Please report all spam threads, posts and suspicious members. We receive spam notifications and will take immediate action!
Page 6 of 13 FirstFirst ... 45678 ... LastLast
Results 51 to 60 of 129

Thread: ASRock Z87 Extreme 6 UEFI 1.90




  1. #51
    parsec's Avatar
    parsec is offline Senior Member
    Join Date
    Jun 2012
    Location
    Third stone from the sun
    Posts
    4,824

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    Quote Originally Posted by Shiari View Post
    I can't even find the "HWiNFO64 Vcore" value. But that value sounds like the CPU VCORE from HWMonitor, and older CPU-Z core voltage. I think this particular value is meaningless, at least for the Extreme 6. It doesn't seem relate to anything set in UEFI, and it isn't needed to achieve an OC anyway. Just ignore this value, please ... makes life simpler.
    I was simply listing all the potential values for core voltages in HWiNFO64, I'm not saying it is valid, nor do I pay attention to it. It matches the "VRM" value in AIDA64. It is found under the "Nuvoton NCT6776F" heading in HWiNFO64, the latest version.

    Quote Originally Posted by Shiari View Post
    From what I see here (I don't really want to stress things much atm, we're at 26c ambient at the moment, 10:30 in the morning, puff puff) HWiNFO64 VID == CPU-Z 1.65/1.64.2 Core Voltage == HWMonitor VID == XTU Core Voltage == AIDA64 CPU VID, and this is the most relevant value for overclocking. It relates directly to the "CPU VCore voltage Mode", "Vcore Adaptive Voltage" and "Vcore voltage additional offset" values as they're named in the 1.90 UEFI. Yes, very nice there's so much consistency in the naming here!



    I'm not sure what the value is supposed to represent, but it's pretty obvious it's not reported correctly. It might be an incompatibility, it could be something simply not registering anything and AsRock just having provided a value for it, could perhaps be behaviour of the Winbond/Nuvoton NCT6776 chip which reports these values, maybe a default value when nothing is connected to read it. I don't know, and tbh, I don't really care either.



    You can trust the "CPU Core Voltage" values I've mentioned, I'll repeat them here: HWiNFO64 VID == CPU-Z 1.65/1.64.2 Core Voltage == HWMonitor VID == XTU Core Voltage == AIDA64 CPU VID
    IMO, therein lies the potential problem, they are all the same on our board. They all rely on the output of the Nuvoton sensor chip, apparently, even IXTU. Then compare that to the results SLK has with his Gigabyte board, and other results in the Gigabyte Haswell OC thread started by Sin at OCN. Do you believe the Gigabyte readings are wrong?

    I've stopped OCing my Haswell CPU simply because I don't trust the UEFI. That is partially based upon what I've seen in screenshots of other board's UEFIs, the information they provide alone is night and day compared to this board.

    SLK, just tried CPU-Z 1.64.0, thanks for the links. It shows... 0.940V, the same reading as the HWiNFO64 "Vcore" value, and the AIDA64 VRM value! All the other programs show my VIDs as 0.750V at idle. So that yet again is another example of who knows what is correct for a specific board. No offense SLK, just saying...

    Let's talk VID, what is it? It's an eight bit value used by the CPU as a signal to request the voltage the CPU needs in a certain situation, idle, under load, etc. The eight bit VID can be translated into a voltage value, since it represents a voltage value, Intel has a table mapping eight bit VIDs to voltages. One little detail that is often ignored is each CPU will have slightly different VID to voltage mapping, since each CPU is different. The VIDs are adjusted for each CPU during the manufacturing process.

    So what are these programs using to display a "CPU core voltage"? The actual voltage measured in some way? Or are they reading the VID eight bit value, and mapping that to a voltage value? If it's the VID to voltage mapping, is the mapping table generated somehow for every CPU, or is it just the static, base table that Intel provides in the Haswell spec document?

    I noticed that the Gigabyte UEFI displays a base voltage for the CPU, which ought to be a certain VID to voltage translation, based upon the specific "base VID" (my terminology) for a CPU. The lower that value is, the better, and predicts how well your Haswell CPU will OC. Do we have such a reading in the ASR UEFI? I wish...

    My decision to not OC my Haswell CPU is simply based on all the uncertainties that exist at this point, IMO, among other factors. If I need an OC'd system, I'll use my i7-2600K or i5-3570K PCs. I don't need to delid my i7-2600K for it to run at 4.6GHz, or even 5GHz.

  2. #52
    Join Date
    Dec 2011
    Location
    Pittsburgh
    Posts
    28

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    I am sure the bios will mature down the road enabling you to overclock with much more ease. I had a feeling you were going to get the 0.94v in the bios like when I first bought my launch Z87 Extreme 4 it happened in CPU-z as well. There may be a program out there somewhere that will show you the actual vcore values, but Asrock should just fix their bios IMO.

  3. #53
    Join Date
    Jul 2013
    Location
    chicago
    Posts
    10

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    Quote Originally Posted by Shiari View Post
    For gaming there is really no noticeable benefit. Games need the GPU, games need the processor feeding the GPU, and the processor doing calculations (e.g. for AI). None of these are memory intensive tasks. Rendering, movie or image processing or other tasks that require loads of memory are the most likely to see any benefit, and even then a higher processing speed still trumps the faster memory.

    Benchmarks may show improvements from faster memory, but most of those are not real world applications and made to highlight the improvements from faster memory access.

    What is your uncore (cache clock) set at btw? Generally the recommendation seems to be to keep that within .3Ghz of the CPU clock.
    the profile sets it at cpu 48, voltage 1.38; uncore 38, voltage 1.30. I woke up and shut xtu off, 4 hours in, im in an un-airconditioned basement and it wasnt helping me sleep. Im about to play some bf3 and crysis 3, if they dont crash im pretty confident in the clock, will try to lower the voltage. I hope they update the bios to allow proper throttling, idling at 38C with this voltage is 10C more than my 780.

