.
I recommend you not to change the clock of the video card. To do this in the window,
of the program "PLL setup", it selects the box "Ignore GSB/PCI." In this way
you won't change the speed of the bus AGP and PCI.
Know however that the increase of speed that you will have, will be modest.
The limits of today's PC are these.
Attachment: gamexp.zip (447k)
Attachment: ClockGen.zip (328k)
Hi all,
here are some result of my investigations based on your inputs :
In the bios I have a function which is called "CPU TM function"
When this function is disabled it seems that Core speed and multiplier dont change.
Strangly the Fritzmark was around 6.5k which is sinificantly lower than when I just received the computer and tested at around 7.5k.
Please notice that I came back to factory frequency of 2.67 GHz (10x266)
When setting back the "CPU TM function" to enabeled, the Core speed coult jump from 2667 to 1600 and Multiplier from 10x to 6x
strangly this time I obtained a better Fritzmark of around 7.1k which is much closer to what I had when just received my computer. (around 7.5k)
So should I assume that I have termal temperature problems ?
I did never open the case to remove dust in my computer, so maybe this can help ?! But it is not very convinient to do especially as I have to remove all cables ans so on...
Do you think this is linked to termal problem ?
What can I do ?
I mean I dont think that something has to be changed on the cooling except dust as the computer was working very fine the first months.
thanks again for your help and answers
best regards
here are some result of my investigations based on your inputs :
In the bios I have a function which is called "CPU TM function"
When this function is disabled it seems that Core speed and multiplier dont change.
Strangly the Fritzmark was around 6.5k which is sinificantly lower than when I just received the computer and tested at around 7.5k.
Please notice that I came back to factory frequency of 2.67 GHz (10x266)
When setting back the "CPU TM function" to enabeled, the Core speed coult jump from 2667 to 1600 and Multiplier from 10x to 6x
strangly this time I obtained a better Fritzmark of around 7.1k which is much closer to what I had when just received my computer. (around 7.5k)
So should I assume that I have termal temperature problems ?
I did never open the case to remove dust in my computer, so maybe this can help ?! But it is not very convinient to do especially as I have to remove all cables ans so on...
Do you think this is linked to termal problem ?
What can I do ?
I mean I dont think that something has to be changed on the cooling except dust as the computer was working very fine the first months.
thanks again for your help and answers
best regards
CPU TM function is CPU Thermal Monitor. This is the one that can be disabled. There is another one thats not in the BIOS and you cannot disable.
Use your motherboard utility to monitor CPU temp or use CoreTemp while running Fritzmark http://www.thecoolest.zerobrains.com/CoreTemp/CoreTemp.zip
Make sure you disable EIST and C1E in the BIOS (check my previous post)
You MUST monitor the CPU temp while running stress test to make sure it is OK
Use your motherboard utility to monitor CPU temp or use CoreTemp while running Fritzmark http://www.thecoolest.zerobrains.com/CoreTemp/CoreTemp.zip
Make sure you disable EIST and C1E in the BIOS (check my previous post)
You MUST monitor the CPU temp while running stress test to make sure it is OK
quote
[This is not my point.
When Enhanced Halt State (C1E) is Enabled, which is the default in my BIOS, then Identical effects as Bouddha occurs.
Mr. Bouddha, please do the following:
1. Get into the BIOS
2. Disable Enhanced Halt State (C1E)
3. Disable Enhanced Intel SpeedStep (EIST)
4. Disable Thermal Monitor 2 (TM2) (but I recommend you enable it later)
5. Save, reboot, run CPUZ, run CoreTemp or your mobo CPU temp tool, then run Fritz mark
6. Monitor you CPU temp and speed
7. You can enable TM2 and retest
Hope things workout well for you]
Isn't the "CPU TM function " the "Thermal Monitor 2" you mentionned that should not be leaved disabled for ever but just for test ?
regards
[This is not my point.
When Enhanced Halt State (C1E) is Enabled, which is the default in my BIOS, then Identical effects as Bouddha occurs.
