Ryzen 7000 / Zen 4 Owner's Club

[Hurricane28]

5 hours ago, neurotix said:

 

For my 6400MHz OC when I was still running it on B-Die sticks, I found out I *required* 1.25v SoC voltage to run my memory at that speed with my super tight timings.

 

Make sure it's getting 1.25v *at least*, when I set my board to 1.25v it actually gives it 1.28v which is increasing heat, I tried lowering it to 1.23 but then got lockups and reboots.

 

You shouldn't need or have to use 1.3v unless you crash randomly again,, I would leave it alone because it's just going to add extra heat.

 

Glad you found the issue!

 

 

For me it shows only 1.2V for the Soc while i set it to 1.25V in BIOS.. weird, Guess i have to raise it more. 

 

[Hurricane28]

I set Soc voltage to 1.3v to check but in HWINFO64 it stays at 1.2v guess this bios locks the voltage or something. 

 

OR xmp doesn't allow it for some reason. Gonna Ask Gigabyte about this weird thing. 

[J7SC_Orion]

5 hours ago, Hurricane28 said:

I set Soc voltage to 1.3v to check but in HWINFO64 it stays at 1.2v guess this bios locks the voltage or something. 

 

OR xmp doesn't allow it for some reason. Gonna Ask Gigabyte about this weird thing. 

 

In addition to HWInfo, Zen Timings shows what looks like real-time and  fairly accurate SOC_V, keeping in mind that SOC also has its own LLC in bios (VRM Options).  Gigabyte bios should allow for up to 1.3 V set SOC_V but I keep mine at ~ 1.25 V max, and medium LLCs.

[Hurricane28]

5 hours ago, J7SC_Orion said:

 

In addition to HWInfo, Zen Timings shows what looks like real-time and  fairly accurate SOC_V, keeping in mind that SOC also has its own LLC in bios (VRM Options).  Gigabyte bios should allow for up to 1.3 V set SOC_V but I keep mine at ~ 1.25 V max, and medium LLCs.

 

Idk what AMD is doing but in Zentimigns and HWINFO64 it reads 1.2v but in aida64 it reads the correct values.

 

What is this weird AMD fuckery the whole time. 

 

[neurotix]

18 hours ago, J7SC_Orion said:

 

In addition to HWInfo, Zen Timings shows what looks like real-time and  fairly accurate SOC_V, keeping in mind that SOC also has its own LLC in bios (VRM Options).  Gigabyte bios should allow for up to 1.3 V set SOC_V but I keep mine at ~ 1.25 V max, and medium LLCs.

 

Yeah 1.25v max is what I do and leave Loadline Calibration on Auto for my board.

 

I'd imagine if you set 1.3v and max Loadline you could probably kill your chip, even with the 1.3v limit.

[Hurricane28]

21 minutes ago, neurotix said:

 

Yeah 1.25v max is what I do and leave Loadline Calibration on Auto for my board.

 

I'd imagine if you set 1.3v and max Loadline you could probably kill your chip, even with the 1.3v limit.

 

Ye man, i only want 1.250V which is needed for my 6400 but the BIOS simply will not allow it. 

 

Its stuck at 1.2v ive read on the net.. I already contacted Gigabyte about it.

 

[neurotix]

If the system is no longer crashing though when you set it to 1.25v despite it showing as 1.2v, it is probably at 1.25v? And everything is good despite the erroneous readings?

 

Again, I just went with 1.25v from the start, because I found a set of timings for the Flare X B-Die I was using and they recommended it be at 1.25v. No crashes or anything.

 

I tried to lower it one time to 1.23v to reduce heat and I was getting Bsods and reboots. Too bad none of us thought or suggested or asked what your SoC voltage was sooner.

[Hurricane28]

11 hours ago, neurotix said:

If the system is no longer crashing though when you set it to 1.25v despite it showing as 1.2v, it is probably at 1.25v? And everything is good despite the erroneous readings?

