Overclock.net banner
121 - 140 of 631 Posts
Discussion starter · #121 ·
Hello,

I’ve got a weird issue. I’m trying to setup all voltages before overclocking.
I do Optimized settings, save, reboot, then NGU, D2D….

When I get to the VDD2 value, if I put 1.45, the system doesn’t boot.

I don’t understand why.

There is no OC, no XMP.
This is "CPU VDD2" or VDD2 in the DRAM Voltages?
 
Discussion starter · #123 ·
CPU VDD2 in the bios yes !
It may be booting with it not set as a result of your kit requiring more, and the auto default supplying more. Not sure without seeing Hwinfo. Your kit being higher frequency may be needing more than 1.5v.

I am running this for my 8600mhz overclock

CPU VDDQ = at least 1.5v
VccIO = 1.35v
VccClk = 1.35v
CPU VDD2 = 1.57v.
CPU System Agent = 1.45v

VDD2 should always be slightly higher than VDDQ.

So if you have VDDQ at 1.45 then VDD2 should be 1.47.
You can move up 100mv increments until it posts and is stable, or just set to 1.5/1.57.

HWinfo doesnt seem to show CPU VDDQ, but does show CPU VDD2

Image
 
Hey Guys, when I do a cold boot the PC powers up for a second or two then turns off then powers up again and everything is fine.. I thought I fixed it last night but today it is doing it again..

I was wondering if anybody knows why this is doing that.. I checked the pins and they were fine as is the CPU. I checked with a single stick of memory it did the same thing.. is there something in BIOS causing this to startup funny?
 
  • Rep+
Reactions: ArrogantHair
It may be booting with it not set as a result of your kit requiring more, and the auto default supplying more. Not sure without seeing Hwinfo. Your kit being higher frequency may be needing more than 1.5v.

I am running this for my 8600mhz overclock

CPU VDDQ = at least 1.5v
VccIO = 1.35v
VccClk = 1.35v
CPU VDD2 = 1.57v.
CPU System Agent = 1.45v

VDD2 should always be slightly higher than VDDQ.

So if you have VDDQ at 1.45 then VDD2 should be 1.47.
You can move up 100mv increments until it posts and is stable, or just set to 1.5/1.57.

HWinfo doesnt seem to show CPU VDDQ, but does show CPU VDD2

View attachment 2703980
VccDDQ under memory timings is the VDDQ unless im confused.
 
What does the increasing the VccCLK do? I cant find much info about it. mine is running at 1.1v
 
Discussion starter · #127 ·
VccDDQ under memory timings is the VDDQ unless im confused.
There are two.

CPU VDDQ
VDDQ under Dram Voltages.

The following are CPU level settings that give you voltage headroom by increasing v to rails
CPU VDDQ = at least 1.5v
VccIO = 1.35v
VccClk = 1.35v
CPU VDD2 = 1.57v.
CPU System Agent = 1.45v

Under Dram Voltages
VDDQ and VDD2 apply v to the ram sticks. Default auto.

So you increase the 1st to give headroom.
If auto posts, you are good.
If not then you have to bump up the 2nd Dram voltages.
So if Kit 1.45v, you can try 1.5v etc.

Beginner at ram overclocking, and there is no documentation with the Gigabyte boards, so this how I think it works. Open to corrections.

For my Oc from 8200 to 8600 - I set the 1st values as described above.
Then I set the Dram voltages VDDQ to 1.5 and VDD2 to 1.45v - per recommendation of Leonman44. (auto worked for F16c and earlier, but F16 broke something and I had to apply to the Dram)
For my sticks these appear as VDD (swa) and VDDQ (swb) in Hwinfo.

Image
 
Discussion starter · #128 ·
VccCLK - Power for the internal clock PLL - according to Intel datasheet.
VccIO - Power for intern I/O Buffers.

I had to bump VccIO for mine to get stable. Recommended to me - so I have no idea of mechanics.
 
  • Rep+
Reactions: Wribbity
Discussion starter · #129 ·
Hey Guys, when I do a cold boot the PC powers up for a second or two then turns off then powers up again and everything is fine.. I thought I fixed it last night but today it is doing it again..

I was wondering if anybody knows why this is doing that.. I checked the pins and they were fine as is the CPU. I checked with a single stick of memory it did the same thing.. is there something in BIOS causing this to startup funny?
I have the same issue and reported it to Gigabyte. Their support are incompetent and blame the overclock. Still have a ticket in. Hope to get an answer or fix.