Mr. Bouddha, please do the following:
1. Get into the BIOS
2. Disable Enhanced Halt State (C1E)
3. Disable Enhanced Intel SpeedStep (EIST)
4. Disable Thermal Monitor 2 (TM2) (but I recommend you enable it later)
5. Save, reboot, run CPUZ, run CoreTemp or your mobo CPU temp tool, then run Fritz mark
6. Monitor you CPU temp and speed
7. You can enable TM2 and retest
Hope things workout well for you]
Isn't the "CPU TM function " the "Thermal Monitor 2" you mentionned that should not be leaved disabled for ever but just for test ?
regards
Hi,
It is a known bug with CPUz with the multiplier jumping back and forth, there are other programs that work fine such as NuclearMC http://nuc-rus.narod.ru/eng.htm ... Running my Quad at 4.1Ghz it is showing 23.05 fritzmark and 11063 NPS I have taken a pic but you cant upload em here :(
Regards
Tony
It is a known bug with CPUz with the multiplier jumping back and forth, there are other programs that work fine such as NuclearMC http://nuc-rus.narod.ru/eng.htm ... Running my Quad at 4.1Ghz it is showing 23.05 fritzmark and 11063 NPS I have taken a pic but you cant upload em here :(
Regards
Tony
And can I do something to avoid that ?
regards
regards
Hi Lukas,
Thanks for the link :) http://img264.imageshack.us/img264/3075/runfb7.jpg
Thanks for the link :) http://img264.imageshack.us/img264/3075/runfb7.jpg
Hi,
Above I just got the comp running at 4.3Ghz and although it runs some benchies it wont complete the Fritz benchmark at that speed, I dont want to sling any more volts at it without having a replacement in case I kill the thing :) ... The memory is running at around 516Mhz
Regards
Above I just got the comp running at 4.3Ghz and although it runs some benchies it wont complete the Fritz benchmark at that speed, I dont want to sling any more volts at it without having a replacement in case I kill the thing :) ... The memory is running at around 516Mhz
Regards
That looks really great. What kind of cooling do you use and what is your VCore ?
Hi,
For 4.3Ghz had to use 1.6V normally I just go 1.5ish .... AMD Performance figures have turned out to be ... now what is the word I am looking for? BULLSH!T http://blogs.zdnet.com/Ou/?p=567
For 4.3Ghz had to use 1.6V normally I just go 1.5ish .... AMD Performance figures have turned out to be ... now what is the word I am looking for? BULLSH!T http://blogs.zdnet.com/Ou/?p=567
This has been well known in industry circles for many months now. AMD is having problems with excessive leakage with their 65 nm process making it impossible for them to get high clock speeds (and forgetabout ocing the chips). It will probably take them a couple more steppings to get this under control. They can't afford to wait though, so the initial chips will be rather shitty.
Alan
Alan
![[is]](/mwf/v12/flags/is.png "Iceland")
> I have taken a pic but you cant upload em here :-(
It's a good idea to have a look at the Forum FAQ board every now and then :). It contains some useful information for forum users:
How do I attach images and other files to posts?
Then of course there is the Help button which is another good source of information.
Hi Dadi,
You are of course correct :) I was being lazy ... Thx for the link it should prove useful
Regards
Tony
You are of course correct :) I was being lazy ... Thx for the link it should prove useful
Regards
Tony
If your computer case is not sealed with "Don't open, unless you void your guarantee" sticker, then it is definitely reasonable to open the case and check, what is the dust situation there. The cooler is a great dust magnet, you will wonder if you have a look! A dust cover reduces drastically the heat dissipation.
For cleaning some dust it is usually not necessary to remove any cables inside the box. Cables outside the box fit only the correct place (except maybe keyboard and mouse, that can be occasionally switched, but this does no harm to your computer and you can just switch those back if they are not in correct place).
A narrow brush with long soft hair is the best tool to get rid of dust from cooler.
Compressed air that is sold in small ballons is useful too.
Temperature of your CPU should be normally less than 65º C under heavy load (the lower the better). Temperatures above 70º C might damage the CPU.
For cleaning some dust it is usually not necessary to remove any cables inside the box. Cables outside the box fit only the correct place (except maybe keyboard and mouse, that can be occasionally switched, but this does no harm to your computer and you can just switch those back if they are not in correct place).
A narrow brush with long soft hair is the best tool to get rid of dust from cooler.
Compressed air that is sold in small ballons is useful too.
Temperature of your CPU should be normally less than 65º C under heavy load (the lower the better). Temperatures above 70º C might damage the CPU.
Sure, but USB devices if you switch from a place to an other I have already faced some "small" issues.
As per dust remaining, I will open the computer and remove, as I cannot accept this frequency change and have to do something to avoid it.
best regards
As per dust remaining, I will open the computer and remove, as I cannot accept this frequency change and have to do something to avoid it.
best regards
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill