Jump to content
HWBOT Community Forums

[Overclocking Support Thread] MSI X99 OC & GAMING Motherboards


mr. robot

Recommended Posts

  • Replies 439
  • Created
  • Last Reply

Top Posters In This Topic

First try to keep your settings as it was and increase SA/IO up to 1.35v.

Also enable BIOS option "Enhanced Memory Controler" and see what happens.

As vdelvec mentioned, try to flash latest MP BIOS available (120) or latest BETA BIOS (131).

 

 

Hello,

i'm running the latest Bios (120). I played with the voltages up to 1.15V without success.

Runing 2666Mhz-15-15-15-1t works stable. Tighter timmings are not working.

I have noticed that the DDR VPP voltages are not equal on auto settings.

DDR VPP AB is 2.224V and DDR VPP CD is 1.936V.

I think they should be equal.

Is it normal?

What voltage values can i set to fix it?

Link to comment
Share on other sites

Hello,

i'm running the latest Bios (120). I played with the voltages up to 1.15V without success.

Runing 2666Mhz-15-15-15-1t works stable. Tighter timmings are not working.

I have noticed that the DDR VPP voltages are not equal on auto settings.

DDR VPP AB is 2.224V and DDR VPP CD is 1.936V.

I think they should be equal.

Is it normal?

What voltage values can i set to fix it?

 

Dont mess with DDR VPP, they are perfectly fine.

However I suspect 1.15v SA/IO might not be enough as I was saying previously.

Bump them up to 1.35v (no risk at all, it's within Intel specs) and see what happens. ;)

Edited by pepinorang
Link to comment
Share on other sites

Dont mess with DDR VPP, they are perfectly fine.

However I suspect 1.15v SA/IO might not be enough as I was saying previously.

Bump them up to 1.35v (no risk at all, it's within Intel specs) and see what happens. ;)

 

I only changed the MB. My 5960X never needed such high voltages on IO and Sa.

IO was on auto = ~ 1,05V and SA 0,9157V.

Do you think that this completely changed after the MB replacement?

 

Micron an Samsung Specs say that DDR VPP set value is 2.5V. I'm not a DDR specialist but

1,9V -> 2,5V is not a little difference.

Link to comment
Share on other sites

I only changed the MB. My 5960X never needed such high voltages on IO and Sa.

IO was on auto = ~ 1,05V and SA 0,9157V.

Do you think that this completely changed after the MB replacement?

 

Micron an Samsung Specs say that DDR VPP set value is 2.5V. I'm not a DDR specialist but

1,9V -> 2,5V is not a little difference.

 

SA/IO needed can vary from board to board regardless of CPU, just try it and see if there is any improvement, if not we can move on to something else! ;)

 

The difference between DDR VPP AB/CD is due to trace layout.

As far as I know no one ever needed to touch those, even for 3000 C10/3200 C12 or DDR4-4420 validation, thus I don't see why you would.

But if you wish you could try, we learn things everyday they say. :)

Link to comment
Share on other sites

Been done since A21 on XPOWER.

However I just posted BETA BIOS A31 in first page, check it out!

 

Just tried A31 for a brief moment. The interface in BIOS looks quite cool but there are some issues with it.

 

- My Soundblaster ZxR is not recognized at all. Working fine on A20 BIOS. The card does get power though as I can hear the relay click.

 

- System seems to hang at A9 postcode when devices are connected to the Asmedia USB3.1 ports. It immediately resumes booting when unplugging these devices, no reset button or anything needed.

 

- System seems to hang on windows boot with some devices connected (Vendor ID 1DD2) to the Intel USB3.0 ports. Only continuing to boot when pressing the reset button.

 

- With the exact same OC settings A31 lacks a good bit of memory bandwidth in synthetic benchmarks. There also seems to be less real-world performance in very memory intensive workloads.

 

- Motherboard temperature is reported as 96°C in BIOS which hopefully isn't true.

