Please report all spam threads, posts and suspicious members. We receive spam notifications and will take immediate action!
Page 4 of 13 FirstFirst ... 23456 ... LastLast
Results 31 to 40 of 129

Thread: ASRock Z87 Extreme 6 UEFI 1.90




  1. #31
    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

    I believe I verified what I think I noticed when using manual/Override voltage settings, in a review of the i7-4770K CPU. If this is wrong, I'll be glad to learn otherwise.

    IF manual/Override voltage settings are used, all CPU power saving options are disabled. Set the CPU core voltage to 1.200V manually, with or without a Voltage Offset, and the CPU core voltage at idle will be 1.200V, and will not go lower than that regardless of C-State settings, Windows Minimum processor state, etc. That is the rule for the Haswell FIVR.

    Adaptive voltage is much more complex. At stock max multiplier (38), if I take about the lowest voltage values for Override/manual Voltage and Offset that allows the PC to boot, and just change to Adaptive Voltage with the same voltage and offset values, the PC won't boot! Even when I increased the Adaptive CPU core voltage value to 1.0V, the PC would not boot. That really surprised me.

    Again with a multiplier of 38, with Adaptive voltage and Offset set to Auto, if I turn off the CPU power saving options by setting Windows Minimum processor state to 100%, at idle or moderate real usage load, the CPU core voltage is 1.200V, +/- 0.001V. So 1.200V seems to be a standard or default-like voltage for at least my Haswell CPU (i5-4670K.)

    Using a total of 1.200V with Override/manual Voltage and Offset, the PC was stable at 4.3GHz in AIDA64 with the CPU and FPU stress tests checked, which for me gives the highest core temperatures. The step to 4.4GHz was not stable with less than 1.26V.

    Back to multiplier of 38, Adaptive voltage and Offset set to Auto, CPU Input voltage of Fixed 1.9V (memory at 2133 in all of the stock/38 multiplier tests), with the newest AIDA64 (used for all tests above), build 2529 Beta. Ran the test with CPU and FPU stress tests checked. Maximum CPU core voltage was 1.293V according to AIDA64.

    So without any OC and Adaptive voltage and offset on Auto, I was just a hair under 1.3V CPU core voltage. Core temps were in the mid 80's C, max CPU package power was 80.65W.

    Given this, as I believe Shiari said, how can we consider a 1.3V CPU core voltage as excessive for Haswell CPUs? Assuming that the ASRock board is not an influence on the CPU core voltage, which is reasonable IMO, then the FIVR is adjusting core voltage at it sees fit. What do we trust?

    According to the Haswell datasheet from Intel, the SVID signal that controls the core voltage has a range of values that translates to 0.500V - 3.040V. I'm not saying that the CPU could survive 3V or 2V with normal cooling, but it would seem to me that a core voltage of 1.3V is not a taboo level. If 1.2V is a nominal value at stock clocks, 1.3V is less than 10% greater, and cannot be overly high, IMO.

  2. #32
    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 parsec View Post
    I believe I verified what I think I noticed when using manual/Override voltage settings, in a review of the i7-4770K CPU. If this is wrong, I'll be glad to learn otherwise.

    IF manual/Override voltage settings are used, all CPU power saving options are disabled. Set the CPU core voltage to 1.200V manually, with or without a Voltage Offset, and the CPU core voltage at idle will be 1.200V, and will not go lower than that regardless of C-State settings, Windows Minimum processor state, etc. That is the rule for the Haswell FIVR.
    That's right, does exactly what it says on the tin: "use this voltage". The CPU should still downclock, but there's very little if any gain because you're feeding a set amount of power into the CPU. This is however the easiest way to find the voltage you need to get stable at a certain clock, and/or determine whether your CPU is a dog or a good clocker.

    Quote Originally Posted by parsec View Post
    Adaptive voltage is much more complex. At stock max multiplier (38), if I take about the lowest voltage values for Override/manual Voltage and Offset that allows the PC to boot, and just change to Adaptive Voltage with the same voltage and offset values, the PC won't boot! Even when I increased the Adaptive CPU core voltage value to 1.0V, the PC would not boot. That really surprised me.
    I'm not sure what voltages you're using here. I think I had some issues booting with some made up numbers as well, though I can't exactly remember. I've not tried lowering it (and raising the offset correspondingly) but I believe my boot issues where with a 1.0v/0.3v ... right now (from memory) I'm at roughly 1.2v/0.1v I think.

    Quote Originally Posted by parsec View Post
    Again with a multiplier of 38, with Adaptive voltage and Offset set to Auto, if I turn off the CPU power saving options by setting Windows Minimum processor state to 100%, at idle or moderate real usage load, the CPU core voltage is 1.200V, +/- 0.001V. So 1.200V seems to be a standard or default-like voltage for at least my Haswell CPU (i5-4670K.)
    I think it's the board deciding that, not the CPU.

