Core i7 965 EE Unlocked turbo thread

[Ph0b0s]

I wanted to start a thread about a feature of the new core i7 965 extreme edition that has gone a bit under the radar. Well from my searches around the internet anyway. That feature is ability of the 965 to do unlocked dynamic overclocking using the turbo mode.

This has been covered by some websites in brief detail, but I though it would be good to have a central place to document the feature, see how well it really works, once people get to test it and which motherboards support the feature, as not all X58 motherboards do (as far as I can tell).

Now what is the feature I am on about?

Bear in mind this is based just on what I have read and some assumptions, if there are people using this or who have better info, please reply and point out any errors...

All I7 processor can engage a turbo mode, which allows the the processor to be over clocked dynamically by up-ing the CPU ratio by 1 - 2 x 133 Mhz (base clock) within a TDP of 130W. With the 965 which has an unlocked ratio and with a supporting motherboard, the ratio increase and TDP limits can be decided by the user.

This means instead of a max increase of 266 Mhz in turbo mode, the increase could be anything, 400mhz, 533mhz, etc. And it would be applied only when needed.

This gives the tantalizing prospect (well to me anyway) of user customisable dynamic overclocking of the processor. This also gives the extra benefit of not having to turn off the power saving features of the processor (which I believe you have to do when putting in place a static over clock).

The page below is the only website I have seen that has talked about and used the feature in any great detail. There is also a picture of the area in, a supporting bios, of the settings for the multiplier for the amount of core being used and the TDP limit. They were able to get the processor to clock to 3.7Ghz dynamically.
http://www.tomshardware.com/re...i7-Nehalem,2057-9.html

So far the only two motherboards that seem to support this set-up are:

MSI Eclipse
Intel Smackover

Support is defined as being able to modify the turbo ratio limits for the amount of cores in use and the TDP. If anyone knows of any other motherboards that offer support, please let me know and I will add them to the list.

Also anyone with a 965 processor and using this feature please let us know of your experiences with it.

That's it. Comments please. BTW I'm not desperate to get into a discussion about dynamic vs static overclocking as obviously you will probably get higher overclocks (if that is what floats your boat) from static overclocking. I just really like the idea of dynamic overclocking as a new feature and wanted to start a discussion about it.....

 

[Ph0b0s]

Some better and more well written details on the turbo and unlocked turbo:

http://www.heise-online.co.uk/...now-available--/111986

"Information about the Turbo Boost feature can be found in the Specification Update; all three Core i7 configurations are reportedly factory set to increase their clock speeds by a maximum of 133 MHz (once the feature is enabled in the BIOS setup) if one, two or three of their physical cores aren't operating at maximum load and while they have neither reached their Thermal Design Power (TDP) of 130 Watts nor their maximum operating temperature; if only one core is under load, they can in the same conditions increase the speed by doubling the amount, that is 266 MHz, and increase the core voltage."

"These Turbo Boost levels can be chosen freely in the Core i7 965 Extreme Edition, and the TDP limit can also be adjusted ? if the mainboard offers a BIOS setup capable of it; in "normal" processors, the TDP limit is set at a fixed value, which some observers interpret as a mechanism for limiting overclocking. An integrated power control unit (PCU) makes sure that Core i7 processors operate within their TDP limit, maximum operating temperatures and maximum thermal design current."


From http://techreport.com/articles.x/15818/2

"One trick that this microcontroller enables is the oh-so creatively named "Turbo mode" built into the Core i7. This feature pushes the active cores beyond their baseline clock frequencies when the CPU isn't at full utilization. Turbo mode operates according to some simple rules. In the event that a single-threaded application is occupying one core while the rest are idle, Turbo mode will raise clock speeds by as much as two full "ticks" beyond the baseline. For instance, for our Core i7-965 Extreme processor, Turbo mode could raise the multiplier from 24 to 26, or the core clocks from 3.2 GHz to 3.46 GHz, since the base clock in Core i7 systems runs at 133 MHz. With two or more threads active, Turbo mode will only raise clock speeds by one tick. All of this happens automatically using the same basic P-state mechanism as SpeedStep."

"The additional clock frequency headroom comes from the fact that a less-than-fully-occupied Core i7 may not run up against the limits imposed by its thermal design power, or TDP?the chip's specified power envelope. We've seen a processor running eight instances of Prime95 stay at "one tick up" for a sustained period of time with good cooling. Then again, Intel has set CPU core voltages individually at the factory for some time now, and it's quite possible that some chips may not be able to sustain Turbo acceleration within their specified power envelopes for any length of time. As I understand it, that may simply be the luck of the draw, with only the baseline clock speed guaranteed."

"Interestingly enough, because the Core i7-965 Extreme Edition doesn't have a locked upper multiplier, the CPU can be overclocked by tweaking the Turbo mode settings in the BIOS. Intel's DX58S0 "Smackover" (uh huh) motherboard exposes control over the maximum clock multipliers for one, two, three, and four occupied cores, as well as the ability to adjust the TDP limit in watts and the current limit in amps. You'll probably want a good aftermarket cooler if you plan to play with these settings. If that's too fancy for your tastes, one may also choose to disable Turbo mode and overclock via the usual ways, as well?either by raising the multiplier on an Extreme Edition or by cranking up the base clock on any Core i7. "

 

[Idontcare]

My thoughts on the dynamic vs. static for the 965 is that regardless whether you tease your rig into being 4GHz stable by static clocking methods or by tweaking it with dynamic methods your max clockspeed is still going to be limited by the weakest (clockspeed stability POV) of the four cores.

If dynamic OC allowed you to specify which core was to be the dominate core then you could tease more overclock out of your rig by letting the weaker core remain at a lower base clockspeed and set the higher performing core to the higher allowed clockspeed with dynamic method.

But as it stands now if your rig can support 4GHz with one thread fully loaded while using the turbo mode then it can also support it with the base clock set to 4GHz for all cores provided you have a method of removing the heat. Thread migration makes sure all four cores eventually get tested at 4GHz (in this example) with turbo or without as the thread hops around.

So if dynamic overclock is stable then so too will be static, and for one thread the power consumption will be the same too because the PCU is shutting down the idle cores anyways regardless whether they are clocked to 3.2GHz or 4GHz.

The power consumption difference comes when the chip is loaded more than one thread, if your are static overclocked then all the threads will run at the 4GHz (this example) and consume more power (thus requires better cooling) versus the dynamic overclock situation where the cores run 3.2GHz when more than one is loaded in this example (and thus lower power, lower performance, but easier to handle the thermals).

 

[BTRY B 529th FA BN]

Originally posted by: Idontcare
My thoughts on the dynamic vs. static for the 965 is that regardless whether you tease your rig into being 4GHz stable by static clocking methods or by tweaking it with dynamic methods your max clockspeed is still going to be limited by the weakest (clockspeed stability POV) of the four cores.

If dynamic OC allowed you to specify which core was to be the dominate core then you could tease more overclock out of your rig by letting the weaker core remain at a lower base clockspeed and set the higher performing core to the higher allowed clockspeed with dynamic method.

But as it stands now if your rig can support 4GHz with one thread fully loaded while using the turbo mode then it can also support it with the base clock set to 4GHz for all cores provided you have a method of removing the heat. Thread migration makes sure all four cores eventually get tested at 4GHz (in this example) with turbo or without as the thread hops around.

So if dynamic overclock is stable then so too will be static, and for one thread the power consumption will be the same too because the PCU is shutting down the idle cores anyways regardless whether they are clocked to 3.2GHz or 4GHz.

The power consumption difference comes when the chip is loaded more than one thread, if your are static overclocked then all the threads will run at the 4GHz (this example) and consume more power (thus requires better cooling) versus the dynamic overclock situation where the cores run 3.2GHz when more than one is loaded in this example (and thus lower power, lower performance, but easier to handle the thermals).