  4. #54
    Shiari is offline Member
    Join Date
    Jun 2013
    Location
    Nottingham
    Posts
    143

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    Quote Originally Posted by dehenry2 View Post
    the profile sets it at cpu 48, voltage 1.38; uncore 38, voltage 1.30. I woke up and shut xtu off, 4 hours in, im in an un-airconditioned basement and it wasnt helping me sleep. Im about to play some bf3 and crysis 3, if they dont crash im pretty confident in the clock, will try to lower the voltage. I hope they update the bios to allow proper throttling, idling at 38C with this voltage is 10C more than my 780.
    Make sure your power profile has 0% set for the advanced settings under Processor power management > Minimum processor state setting. Besides that you may need to use adaptive voltage rather than voltage override (which I believe is set by the OC profiles). You'll need to dial in the adaptive voltages according to what you need for a stable OC. I use 1.09V with a 150.4mV offset for my 4.6 OC which brings it to 1.2590V under load (I guessed at these values ... about 0.1V seemed a reasonable offset, perhaps I can spread it to something like 0.99V/250mV safely but I haven't tried that yet). Currently my idle temps are 29-32c with a 26c ambient temperature, and a reported TDP of only 6W.

  5. #55
    Join Date
    Jul 2013
    Location
    chicago
    Posts
    10

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    4.8 crashed in BF3 after 10 mins, 4.7 is stable with the same 1.38 voltage. Im gonna stick with 4.7 and try to drop the voltage down.

  6. #56
    Join Date
    Jul 2013
    Location
    chicago
    Posts
    10

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    I switched to adaptive voltages, 1.275 cpu with .025 offset and 1.050 cache with .150 offset. I also changed the cache multi to 42, because of a thread i found Official Intel Haswell 4670/4770(K) thread - Guru3D.com Forums prime is working so im gonna try crysis 3, hopefully i can drop the voltage more.



  7. #57
    Join Date
    Jul 2013
    Location
    chicago
    Posts
    10

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    no dice, multiple BSODs in crysis 3, clock interrupt secondary processor...these settings are stable at 4.7 for crysis 3...with the voltage this high I may as well try to get 5.0....anyone seen any threads pushing voltage beyond 1.4?


  8. #58
    Join Date
    Jul 2013
    Location
    chicago
    Posts
    10

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    how do the adjusted voltage and offset values interact? my voltage is dropping now idling at desktop, .873 lowest, but temps are still 7c higher than stock idle...

  9. #59
    Shiari is offline Member
    Join Date
    Jun 2013
    Location
    Nottingham
    Posts
    143

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    I've had a play with the adaptive core voltage and offset, but not much luck. Lowering the "base" value (can technically go as low as 0.8) and increasing the offset by the same amount seems to raise (eh? makes no sense) the core voltage, I've seen it go up to about 1.5V! Though the CPU was mostly idle, so no excessive heat and no harm done. Had to undo it though, so now I'm back at 1.096V/150mv which brings me to ~1.26V at load. I can't seem to get the idle core voltage lower than about 0.86V, which very much appears to be a limitation of the board. Or I'm doing something utterly wrong with these settings, wouldn't know what though. But with a reported TDP at idle of 5-7W and 24-25c core temperatures at 22c ambient I'm happy with these values.

    Also, need to do more stability testing still, but it seems I'm able to nudge my clocks up to 4.7 while keeping everything else the same :)

  10. #60
    parsec's Avatar
    parsec is offline Senior Member
    Join Date
    Jun 2012
    Location
    Third stone from the sun
    Posts
    4,824

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    Quote Originally Posted by Shiari View Post
    I've had a play with the adaptive core voltage and offset, but not much luck. Lowering the "base" value (can technically go as low as 0.8) and increasing the offset by the same amount seems to raise (eh? makes no sense) the core voltage, I've seen it go up to about 1.5V!
    Yes, I had the same experience. With the sum of adaptive core voltage and offset the same, subtracting from core and adding that same amount to offset resulted in a higher overall voltage. So the algorithm seems to be, if the base Adaptive voltage is set "to low", ignore that and use whatever the CPU thinks it needs, and the larger the offset is, the more it can be surpassed.


    Quote Originally Posted by Shiari View Post
    Though the CPU was mostly idle, so no excessive heat and no harm done. Had to undo it though, so now I'm back at 1.096V/150mv which brings me to ~1.26V at load. I can't seem to get the idle core voltage lower than about 0.86V, which very much appears to be a limitation of the board. Or I'm doing something utterly wrong with these settings, wouldn't know what though. But with a reported TDP at idle of 5-7W and 24-25c core temperatures at 22c ambient I'm happy with these values.

    Also, need to do more stability testing still, but it seems I'm able to nudge my clocks up to 4.7 while keeping everything else the same :)
    You no doubt have Minimum processor state set to 0%, and all C-States enabled or Auto, right? Using Adaptive voltages, I can get idle/800MHz core voltage/VID down to ~0.750V and CPU Package power (your TDP I assume) under 2 Watts (under 1W even!), but my core temps are higher, ~30C. I agree the "penalty" of using Override voltages is not much, but I really wonder if the sub 0.2V idle core voltages I've seen are real. I should post pics of my apparent CPU power usage, since it's a see it to believe it kinda thing... if correct.

    I also wonder when the ultra low power states are active, if the core voltage/VID is ignored by the FIVR, and an even lower voltage is provided to the CPU. So while CPU-Z, etc, are using the VID to display the core voltage, that's not the true core voltage in that situation.

Page 6 of 13 FirstFirst ... 45678 ... LastLast

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •