Ryzen memory and overclocking
Ryzen memory and overclocking
Ok guys maybe someone can explain what's going on. First of my specs are:
CPU- Ryzen 1700x
MOBO- MSI X370 Gaming pro carbon
GPU- GTX 1080
RAM- Corsair vengeance LPX 3000Mhz 2x8gb
PSU- 650W EVGA G3
I updated my bios from version 1.4 to 1.5. This naturally erased the overclock I had previously so I set the multiplier to 3.9ghz with everything else set to auto. I booted back into Win10 and evrything was ok just like before. Given the bios update incluided memory stability improvements I decided to try getting it above the 2133 it's been running at. I set it to 2667 again everything auto and it booted fine. It obviously wasn't stable because when I restarted the pc to go back to bios it did a memory reset. So I adjusted the RAM voltage to 1.35v and it just did a memory loop. So it put it back down to 2133mhz. Here's the confusing bit. Any time I either adjust the voltage for my CPU or RAM bios shows my OC at 3.9ghz Windows then shows 3.4ghz CPUZ and Ryzen master show the same. Not to mention the bios, Ryzen master and cpuz both show that the voltage adjustment has no effect. Anyone know the reason for this?
I reset my bios and applied the 3.9ghz OC and left the memory at 2133 until the next bios update.
Insomniac Jack here – maybe someone can shed some light on what’s happening. To recap my setup:
CPU: Ryzen 1700x
Motherboard: MSI X370 Gaming pro carbon
GPU: GTX 1080
RAM: Corsair vengeance LPX 3000Mhz 2x8gb
Power Supply: 650W EVGA G3
I upgraded my BIOS from version 1.4 to 1.5, which wiped out the overclock I had set before. So I adjusted the multiplier to 3.9GHz and left everything else auto. After restarting in Windows 10, everything worked just like before. With the BIOS update bringing memory stability improvements, I tried pushing it above 2133MHz again. I set it to 2667 with auto mode, and it booted normally. But when I restarted to go back to the BIOS, it performed a memory reset. I then changed the RAM voltage to 1.35V, which caused it to loop back down to 2133MHz.
The confusing part is this: whenever I tweak the CPU or RAM BIOS settings, the overclock value changes in Windows – sometimes showing 3.9GHz, other times 3.4GHz for the CPU, and Ryzen Master matching it. The BIOS itself doesn’t reflect these changes, and even the Ryzen Master and CPU-Z show no impact. Anyone know why this is happening?
I reset the BIOS, applied the 3.9GHz setting, kept the RAM at 2133MHz until the next update, and then got the next BIOS release. That update includes AGESA 1.0.0.6 from AMD, which should help Hynix chips, though my chips are Hynix and I’m fine running them at full speed.
On the RAM side, have you tried increasing the NB (SoC) voltage a bit? My current setting is 1.100V, which really helps with memory overclocking. I also set the CPU LLC to level 4 and the NB LLC to level 5. Since we’re using the same board and chips, we should be okay.
Seanie280672 :
Insomniac Jack :
Ok guys maybe someone can explain what's going on. First of my specs are:
CPU- Ryzen 1700x
MOBO- MSI X370 Gaming pro carbon
GPU- GTX 1080
RAM- Corsair vengeance LPX 3000Mhz 2x8gb
PSU- 650W EVGA G3
I updated my bios from version 1.4 to 1.5. This naturally erased the overclock I had previously so I set the multiplier to 3.9ghz with everything else set to auto. I booted back into Win10 and evrything was ok just like before. Given the bios update incluided memory stability improvements I decided to try getting it above the 2133 it's been running at. I set it to 2667 again everything auto and it booted fine. It obviously wasn't stable because when I restarted the pc to go back to bios it did a memory reset. So I adjusted the RAM voltage to 1.35v and it just did a memory loop. So it put it back down to 2133mhz. Here's the confusing bit. Any time I either adjust the voltage for my CPU or RAM bios shows my OC at 3.9ghz Windows then shows 3.4ghz CPUZ and Ryzen master show the same. Not to mention the bios, Ryzen master and cpuz both show that the voltage adjustment has no effect. Anyone know the reason for this?