Man, seriously, who do you work for. j/k

 

[Idontcare]

Originally posted by: BTRY B 529th FA BN
Man, seriously, who do you work for. j/k

Self-employed. I'm afraid to ask, but why do you ask?

 

[BTRY B 529th FA BN]

Originally posted by: Idontcare

Originally posted by: BTRY B 529th FA BN
Man, seriously, who do you work for. j/k

Self-employed. I'm afraid to ask, but why do you ask?

Just your extensive knowledge prompts me to ask - hah

 

[Idontcare]

Originally posted by: BTRY B 529th FA BN

Originally posted by: Idontcare

Originally posted by: BTRY B 529th FA BN
Man, seriously, who do you work for. j/k

Self-employed. I'm afraid to ask, but why do you ask?

Just your extensive knowledge prompts me to ask - hah

Ah, well that is much kinder than saying "cuz you post here all the freaken time and I'm curious what employer lets you do that instead of working ".

What knowledge of I have in this field comes from working at Texas Instruments in the CMOS development R&D group for a number of years. Calling it extensive though is being overly generous

I happen to have a lot of equally ill-formed opinions too, they just get all gussied up with these fancy engineering words so most folks tend to assume I must know something if I can talk the talk so much.

 

[Ph0b0s]

Thanks for the responses. As I said at the beginning, I did not want to get into a discussion about static vs dynamic over-clocking. But it looks like that is what has happened anyway....

Originally posted by: Idontcare

My thoughts on the dynamic vs. static for the 965 is that regardless whether you tease your rig into being 4GHz stable by static clocking methods or by tweaking it with dynamic methods your max clockspeed is still going to be limited by the weakest (clockspeed stability POV) of the four cores.

If dynamic OC allowed you to specify which core was to be the dominate core then you could tease more overclock out of your rig by letting the weaker core remain at a lower base clockspeed and set the higher performing core to the higher allowed clockspeed with dynamic method.

But as it stands now if your rig can support 4GHz with one thread fully loaded while using the turbo mode then it can also support it with the base clock set to 4GHz for all cores provided you have a method of removing the heat. Thread migration makes sure all four cores eventually get tested at 4GHz (in this example) with turbo or without as the thread hops around.

I think you are saying that static is better than the dynamic mode because in the static case all cores are overclocked and in the turbo mode case only one core is overclocked so threads on the un-overclocked cores would suffer by comparision? I knew my initial post was not written well enough...

It is only is 920 and 940 that will go into turbo mode if only one core is in use. As far as I understand it, with unlocked turbo on the 965 all cores can be dynamically over clocked by the same amount by setting a higher multiplyer for if all cores are in use. Giving the same result as changing the base clock, but dependant on load etc i.e dynamically.
From http://www.techradar.com/news/...ock-the-core-i7-480411

"Where things start to get tricky is the way Turbo Mode gives you access to individual multiplier settings for each core on the Extreme chip. You can set the maximum multiplier setting for each core depending on load, allowing you to tune performance for one, two, three and four core performance. What's more, the Extreme model also gives you access to the chip's thermal and electrical current limiters."

Also see intel bios screenshot here http://www.tomshardware.com/ga...5333-0-2-3-1-png-.html

Originally posted by: Idontcare

So if dynamic overclock is stable then so too will be static, and for one thread the power consumption will be the same too because the PCU is shutting down the idle cores anyways regardless whether they are clocked to 3.2GHz or 4GHz.

I also don't think a CPU that being statically overclocked will have the same idle power consumption as a dynamically overclocked CPU as when doing static overclocks the powersaving fuctions are normally turned off.

See the idle power consumption difference of an overclocked 920 vs stock 920 here. http://www.bit-tech.net/hardwa...g-intel-core-i7-920/14

Again I am not an authority on this and am just going by what I have read, but I would assume that in the unlocked turbo mode dynamic overcloking enabled senario the CPU would keep it's stock idle power consumption since at idle turbo mode would disengage and intel's power saving functions would kick in. Another reason I am interested in this feature. Nearly as good as static overclocking for speed and also sotck idle power consumption all looked after automatically.

But again this is just theory at the moment, so please any one with a 965 using this feature please lets us know whether this is all it is cracked up to be.

 

[Idontcare]

Originally posted by: Ph0b0s
As I said at the beginning, I did not want to get into a discussion about static vs dynamic over-clocking. But it looks like that is what has happened anyway....

Doh! Completely missed that. Sorry.

Originally posted by: Ph0b0s
"Where things start to get tricky is the way Turbo Mode gives you access to individual multiplier settings for each core on the Extreme chip. You can set the maximum multiplier setting for each core depending on load, allowing you to tune performance for one, two, three and four core performance. What's more, the Extreme model also gives you access to the chip's thermal and electrical current limiters."

I had not realized the turbo mode could be customized on a per-core basis for the 965. That degree of freedom surely does add an advantage as it does enable the user to maximize the clockspeed for each core individually so the weakest core doesn't limit the max clockspeed of the strongest core.

The next limiter in this approach of using dynamic overclock for maximizing individual core clockspeeds is that they all share a common Vcc. It would be much more helpful to be able to set Vcc by core too to better maximize the clockspeed of each core.

Originally posted by: Ph0b0s
I also don't think a CPU that being statically overclocked will have the same idle power consumption as a dynamically overclocked CPU as when doing static overclocks the powersaving fuctions are normally turned off.

As I understand it that's not a requirement of the overclock though, that just happens because the user elected to disable the power savings.

Not too unlike how people disable EIST and C1E to overclock their Conroe and Penryn rigs. Not necessary in my experience, and function just fine on my QX6700 when I statically overclock by changing multipliers.

Now what does monkey with power-savings on the Core2 series of processors is when you manually set the Vcc in the BIOS, the BIOS then refuses to allow the chip to activate a lower VID when it goes to idle. That may be what is happening with the power consumption results you linked to regarding the 920. In this case it will impact the 965 too (setting higher Vcc to enable higher stable clockspeed, resulting in powersavings features being disabled/over-ruled by the BIOS during idle).

 

[soccerballtux]

Mmmmm asynchronous processing...

 

[Martimus]

There are a lot of linked articles and reading, but alas I don't have the time to read through it all right now. I do have a favor to ask in the mean time though: is the Dynamic overclocking you are talking about similar to the ability to overclock each core seperately on a Phenom processor, or are there some major differences between the two?

 

[Idontcare]

Originally posted by: Martimus
There are a lot of linked articles and reading, but alas I don't have the time to read through it all right now. I do have a favor to ask in the mean time though: is the Dynamic overclocking you are talking about similar to the ability to overclock each core seperately on a Phenom processor, or are there some major differences between the two?

Martimus of what little I know of the two CPU's the answer would appear to be "yes", at least at a superficial "they are labeled similarly on a powerpoint slide block diagram" level.

Hopefully someones in the know on either processor could illuminate the subject.

It was news to me that i7 allowed user-configurable asynchronous clockspeeds thru the dynamic clock scheduler. So I am completely behind the curve here.

 

[PlasmaBomb]

The one problem I had noticed in a review was that with the dynamic overclocking on the 920 and 940 is that it didn't really help single threaded apps because the OS bounced the thread around...

 

[Ph0b0s]

Originally posted by: PlasmaBomb
The one problem I had noticed in a review was that with the dynamic overclocking on the 920 and 940 is that it didn't really help single threaded apps because the OS bounced the thread around...

Turbo mode on the 920 and 940 is a completely different animal to how it can be done on the 965 due to the unlocked nature of the proccessor. I myself am not impressed by turbo mode on the 920 and 940 for the reasons you mention, i.e vista etc will always try to keep maximum amount of cores in use. I.e if a core is in use and another thread starts it will be moved to a free core first instead of being put on an already used core. I suppose this makes sense as it means the thread gets a whole core to itself. This is my understanding of how a new OS like vista allocates threads.