    Quote Originally Posted by parsec View Post
    Using a total of 1.200V with Override/manual Voltage and Offset, the PC was stable at 4.3GHz in AIDA64 with the CPU and FPU stress tests checked, which for me gives the highest core temperatures. The step to 4.4GHz was not stable with less than 1.26V.
    I think 1.26V @ 4.4 is reasonable, seems to be a mediocre to average clocker. You might be able to get it to 4.5 with ~1.3V, but higher than that might be difficult. Better cooling might help allowing for a bit more voltage.

    Quote Originally Posted by parsec View Post
    Back to multiplier of 38, Adaptive voltage and Offset set to Auto, CPU Input voltage of Fixed 1.9V (memory at 2133 in all of the stock/38 multiplier tests), with the newest AIDA64 (used for all tests above), build 2529 Beta. Ran the test with CPU and FPU stress tests checked. Maximum CPU core voltage was 1.293V according to AIDA64.

    So without any OC and Adaptive voltage and offset on Auto, I was just a hair under 1.3V CPU core voltage. Core temps were in the mid 80's C, max CPU package power was 80.65W.
    1.3V at 3.8GHz? Ouch. Yeah that'll push the temperatures.

    Quote Originally Posted by parsec View Post
    Given this, as I believe Shiari said, how can we consider a 1.3V CPU core voltage as excessive for Haswell CPUs? Assuming that the ASRock board is not an influence on the CPU core voltage, which is reasonable IMO, then the FIVR is adjusting core voltage at it sees fit. What do we trust?
    1.3V is really quite okay, as long as the temperatures are reasonable. 1.35V effective is about the max I would go myself, as long as the CPU isn't boiling yet. And yeah, the FIVR will up the core voltage if it feels it needs to, though unknown is exactly what is triggering that. My guess would be that it checks what parts of the CPU are active (e.g. AVX, FPU) and ups the voltage accordingly.

    Quote Originally Posted by parsec View Post
    According to the Haswell datasheet from Intel, the SVID signal that controls the core voltage has a range of values that translates to 0.500V - 3.040V. I'm not saying that the CPU could survive 3V or 2V with normal cooling, but it would seem to me that a core voltage of 1.3V is not a taboo level. If 1.2V is a nominal value at stock clocks, 1.3V is less than 10% greater, and cannot be overly high, IMO.
    If you had LN2 cooling or such you could up the voltage a lot more. It is all about the temperatures, not about the voltages. The voltage simply causes temperatures to rise, and the high temperature can destroy the chip or degrade it over time.

    Your chip might still fail to be stable at some clock even with a high voltage, but that's the silicon lottery for ya.

  3. #33
    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

    Shiari, I started to reply to your last post, but then I was side tracked by something I found in a Gigabyte mother board, Haswell OC guide, a thread at OCN. It contradicts what we agree on about using a manual voltage with Haswell, and the affect it has on the CPU core voltage. The following is from that guide, not just a random post:


    "However now that Haswell has a built in VRM there is almost little reason to use offset mode b/c if you enable C3 power state and EIST the CPU multiplier and voltage drop together even if you have set a manual voltage without offset."

    This seems wrong given what we have seen, but I really never tried manual/Override voltage setting without an Offset value. So I decided to give it a try.

    I then found in our board's UEFI, it is impossible to directly enter an Override Voltage Offset of 0.0V. If you do, it changes to Auto. I had set the Override Voltage to 1.20V, multiplier at 38, just to avoid any forced settings when higher multipliers are used, and simply to see what the behavior is at a stock multiplier.

    Back to Override Voltage Offset. When set to Auto, if I press the '+' key, it changed to 0.001V. Pressing the '-' key from 0.001V, sets it back to Auto. So does that mean Auto in this case is 0.0V?

    Back to the CPU multiplier and voltage drop together even if you have set a manual voltage without offset, with the Override voltage and Offset set to 1.2 and Auto respectively. I had C-States enabled, and Minimum processor state set to 0%. The result was NO different than using a non-zero Offset voltage, CPU core frequency dropped to 800MHz, CPU core voltage stayed at 1.20V. Checked the core voltage with three different programs, all showed 1.20V with the core clocks at 800MHz.

    So what do we make of this? Is that guide wrong? Is our UEFI different than Gigabyte boards? What am I missing?

  4. #34
    Join Date
    Dec 2011
    Location
    Pittsburgh
    Posts
    28

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    Quote Originally Posted by parsec View Post
    So what do we make of this? Is that guide wrong? Is our UEFI different than Gigabyte boards? What am I missing?


    I am about to test that right now. I now have a Gigabyte Z87X-D3H.

  5. #35
    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 SLK View Post
    I am about to test that right now. I now have a Gigabyte Z87X-D3H.
    I am extremely interested to know the results! Have you seen that thread at OCN I mentioned?

    So you're trying another board? Or bailed out on the ASR board? Whatever, I have the same thing in mind, I sure don't blame you, just hope you'll post your results.

  6. #36
    Join Date
    Dec 2011
    Location
    Pittsburgh
    Posts
    28

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    Fixed voltage 1.16v at 4.2ghz. All cstates, C1E and EIST on auto and it DOES throttle both voltage and speed. Another good thing is as long as you know how well your uncore does, this board is super easy to overclock. Just set the voltage and it doesn't even overshoot on AVX either unless you choose auto or offset. My temps barely hit over 70c with AVX LINX, P95 and IBT.