I reset my bios and applied the 3.9ghz OC and left the memory at 2133 until the next bios update.
The next bios update will have the AGESA 1.0.0.6 in it from AMD and is supposed to help Hynix chips support, although mine are Hynix and I dont have any problems running mine at full speed.
Anyway, onto the problem, whenever your bios resets like that, although the bios says 3900 for your cpu, its actually quietly reset that too, so what I do is I dial mine down to 3800, reboot back to the bios, set it back to 3900 and reboot, its fine then.
With regards to your ram, have you tried to raise the NB (SoC) voltage a little, mine is currently set to 1.100v, this helps a lot with memory overclocking, I also have LLC for the CPU at level 4 and LLC for the NB at Level 5, you and me have the same board and chip.
I would but it seems when I manually set a voltage it never actually takes hold. Even with my overclock if I manually set it to 3.5 volts and boot back into the BIOS it says its at 3.6. The same thing happens with my RAM. When I set the voltage to 1.35 volts and boot back it shows something completely different. I'm not sure whether I had this issue before the update. I didn't pay that much attention.
Insomniac Jack :
Seanie280672 :
Insomniac Jack :
Ok guys maybe someone can explain what's going on. First of my specs are:
CPU- Ryzen 1700x
MOBO- MSI X370 Gaming pro carbon
GPU- GTX 1080
RAM- Corsair vengeance LPX 3000Mhz 2x8gb
PSU- 650W EVGA G3
I updated my bios from version 1.4 to 1.5. This naturally erased the overclock I had previously so I set the multiplier to 3.9ghz with everything else set to auto. I booted back into Win10 and evrything was ok just like before. Given the bios update incluided memory stability improvements I decided to try getting it above the 2133 it's been running at. I set it to 2667 again everything auto and it booted fine. It obviously wasn't stable because when I restarted the pc to go back to bios it did a memory reset. So I adjusted the RAM voltage to 1.35v and it just did a memory loop. So it put it back down to 2133mhz. Here's the confusing bit. Any time I either adjust the voltage for my CPU or RAM bios shows my OC at 3.9ghz Windows then shows 3.4ghz CPUZ and Ryzen master show the same. Not to mention the bios, Ryzen master and cpuz both show that the voltage adjustment has no effect. Anyone know the reason for this?
I reset my bios and applied the 3.9ghz OC and left the memory at 2133 until the next bios update.
The next bios update will have the AGESA 1.0.0.6 in it from AMD and is supposed to help Hynix chips support, although mine are Hynix and I dont have any problems running mine at full speed.
Anyway, onto the problem, whenever your bios resets like that, although the bios says 3900 for your cpu, its actually quietly reset that too, so what I do is I dial mine down to 3800, reboot back to the bios, set it back to 3900 and reboot, its fine then.
With regards to your ram, have you tried to raise the NB (SoC) voltage a little, mine is currently set to 1.100v, this helps a lot with memory overclocking, I also have LLC for the CPU at level 4 and LLC for the NB at Level 5, you and me have the same board and chip.
I would but it seems when I manually set a voltage it never actually takes hold. Even with my overclock if I manually set it to 3.5 volts and boot back into the BIOS it says its at 3.6. The same thing happens with my RAM. When I set the voltage to 1.35 volts and boot back it shows something completely different. I'm not sure whether I had this issue before the update. I didn't pay that much attention.
I just loaded XMP profile for my ram, left everything else alone, excpet ram voltage which I set to 1.37v
For the CPU I set 1.375v for 3.9ghz, pretty much sits there, I think it says something like 1.374v in the bios, but thats good enough.
and for the CPU NB voltage I set 1.100v and I think that sits around 1.098v again, close enough.