Originally posted by: : Martimus
There are a lot of linked articles and reading, but alas I don't have the time to read through it all right now.

Have an article again I can point people to, but won't do so again as mentioned above I do go on a bit and post a lot of articles. I must be spending too much time on wikipedia and have picked up their habit of not stating anything unless I can site / link to the info.

This behaviour of thread allocation works to make the turbo mode useless as it gets disengaged onces lots of cores start being used on the 920 and 940 as also the TDP envelope for them is locked as well.

Originally posted by: : Martimus
I do have a favor to ask in the mean time though: is the Dynamic overclocking you are talking about similar to the ability to overclock each core seperately on a Phenom processor, or are there some major differences between the two?

On the 965 you get the option to change the TDP and to set the following (on a supporting motherboard:

A maximum mulitplier for the case of 1 core in use.
A maximum mulitplier for the case of 2 cores in use.
A maximum mulitplier for the case of 3 cores in use.
A maximum mulitplier for the case of 4 cores in use.

I don't know how this compares with the Phenom. And I don't know if the overclock for each core on the Phenom is dynamic. Notice the multipliers and TDP settings you configure on the 965 are maximums and the proccessor will clock up to them or stay at stock depending on load etc.

Hope this answers you your questions, even if it is a long read. Must work on my putting everything into one paragraph skills.

Originally posted by: : Martimus
-------------------------
Harvey = awsome

I'm a Harvey, so thanks for that.

 

[Martimus]

Thank you both for your answers. It does appear similar to the AMD feature, although it is probably more useful as the Phenoms don't overclock very well.

Edit: After reading through the articles and the entire original post, it looks like the biggest difference between the two is that Turbo mode automatically raises the clock speed of individual cores, where I believe that you have to do it manually on an AMD processor. I think that the Fusion app would automatically overclock the processor, but I believe that was the entire processor and not individual cores like the Turbo feature. I may have this all wrong though.

 

[Idontcare]

Originally posted by: Ph0b0s
On the 965 you get the option to change the TDP and to set the following (on a supporting motherboard:

A maximum mulitplier for the case of 1 core in use.
A maximum mulitplier for the case of 2 cores in use.
A maximum mulitplier for the case of 3 cores in use.
A maximum mulitplier for the case of 4 cores in use.

I don't know how this compares with the Phenom. And I don't know if the overclock for each core on the Phenom is dynamic. Notice the multipliers and TDP settings you configure on the 965 are maximums and the proccessor will clock up to them or stay at stock depending on load etc.

Ph0b0s I hope you can see, and agree, that what you write (quoted above) is distinctly different from, and entirely not the same as, what you wrote even further above in the thread (quoted below):

Originally posted by: Ph0b0s
"Where things start to get tricky is the way Turbo Mode gives you access to individual multiplier settings for each core on the Extreme chip. You can set the maximum multiplier setting for each core depending on load, allowing you to tune performance for one, two, three and four core performance. What's more, the Extreme model also gives you access to the chip's thermal and electrical current limiters."

You are now coming around full-circle and are talking about dynamic clocking in the same reference that I was originally thinking it was when I posted the following:

Originally posted by: Idontcare
But as it stands now if your rig can support 4GHz with one thread fully loaded while using the turbo mode then it can also support it with the base clock set to 4GHz for all cores provided you have a method of removing the heat. Thread migration makes sure all four cores eventually get tested at 4GHz (in this example) with turbo or without as the thread hops around.

So I guess I am a little baffled now (again) how it is that the unlocked dynamic multipliers allows you to extract higher overclocks. I can now only see the argument applying in the corner case where the overclock is thermally limiting the overclock (i.e. processor TDP is exceeding the thermal dissipation capability of the HSF) when you have overclocked one or more cores (either dynamically or statically).

Just to repeat, and add clarity, what I am saying is that "A maximum multiplier for the case of 1 core in use." != "A maximum multiplier for the case of core 1." In one case thread migration means the "1 core" can be and will be any of the cores at any given time, whereas "core 1" is a hardware designated core.