 

- Artwork in the BIOS seem to come from one of the Gaming series boards (red instead of yellow)

 

 

Since some of my USB devices have been a mess in general with this board (regardless of BIOS version), is there some way for me (or maybe the BIOS engineers) to disable some vendor ID devices on boot?

Link to comment
Share on other sites

Just tried A31 for a brief moment. The interface in BIOS looks quite cool but there are some issues with it.

 

- My Soundblaster ZxR is not recognized at all. Working fine on A20 BIOS. The card does get power though as I can hear the relay click.

 

- System seems to hang at A9 postcode when devices are connected to the Asmedia USB3.1 ports. It immediately resumes booting when unplugging these devices, no reset button or anything needed.

 

- System seems to hang on windows boot with some devices connected (Vendor ID 1DD2) to the Intel USB3.0 ports. Only continuing to boot when pressing the reset button.

 

- With the exact same OC settings A31 lacks a good bit of memory bandwidth in synthetic benchmarks. There also seems to be less real-world performance in very memory intensive workloads.

 

- Motherboard temperature is reported as 96°C in BIOS which hopefully isn't true.

 

- Artwork in the BIOS seem to come from one of the Gaming series boards (red instead of yellow)

 

 

Since some of my USB devices have been a mess in general with this board (regardless of BIOS version), is there some way for me (or maybe the BIOS engineers) to disable some vendor ID devices on boot?

 

For this kind of issues, please directly report to https://register.msi.com/

If you never did it, they are probably not aware and thus cannot fix! ;)

Link to comment
Share on other sites

Reported to MSI tech support now, hopefully this can be fixed for all boards as the issue isn't limited to this board alone.

 

Thanks man!

 

If the devices are not listed on the QVL list, it may take some time as they have to request it from third party vendors or buy it (in case it's from competitors) before investigating.

Link to comment
Share on other sites

X99A GODLIKE GAMING officially became 1st motherboard worldwide to support DDR4-3200 128Gb kit!

X99A XPOWER AC, X99A GAMING 9 ACK and X99A GODLIKE CARBON will follow soon. :)

 

http://www.guru3d.com/news-story/g-skill-offers-ddr4-3200mhz-128gb-(16gbx8)-memory-kit.html

 

cf15d62d54f834ed5dbe16e49d8e066b.jpg

 

That's good with tight timings.

But will 3200 run with cache high speed?

I may have found the problem with OC ring running with 3200 ram.

Link to comment
Share on other sites

Those voltages exists on every X99 with "special socket" design but are named differently from vendor to vendor. ;)

 

Ring OC1/2 help with Ring OC, while DDR OC1/2/3 help with memory OC.

For Ring OC, you only need to take care of vRing and Ring OC2 voltage.

You don't need to change Ring OC1, AUTO will work fine as this voltage don't help past 0.95v.

 

To improve stability at or above DDR4-3200 and/or for memory high frequency validations, just enable the "Enhanced Memory Controler" option in the BIOS, it will auto-adjust DDR OC1/2/3 for you (more volts don't seem to help).

 

Below is the BIOS auto-rule for Ring OC that I had posted 3-4 pages ago in this thread, it might help you.

FYI, up to 1.45v 24/7 has been fine for me since Haswell-E launch until now, you would be surprised to know how high some other boards auto-set these "special socket" voltages! ;)

attachment.php?attachmentid=3530&d=1448349323

 

 

 

What ram, what settings? Too little info to help you here. :)

 

3200 is working.

Every setting is Auto:

4.5 CPU

4.5 Ring

3200 Memory

Enhanced mem controller

 

Not sure why I was having trouble, but recently did this:

 

turned pc off for many hours, unplugged from wall socket.

reset cmos

checked power cables for cpu (maybe one of those was loose)

removed a video card

left most settings at auto in other areas of uefi

 

 

*Testing stability now, although I'm sure 4.5 @ 1.250 is not enough, but will be ok at 4.4

4.5 uncore increases temp about 5c

 

** blokes at overclock.net saying 1.45 ring is suicidal.

Edited by GunnzAkimbo
Link to comment
Share on other sites

After clear CMOS it is normal that board boot once and then reboot before loading OS, but that does not seems like what you describe.