    I bailed out on both my extreme 4 and extreme 6. I actually paid much less for the D3H and I am happy as a clam now. ASrock needs to fix the bios asap.


    Yes, I saw the thread over at OCN. Sin convinced me to get the board. :)
    Last edited by SLK; 07-07-2013 at 12:02 PM.

  7. #37
    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 SLK View Post
    Fixed voltage 1.16v at 4.2ghz. All cstates, C1E and EIST on auto and it DOES throttle both voltage and speed. Another good thing is as long as you know how well your uncore does, this board is super easy to overclock. Just set the voltage and it doesn't even overshoot on AVX either unless you choose auto or offset. My temps barely hit over 70c with AVX LINX, P95 and IBT.

    I bailed out on both my extreme 4 and extreme 6. I actually paid much less for the D3H and I am happy as a clam now. ASrock needs to fix the bios asap.


    Yes, I saw the thread over at OCN. Sin convinced me to get the board. :)
    Thank you for this, I'm glad your new board is working well... or should I say, correctly?

    I was afraid that is the situation with this ASRock board/UEFI

    I've stopped trying to OC it at all, all the stress on the other devices with the BSODs and freezing, and all the restarts afterwards. I can't even look at the "Unsafe shutdown count" SMART data values on my SSDs, more salt in my wounds...

    Yes, I see your new board for $10 less than the Z87 Ex 4, after rebate, at the same retailer. I'm past the retailer's return period on my ASR board...

    I'm not a hardcore OC person, and I know this board is not an extreme OC board, but I do expect the basic features and options to function correctly.

    Glad you found Sin's thread, it has taught me several things, and has also depressed me...

    A shame really, some things work so well on this board, but not the CPU and UEFI interaction. When I saw the CPU input voltage labeled as the Vcore in the UEFI, I just knew that was a bad sign. I decided to let it go... hindsight is 20/20.

    Time to explore my options.

  8. #38
    Join Date
    Dec 2011
    Location
    Pittsburgh
    Posts
    28

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    Wait... I just realized something. Read this thread here... The Gigabyte Z87/Haswell Overclocking(OC) Guide

    CPU-Z 1.65 does not read Vcore on my Gigabyte, it shows the "entered" voltage from what Forceman is saying. If I load up 1.64.0 then it shows idle and load volts perfect. I go into HWinfo64 and it shows the Vcore properly as well but the VID stays the same and does not throttle I am betting that CPU-Z is reporting the VID and not the actual Vcore on the Gigabyte board and I bet the same is happening on the ASrock but its not reporting Vcore at all and you are just seeing VID. The reason why I came to this conclusion is because my Idle temps on Fixed voltage on my extreme 6 was practically the same if it were to throttle the voltage. I bet Asrock has the sensors showing incorrectly throughout all the applications.

  9. #39
    Join Date
    Jul 2013
    Location
    chicago
    Posts
    10

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    Quote Originally Posted by SLK View Post
    Wait... I just realized something. Read this thread here... The Gigabyte Z87/Haswell Overclocking(OC) Guide CPU-Z 1.65 does not read Vcore on my Gigabyte, it shows the "entered" voltage from what Forceman is saying. If I load up 1.64.0 then it shows idle and load volts perfect. I go into HWinfo64 and it shows the Vcore properly as well but the VID stays the same and does not throttle I am betting that CPU-Z is reporting the VID and not the actual Vcore on the Gigabyte board and I bet the same is happening on the ASrock but its not reporting Vcore at all and you are just seeing VID. The reason why I came to this conclusion is because my Idle temps on Fixed voltage on my extreme 6 was practically the same if it were to throttle the voltage. I bet Asrock has the sensors showing incorrectly throughout all the applications.
    HWmonitor is reading my cpuvcore on the extreme6 as fluctuating around .95 but my 4670k vid at 1.224, this is with manual 1.225 set in uefi and 0% min in power management. Im running a custom EK wcing loop, my idle temps are a good 6-7C higher now with manual voltage set at 1.225 and 42 ratio compared to stock. Im having a hell of a time tyring to oc this system, just set it up yesterday. earlier today i had it running 4.4 and got through ten or so prime95 totture test rounds before I rebooted to up it again, now I cant even get 4.2 to pass the first round with at least one core failing...its messing with my 780 oc as well, bioshock is crashing constantly and causing gpu crashes that were stable with stock cpu values. I cant remember what I used for the stable 4.4 earlier and its driving me mad...ive had it up to 1.4v and still cant get it remotely stable again...havent gone above 75C at 1.4 so I know its not a temp problem...
    Last edited by dehenry2; 07-07-2013 at 04:12 PM.

  10. #40
    Join Date
    Jul 2013
    Location
    chicago
    Posts
    10

    Default Re: ASRock Z87 Extreme 6 UEFI 1.90

    Well now prime95 is running at the 4.8 turbo auto setting...the only difference is i didnt set my memory at 1600(which its rated for...) it defaults at 1333 for the profile... I dont know wtf is going on here

Page 4 of 13 FirstFirst ... 23456 ... 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
  •