Set LLC level 4 for the CPU and LLC level 5 for the CPU NB.
On my MSI B350m Mortar motherboard, I needed LLC level 2 for the CPU to keep the voltage locked, but didnt need anything for the NB, it just sat stable at 1.100v all day long, however, last night when I ran IBT on my Pro Carbon X370 motherboard, I noticed the NB voltage was drooping, so I started playing with the NB LLC in the bios until I could run IBT without any drop to voltage at all, came out at level 5 for the NB.
After flashing the bios, did you clear the cmos with the jumper on the board ? this clears any conflicting data between the 2 bioses, (previous and new), ive had all sorts of weird and wonderful problems with boards in my time from forgetting to clear the bios after a flash.
I fixed the CMOS right away when I noticed inconsistent voltages and realized the OC wasn't displaying in Windows after resetting it. Once the reset was done, the OC remained stable but the voltages stayed unusual. It might help to adjust the LLC settings since they're still in Auto mode. When I applied XMP settings, I encountered a RAM boot loop, so I'm hoping the upcoming update will let me use either XMP or at least support a looser timing standard. I've also experimented with the new Try It RAM feature, but it didn't resolve the issue. The extra frames and speed improvements won't be obvious except in benchmarks, and possibly just a few FPS gains that I won't really notice with my current configuration.
Insomniac Jack shared his troubles. He cleared CMOS when he noticed conflicting voltages and saw his OC not appearing in Windows after a reset. Although the OC held after reset, his voltages remained unusual. He suspects the LLC might still be on Auto mode, which could be the problem. Setting XMP to 1 should help, but he's experiencing a RAM boot loop with XMP enabled. He tried the new try it RAM feature without success, though he hopes future updates will support either XMP or looser timings. The extra FPS won't be noticeable except in benchmarks, and possibly a few frames on his current setup. He asks for help enabling XMP profile 1 with specific settings to see if it resolves the issue.
Seanie280672:
Confused Jack shared his troubles. He reset the CMOS when he noticed inconsistent voltages and saw that after resetting the OC it didn’t appear in Windows. Once reset, the OC remained stable but the voltages stayed odd. He thinks adjusting the LLC settings might help since they’re still on auto mode. Setting XMP causes a RAM boot loop, so he’s hoping for an update that lets him use either XMP or a looser timing version. He tried the new try it RAM feature but it didn’t work. He expects future improvements to increase FPS and speed, though he won’t notice right away with his current specs.
He suggests trying XMP profile 1 with a clock speed of 2933mhz, RAM voltage at 1.35v, NB voltage at 1.100v, and RAM speed set to 2666mhz. He plans to save and exit to test. He’s about to finish a YouTube upload in about 15 minutes before trying again.
He also mentions that this time the voltages stayed close to expected levels when manually adjusted, but the boot loop persisted. He’ll wait for the next update to see if things get better.
Insomniac Jack shared his experience with Seanie280672 about CMOS issues and reset problems. He mentioned that after resetting the OC, his system still displayed issues, and voltages remained inconsistent. He considered adjusting the LLC settings and tried XMP profiles but faced a RAM boot loop. He suggested trying the XMP profile 1 with specific voltage settings and noted that the extra FPS would only be noticeable in benchmarks. He also mentioned completing a YouTube upload soon and would try again then.
Seanie280672 shared his troubles with conflicting voltages and an OC that didn't appear in Windows after resetting. He suspects the issue might be related to the LLC being on Auto mode. He's considering adjusting XMP settings, trying the new try it RAM feature, and is open to testing XMP profile 1 with specific voltage and speed changes. He also mentions his current performance improvements are minimal and plans to finish a YouTube upload before revisiting the issue. He notes that previous attempts to fix the problem didn't resolve the boot loop, and he's hoping the next update will help. He also reflects on past experiences with similar problems and the importance of stable voltages for better performance.