If you tried different BIOS it's probably not a BIOS issue, so better ask technical support for some help on this, you can contact them here: https://register.msi.com/home/login :)

 

 

Sorry to get back to you late about this but tech support told me that this cold boot double post shut down and restart issue is normal. I don't believe it though. Think its a bios bug but they will not acknowledge that it is a bug. Since then I have tried a different set of ram and no change and updated to the latest 1.2 bios release. I hope they fix this soon.

 

Thanks

Edited by davidm71_2
spelling error
Link to comment
Share on other sites

Sorry to get back to you late about this but tech support told me that this cold boot double post shut down and restart issue is normal. I don't believe it though. Think its a bios bug but they will not acknowledge that it is a bug. Since then I have tried a different set of ram and no change and updated to the latest 1.2 bios release. I hope they fix this soon.

 

Thanks

 

Yer it's in 1.2 for sure, it happens on my godlike, maybe 133 bios sorts it out.

Link to comment
Share on other sites

Hello,

i'm running the latest Bios (120). I played with the voltages up to 1.15V without success.

Runing 2666Mhz-15-15-15-1t works stable. Tighter timmings are not working.

I have noticed that the DDR VPP voltages are not equal on auto settings.

DDR VPP AB is 2.224V and DDR VPP CD is 1.936V.

I think they should be equal.

Is it normal?

What voltage values can i set to fix it?

 

Helle pepin,i hope you are fine!

In the meanwhile i solved my issue. The readings for ddr4 vpp Voltages were wrong.

Aida doesn't show them right.

 

My Settings for the phase control were wrong. I changed all CPU/RAM from optimized to auto and

tighter Timings are now possible.

 

Thats great and what i expect from a premium board.

Link to comment
Share on other sites

X99A GODLIKE GAMING officially became 1st motherboard worldwide to support DDR4-3200 128Gb kit!

X99A XPOWER AC, X99A GAMING 9 ACK and X99A GODLIKE CARBON will follow soon. :)

 

http://www.guru3d.com/news-story/g-skill-offers-ddr4-3200mhz-128gb-(16gbx8)-memory-kit.html

 

cf15d62d54f834ed5dbe16e49d8e066b.jpg

 

Hi pepinorang..

Testing 64 GB kit because i only have 1 :)

with Patriot Viper 2800C16 and MSI X99S Gaming 7 seems very nice.. straight 3200C16 with memory try it! option in bios can boot easily...

Max OC on air with safety voltage :

3400C16 @1.35V spi32m and XTU passed

3400C14 @1.5V spi32m and XTU passed

 

3400c16spi.jpg

 

3400c14spiok.jpg

Link to comment
Share on other sites

3200 is working.

Every setting is Auto:

4.5 CPU

4.5 Ring

3200 Memory

Enhanced mem controller

 

Not sure why I was having trouble, but recently did this:

 

turned pc off for many hours, unplugged from wall socket.

reset cmos

checked power cables for cpu (maybe one of those was loose)

removed a video card

left most settings at auto in other areas of uefi

 

 

*Testing stability now, although I'm sure 4.5 @ 1.250 is not enough, but will be ok at 4.4

4.5 uncore increases temp about 5c

 

** blokes at overclock.net saying 1.45 ring is suicidal.

 

Glad you got it sorted out!

1.45v Ring for extended periods is indeed suicidal, for bench it is fine as long as CPU is cooled adequately.

For 24/7, I would recommend to not go over 4GHz AUTO or 1.25v Ring / 1.4v Ring OC2.

To clarify, very good chips will be able to run more than 4GHz Ring 24/7 but for most of them anything above 4GHz will be too hot and/or require too much voltages.