I unplug from power supply for a half a minute and then replug in and boot. Happens when my computer goes to sleep. So I now have it set not to sleep.

I believe it is tied to the D2D being above 33. Asus boards have a "Coldboot override" setting that prevents this.
 
I have the same issue and reported it to Gigabyte. Their support are incompetent and blame the overclock. Still have a ticket in. Hope to get an answer or fix.

I unplug from power supply for a half a minute and then replug in and boot. Happens when my computer goes to sleep. So I now have it set not to sleep.

I believe it is tied to the D2D being above 33. Asus boards have a "Coldboot override" setting that prevents this.
Yes the Tech Support I have a ticket in as well. it does shut off when the mobo light is on DRAM and I get a C5 on the led readout and it turns off but turns right back on and runs no issues.. it could be the ram is incompatible but IDK for sure..

I am going to guess they will blame mine on RAM issues.. they tried saying there might be something wrong with the CPU or bent pins. Well I took the CPU very carefully and took pictures there were no bent pins that is for sure..

I did set the NGU and D2D to 32 I am pretty sure it did the same thing so that might not be it.. while I am on that thought I will go change just the D2D setting to 32 and see what happens.. look forward to hearing what Gigabyte Tech support tells you..

Well I set D2D to 32 and rebooted then turned the PC off cleared the power and let it sit for a minute, when I booted up it did the same thing even with D2D at 32? or should I set it lower for a test?
 
Discussion starter · #131 ·
Yes the Tech Support I have a ticket in as well. it does shut off when the mobo light is on DRAM and I get a C5 on the led readout and it turns off but turns right back on and runs no issues.. it could be the ram is incompatible but IDK for sure..

I am going to guess they will blame mine on RAM issues.. they tried saying there might be something wrong with the CPU or bent pins. Well I took the CPU very carefully and took pictures there were no bent pins that is for sure..

I did set the NGU and D2D to 32 I am pretty sure it did the same thing so that might not be it.. while I am on that thought I will go change just the D2D setting to 32 and see what happens.. look forward to hearing what Gigabyte Tech support tells you..

Well I set D2D to 32 and rebooted then turned the PC off cleared the power and let it sit for a minute, when I booted up it did the same thing even with D2D at 32? or should I set it lower for a test?
Nah, if your NGU and D2D are that low, that will indicate it may indeed be ram. The D2D issue is above 33.

That's the thing, I have run Karhu, OCCT and other tests and mine is stable, so I am sure it is the D2D. So when they answer that it is ram OC fault, I am like nooope. When they say lower D2D - I am like - I did not pay $x for an overclocking board not to be able to overclock. lol.

For you, it has to be something small - shouldn't be the cpu, unless - are you using a cpu frame instead of the default bracket? That can cause ram issues if the frame is tightened too much. Has to be hand tight - but only to the point you feel resistance.

If not the above, then it is voltage, timing, compatibility or bios. Your kit is lower frequency, so you would think it would run with no issue. Are you on F16 or earlier? F16 I think they broke something ram related. It is not applying the same timings for XMP/DDR as it did for earlier. I reported this as well. So I manually set mine based off a MemTweakit screenshot I took earlier.

You could try to roll back to F16c or earlier. I have not overclocked ram before, so I am entirely reliant on the other thread.

I will follow up when I hear back from Gigabyte. Really should have got an Asus. My old 9900k ran perfectly fine on a Aorus board, so I thought it would be a good experience.
 
I will follow up when I hear back from Gigabyte. Really should have got an Asus. My old 9900k ran perfectly fine on a Aorus board, so I thought it would be a good experience.
It looks so hard trying to do anything with Gigabyte boards these days. How can they expect anyone to buy their motherboards in the future if they don't give proper support and there is also no presence of them here. I stopped buying Gigabyte boards since my X99 board died randomly with no OC and just happened my local shop was out of Gigabyte boards and offered me to switch to ASUS X99. Never looked back.
Taichi OCF is my first Asrock board, so far seems ok got some bugs here and there but there is strong presence and community so it gives me confidence. Asrock used to be part of ASUS anyway so I'm never leaving this ship.
 
  • Rep+
Reactions: ArrogantHair
I think this is part of Memory training as I had the same issues, but thing have settled down now, and I don't get those power cycles you everyone is talking about.