Phenom allows per-core multipliers to be set on a per-core basis. I.e. you can explicitly set what multiplier Core#0 operates at, Core#1, etc.

edit: fixed quote syntax

 

[Zap]

Originally posted by: Ph0b0s
So far the only two motherboards that seem to support this set-up are:

MSI Eclipse
Intel Smackover

Heh, those are the exact two boards I have access to. On the MSI, EIST has to be enabled, so you can't lock the speeds. On the Intel, you can enable turbo and disable EIST (they call it something like "idle performance mode"). This way you essentially lock the CPU at whatever multiplier you choose. I've had it running at 133x30= 4GHz using this method on the Intel board. Not 100% stable, I get restarts. Seems to be pretty stable around 28x or so. This is with a 200-300mV boost.

 

[Ph0b0s]

Originally posted by: Zap

Heh, those are the exact two boards I have access to. On the MSI, EIST has to be enabled, so you can't lock the speeds. On the Intel, you can enable turbo and disable EIST (they call it something like "idle performance mode"). This way you essentially lock the CPU at whatever multiplier you choose. I've had it running at 133x30= 4GHz using this method on the Intel board. Not 100% stable, I get restarts. Seems to be pretty stable around 28x or so. This is with a 200-300mV boost.

Thanks for the response. Intesting to have some real details rather than theories. Just to confirm. On the MSI you are making changes in the 'turbo boost tech config' section and changing the mutltipliers for '1-core ratio limit', '2-core ratio limit' etc?

And you had all of these set to x28 so the processor would try to overclock whether 1, 2, 3 or four cores are in use?

I was suprised to hear that the turbo overclock in some cases is not dynamic. I was hoping the way it would work was that the overclock would only happen when there is load on the proccessor. Otherwise I hoped the proccessor speed would be stock when not much load, or lower if EIST was enabled. Was not so bothered about locking the speed as I would use a staic overclocking and turn of turbo if that was what I wanted.

So please let us know if the this turbo overclock is dynamic or not, thanks

I am hopiong that more motherboards will start to offer this functionality and they just missed what Intel were doing on their board before launch. I'm hope others like the asus etc will get bios updates to enable it in the future.

 

[IntelUser2000]

Thanks for the response. Intesting to have some real details rather than theories. Just to confirm. On the MSI you are making changes in the 'turbo boost tech config' section and changing the mutltipliers for '1-core ratio limit', '2-core ratio limit' etc?

And you had all of these set to x28 so the processor would try to overclock whether 1, 2, 3 or four cores are in use?

I was suprised to hear that the turbo overclock in some cases is not dynamic. I was hoping the way it would work was that the overclock would only happen when there is load on the proccessor. Otherwise I hoped the proccessor speed would be stock when not much load, or lower if EIST was enabled. Was not so bothered about locking the speed as I would use a staic overclocking and turn of turbo if that was what I wanted.

So please let us know if the this turbo overclock is dynamic or not, thanks

Ph0b0s, you answered your question on the Tomshardware link. Zap told us that he disabled EIST but enabled Turbo, which makes it essentially function as static overclocking.

Tomshardware quote: "Consider this SpeedStep Technology in reverse. In fact, on the Intel platform, enabling Turbo mode requires that SpeedStep also be turned on!"

Edit: After reading through the articles and the entire original post, it looks like the biggest difference between the two is that Turbo mode automatically raises the clock speed of individual cores, where I believe that you have to do it manually on an AMD processor. I think that the Fusion app would automatically overclock the processor, but I believe that was the entire processor and not individual cores like the Turbo feature. I may have this all wrong though.

You are correct.

Turbo mode on the 920 and 940 is a completely different animal to how it can be done on the 965 due to the unlocked nature of the proccessor. I myself am not impressed by turbo mode on the 920 and 940 for the reasons you mention, i.e vista etc will always try to keep maximum amount of cores in use. I.e if a core is in use and another thread starts it will be moved to a free core first instead of being put on an already used core. I suppose this makes sense as it means the thread gets a whole core to itself. This is my understanding of how a new OS like vista allocates threads.