Edited by pepinorang
Link to comment
Share on other sites

Sorry to get back to you late about this but tech support told me that this cold boot double post shut down and restart issue is normal. I don't believe it though. Think its a bios bug but they will not acknowledge that it is a bug. Since then I have tried a different set of ram and no change and updated to the latest 1.2 bios release. I hope they fix this soon.

 

Thanks

 

Is there any BIOS version where this issue does not exist?

Link to comment
Share on other sites

Is there any BIOS version where this issue does not exist?

 

I dunno Pepi, the "double boot phenominon" happens to present itself on Gigabyte and EVFA X99 boards as well. I think its something built in.

 

Also, when I'm running my 5960x at 4.4Ghz CPU and 4.4Ghz Uncore, the board sets Ring to 1.45v. I thought you said that was an OK voltage? Unless I'm misunderstanding something here...

Link to comment
Share on other sites

I dunno Pepi, the "double boot phenominon" happens to present itself on Gigabyte and EVFA X99 boards as well. I think its something built in.

 

Also, when I'm running my 5960x at 4.4Ghz CPU and 4.4Ghz Uncore, the board sets Ring to 1.45v. I thought you said that was an OK voltage? Unless I'm misunderstanding something here...

 

I will check what is this double boot.

OK, now I can confirm that you DO have a super bad short-term memory :P

I never said 1.45v Ring is OK for 24/7, please re-read my posts! ;)

http://forum.hwbot.org/showpost.php?p=429333&postcount=197

http://forum.hwbot.org/showpost.php?p=419168&postcount=109

Link to comment
Share on other sites

I will check what is this double boot.

OK, now I can confirm that you DO have a super bad short-term memory :P

I never said 1.45v Ring is OK for 24/7, please re-read my posts! ;)

http://forum.hwbot.org/showpost.php?p=429333&postcount=197

http://forum.hwbot.org/showpost.php?p=419168&postcount=109

 

Pepi, please forgive me. A few years ago, I was in an accident and I was oxygen deprived for for a bit longer than is considered safe and it affected my short-term memory.

 

I just didn't understand if you thought that the AUTO setting for Ring voltage at 4.4ghz was too high or just fine for 24/7 overclock?

Link to comment
Share on other sites

Settings are now:

 

5960X

4.5Ghz CPU Core 1.25V (manually set)

4.0Ghz Ring 1.25V (should be OK)

3200 memory default of 16-18-18-42

 

I gave HCI memtest a %100 run with 4 instances of 2048MB, passed.

 

 

If i set memory to the Hynix preset of 3200 C16-16-16-39 and set 1.4V, HCI memtest will error out way before %100

Enhanced memory controller is enabled, so not sure if it's the memory sticks maxing out (originally they are 3000mhz kit).

Edited by GunnzAkimbo
Link to comment
Share on other sites

Pepi, please forgive me. A few years ago, I was in an accident and I was oxygen deprived for for a bit longer than is considered safe and it affected my short-term memory.

 

I just didn't understand if you thought that the AUTO setting for Ring voltage at 4.4ghz was too high or just fine for 24/7 overclock?

 

Sorry to hear, I'll keep that in mind but the answer of your question is in my previous posts! ;)

Link to comment
Share on other sites

Settings are now:

 

5960X

4.5Ghz CPU Core 1.25V (manually set)

4.0Ghz Ring 1.25V (should be OK)

3200 memory default of 16-18-18-42

 

I gave HCI memtest a %100 run with 4 instances of 2048MB, passed.

 

 

If i set memory to the Hynix preset of 3200 C16-16-16-39 and set 1.4V, HCI memtest will error out way before %100

Enhanced memory controller is enabled, so not sure if it's the memory sticks maxing out (originally they are 3000mhz kit).

 

None of the Memory Preset are meant to be Memtest stable, they are just a starting base for benching.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...