The issue I have is GCC keeps showing I need to update the NPU driver. No matter how many times I've installed it from GCC, a day or two later the same update shows up again. Ironically that version (32.0.100.3717) isn't even posted on Gigabyte's site, even though it's listed the latest F16 BIOS update (Z890 Aorus Xtreme AI TOP).

The current Intel NPU version I have installed is 32.0.100.3053 (confirmed in Device Manager). GCC shows 32.0.100.3717 as an update, and shows installing when I click to install, yet a few days later it's back. Also, per Device Manager, NOT installed.

Anyone knows what's going on.
 
It looks so hard trying to do anything with Gigabyte boards these days. How can they expect anyone to buy their motherboards in the future if they don't give proper support and there is also no presence of them here. I stopped buying Gigabyte boards since my X99 board died randomly with no OC and just happened my local shop was out of Gigabyte boards and offered me to switch to ASUS X99. Never looked back.
Taichi OCF is my first Asrock board, so far seems ok got some bugs here and there but there is strong presence and community so it gives me confidence. Asrock used to be part of ASUS anyway so I'm never leaving this ship.
I agree with all you say, and it's total BS, and why I also why I stay away from their GPUs. Anyway, Asus isn't much better, any why they were dumped by JayzTwoCents (YouTube video).

All that said, I've been very fortunate to have never had a major issue with any of their boards requiring any RMA issues, and why I still stick with them. But if you've got questions... better look elsewhere as you have a better chance of hitting the lottery 10 times in a role than getting hold of them... or getting any credible help if you do hit the "they answered" jackpot. BTW my last 5 builds including this one has all been with Gigabyte boards - X58A to X87 to Z170 to z270 to Z390 to Z890. The last two boards being the Xtreme series.

Anyway, unfortunately, many companies seem to be going the F the customer route :(
 
  • Rep+
Reactions: fkjchon
Discussion starter · #135 ·
I think this is part of Memory training as I had the same issues, but thing have settled down now, and I don't get those power cycles you everyone is talking about.

The issue I have is GCC keeps showing I need to update the NPU driver. No matter how many times I've installed it from GCC, a day or two later the same update shows up again. Ironically that version (32.0.100.3717) isn't even posted on Gigabyte's site, even though it's listed the latest F16 BIOS update (Z890 Aorus Xtreme AI TOP).

The current Intel NPU version I have installed is 32.0.100.3053 (confirmed in Device Manager). GCC shows 32.0.100.3717 as an update, and shows installing when I click to install, yet a few days later it's back. Also, per Device Manager, NOT installed.

Anyone knows what's going on.
Glad to hear it. Maybe I should change kits.

I noticed that with the NPU, too. I didn't look at the version. Just clicked. Yes, the GCC updates are more frequent than what is posted. My .3717 installed ok. I will watch for the next GCC update.
 
Discussion starter · #136 ·
It looks so hard trying to do anything with Gigabyte boards these days. How can they expect anyone to buy their motherboards in the future if they don't give proper support and there is also no presence of them here. I stopped buying Gigabyte boards since my X99 board died randomly with no OC and just happened my local shop was out of Gigabyte boards and offered me to switch to ASUS X99. Never looked back.
Taichi OCF is my first Asrock board, so far seems ok got some bugs here and there but there is strong presence and community so it gives me confidence. Asrock used to be part of ASUS anyway so I'm never leaving this ship.
I think I had the Asus X99. Never should have left lol.
 
  • Rep+
Reactions: fkjchon
Nah, if your NGU and D2D are that low, that will indicate it may indeed be ram. The D2D issue is above 33.

That's the thing, I have run Karhu, OCCT and other tests and mine is stable, so I am sure it is the D2D. So when they answer that it is ram OC fault, I am like nooope. When they say lower D2D - I am like - I did not pay $x for an overclocking board not to be able to overclock. lol.

For you, it has to be something small - shouldn't be the cpu, unless - are you using a cpu frame instead of the default bracket? That can cause ram issues if the frame is tightened too much. Has to be hand tight - but only to the point you feel resistance.

If not the above, then it is voltage, timing, compatibility or bios. Your kit is lower frequency, so you would think it would run with no issue. Are you on F16 or earlier? F16 I think they broke something ram related. It is not applying the same timings for XMP/DDR as it did for earlier. I reported this as well. So I manually set mine based off a MemTweakit screenshot I took earlier.