Turbo mode will work on ALL Core i7 CPUs regardless of thread allocation because even with all cores pegged having enough thermal headroom will let it clock +133MHz. It's when its running single thread it can clock +266MHz and thread affinity MAY become a problem.

The way to test that is put single thread multiplier higher than 2/3/4 thread multipliers. If it clocks the same in single thread as multi-thread, threads are bouncing around.

But Core i7 shouldn't have that problem: http://www.anandtech.com/cpuch...owdoc.aspx?i=3382&p=12

"I mentioned earlier that the PCU monitors OS performance state requests, so it can actually make intelligent decisions about what power/performance state to go into, despite what the OS is telling it."

Core i7, has a dedicated hardware designed solely for managing power and thermals possibly to overcome the OS interfering as it did with earlier generation CPUs.

 

[Ph0b0s]

Just looking through the manuals for the Asus p6t and Rampage II. You seem to be able to do something simular with them to what can be done on the MSI and Intel boards with the turbo limits. Just not in the bios.

The TurboV software has a place in the advanced section under the 'cpu ratio' section where multipliers can be set, only for the 965 EE though. There are four sets of multipliers denoted (i, ii, iii & iv).

The manual is very brief so I am not sure of how these four setting relate to the turbo limits for the case of one, two, three, or four cores being in use. I can't see anything like on the other boards that allow you to change the TDP or current limits though.

Can anyone with an X58 Asus board and a 965 EE CPU confirm the functionallity or expand on what effect these settings have. I.E are they static overclock settings or are they dynamic if Intel powersaving features are left on in the Bios?

I.e by putting 27 as the multiplier against all the setttings (i, ii, iii & iv) , does the proccessor speed go up to 3.6 ghz (133 x 27) when under-load and then back to 3.2 or lower when idle?

I really quite fancy an Asus board, but only if it can be set to do something like this.....

 

[IntelUser2000]

I.e by putting 27 as the multiplier against all the setttings (i, ii, iii & iv) , does the proccessor speed go up to 3.6 ghz (133 x 27) when under-load and then back to 3.2 or lower when idle?

Come to think of it, WHY would it clock it down when it can shut itself down completely?? That makes the point of dynamic Turbo only better for easier overclocking.

 

[Zap]

Originally posted by: Ph0b0s
So please let us know if the this turbo overclock is dynamic or not, thanks

So far in my testing... On the MSI board it is always dynamic. On the Intel board, I can choose. Both boards let you set turbo multipliers on individual cores.

 

[Ph0b0s]

Thanks for the info. Great to get a confirmation.

"Both boards let you set turbo multipliers on individual cores. "

Just want to clarify this in case of consfusion Is it really setting the muliplier for each individual core?

I know in the config says '1-core ratio limit' etc, but I read this as core loading rather than overclocking individual cores. I.e the multiplier put against the '1-core ratio limit' will be the speed the proccessor works at if only one core is loaded and the others are not or have powered off.

And in the same way the muliplier against '2-core ratio limit' is the max speed the proccessor will clock to if everything is ok (temp, tdp), when two cores are loaded, not mattering whether it is cores 1 & 2, 2 & 4, or another combination of two cores in use. The other two cores not being used or powered down. The two used cores would be clocked to the same speed, I think under load.

Same for the other two settings ,i.e 3 cores in use and 4 / all cores in use.

Anyway you can confirm this by say putting say x27 against 1-core limit, x26 against 2-core, x25 against 3-core and x24 against 4-core. Then as the cores get loaded the proccesor will clock from 3.6 ghz down to 3.2 ghz when all four cores are in use if my assumption is correct.

 

[Nemesis 1]

You guys are all over the place on this. Advice . Who explains this at XS forums better than all others can.

 

[Ph0b0s]

Originally posted by: Nemesis 1
You guys are all over the place on this. Advice . Who explains this at XS forums better than all others can.

Any chance of a link to this amazing thread or threads were this is all explained?