You could try to roll back to F16c or earlier. I have not overclocked ram before, so I am entirely reliant on the other thread.

I will follow up when I hear back from Gigabyte. Really should have got an Asus. My old 9900k ran perfectly fine on a Aorus board, so I thought it would be a good experience.
Yes, I am using a CPU Frame from thermalright it is just barely snug I did not crank it down super tight haha!! I will try a different BIOS version I still think I have F16c.. I also tried to change the different speed rate it still started like it has been..

I thought I fixed it when I turned off the Power limits and set them at auto instead of maxing them out.. it did boot fine for a day then I was adjusting the NGU and D2D and tried OC the ram a bit and it came back where it starts turns off the restarts in one press of the button..
 
It looks so hard trying to do anything with Gigabyte boards these days. How can they expect anyone to buy their motherboards in the future if they don't give proper support and there is also no presence of them here. I stopped buying Gigabyte boards since my X99 board died randomly with no OC and just happened my local shop was out of Gigabyte boards and offered me to switch to ASUS X99. Never looked back.
It's funny because the opposite happened to me. My first two Gigabyte mobos (one in the Core 2 Quad era and the next one Z77) both had weird issues. First one needed RAM sticks to be taken out to boot after a full shutdown. Second one would reboot on occasion while idling. It did seem to improve a bit when I upgraded the CPU from i3-2100 to i7-3770. Then next one was a Gigabyte Z97 which absolutely hated the Corsair RAM that worked mostly fine on the previous Z77. Paired it with Kingston and it worked great after that.

But having heard so much about ASROCK over the years, I gave them a try with the Z790 Sonic. Didn't regret it. For the price it was great and still is. I was going to get the Taichi OCF but Amazon shipped me normal Taichi because the mobo had a sticker saying OCF on one side and just Taichi on the other.

When I got my refund, the Taichi OCF had already dropped $50 in price but I decided not to risk going through the same thing again and went with Gigabyte AORUS Elite Wifi 7 ICE. Again, pleasantly surprised. No showstopping issues and I retired from overclocking it once it gave me a better y-cruncher score with 245KF than Techpowerup's 14900K! Why bother? That's more performance already than I was expecting.

I do believe that someone is smoking something at Gigabyte by releasing super expensive AI TOPS mobos and their Tachyon is basically unobtanium. Just glad that I didn't have to pay too much and the mobo turned out great for me.
 
There are two.

CPU VDDQ
VDDQ under Dram Voltages.

The following are CPU level settings that give you voltage headroom by increasing v to rails
CPU VDDQ = at least 1.5v
VccIO = 1.35v
VccClk = 1.35v
CPU VDD2 = 1.57v.
CPU System Agent = 1.45v

Under Dram Voltages
VDDQ and VDD2 apply v to the ram sticks. Default auto.

So you increase the 1st to give headroom.
If auto posts, you are good.
If not then you have to bump up the 2nd Dram voltages.
So if Kit 1.45v, you can try 1.5v etc.

Beginner at ram overclocking, and there is no documentation with the Gigabyte boards, so this how I think it works. Open to corrections.

For my Oc from 8200 to 8600 - I set the 1st values as described above.
Then I set the Dram voltages VDDQ to 1.5 and VDD2 to 1.45v - per recommendation of Leonman44. (auto worked for F16c and earlier, but F16 broke something and I had to apply to the Dram)
For my sticks these appear as VDD (swa) and VDDQ (swb) in Hwinfo.

View attachment 2704437
I meant in HWINFO
 
I think this is part of Memory training as I had the same issues, but thing have settled down now, and I don't get those power cycles you everyone is talking about.

The issue I have is GCC keeps showing I need to update the NPU driver. No matter how many times I've installed it from GCC, a day or two later the same update shows up again. Ironically that version (32.0.100.3717) isn't even posted on Gigabyte's site, even though it's listed the latest F16 BIOS update (Z890 Aorus Xtreme AI TOP).

The current Intel NPU version I have installed is 32.0.100.3053 (confirmed in Device Manager). GCC shows 32.0.100.3717 as an update, and shows installing when I click to install, yet a few days later it's back. Also, per Device Manager, NOT installed.

Anyone knows what's going on.
Im having the same issue with the NPU driver. I've installed it 3 times now
 
121 - 140 of 631 Posts