 

Again, I just went with 1.25v from the start, because I found a set of timings for the Flare X B-Die I was using and they recommended it be at 1.25v. No crashes or anything.

 

I tried to lower it one time to 1.23v to reduce heat and I was getting Bsods and reboots. Too bad none of us thought or suggested or asked what your SoC voltage was sooner.

 

I think so yes. In HWINFO64 is shows as 1.2v but as you can see in the screenshot earlier it shows as 1.25v or higher. So weird. 

[Hurricane28]

It happened again. 

 

When playing CODMW3 i got stutter issues, fps was high, and in windows i also got USB issues like typing and mouse is getting slow. 

 

I watched at Soc and it was at 1.2v in HAINFO64 but in GCC it shows as 1.250v and Aida64 shows 1.285v something. 

 

Already in contact with Gigabyte cos this is serious issue and it can do harm to the CPU if too much Soc voltage is applied. 

 

Are more people seeing this on Gygabyte boards if so, plz report so i can report to Gugabyte about this. 

[neurotix]

5 hours ago, Hurricane28 said:

It happened again. 

 

When playing CODMW3 i got stutter issues, fps was high, and in windows i also got USB issues like typing and mouse is getting slow. 

 

I watched at Soc and it was at 1.2v in HAINFO64 but in GCC it shows as 1.250v and Aida64 shows 1.285v something. 

 

Already in contact with Gigabyte cos this is serious issue and it can do harm to the CPU if too much Soc voltage is applied. 

 

Are more people seeing this on Gygabyte boards if so, plz report so i can report to Gugabyte about this. 

 

Try reducing your memory frequency from 6400MHz to 6200MHz. You won't lose very much performance, not even 1%, but it might get things stable.

 

When I was still running around those speeds on a different kit, they could not do 6400 stably at all, I'd bsod as soon as I logged into Windows or Linux or Bsod typing my pw etc.

 

If that doesn't work then reduce to 6000MHz although this *will* give you more of a performance loss. Still not even 3% though if that.

 

Hope this helps...

[Hurricane28]

7 hours ago, neurotix said:

 

Try reducing your memory frequency from 6400MHz to 6200MHz. You won't lose very much performance, not even 1%, but it might get things stable.

 

When I was still running around those speeds on a different kit, they could not do 6400 stably at all, I'd bsod as soon as I logged into Windows or Linux or Bsod typing my pw etc.

 

If that doesn't work then reduce to 6000MHz although this *will* give you more of a performance loss. Still not even 3% though if that.

 

Hope this helps...

 

Ye, im still running 6400MHz atm. Tomorrow i try to lower to 6200 and see how it goes. 

 

Still, there are many issues with AMD man. I am in contact with Gigabyte about this as there are issue with voltages and RAM related stuff. I will do some testing and send it back to Gigabyte and they hopefully contact AMD about this cos its a mess atm. 

 

I had an conversation with my retail store today which i called about this problem im having and he said there are many people with the same problems as me. 

He also wanted to participate in an solution cos AMD needs to step up their software engineering cos hardware wise they are dong okay but the BIOS is simply a mess on any board tbh. 

[ENTERPRISE]

14 hours ago, Hurricane28 said:

 

Ye, im still running 6400MHz atm. Tomorrow i try to lower to 6200 and see how it goes. 

 

Still, there are many issues with AMD man. I am in contact with Gigabyte about this as there are issue with voltages and RAM related stuff. I will do some testing and send it back to Gigabyte and they hopefully contact AMD about this cos its a mess atm. 

 

I had an conversation with my retail store today which i called about this problem im having and he said there are many people with the same problems as me. 

He also wanted to participate in an solution cos AMD needs to step up their software engineering cos hardware wise they are dong okay but the BIOS is simply a mess on any board tbh. 

I think its also Motherboard vendor related. I have no issues on my mobo. I think its unfair to assume AMD are fully to blame. Motherboard vendors are free to make their boards and BIOS as they see fit, aside from socket and anything driven by AMD AGESA . There is a lot that mobo vendors do that is completely outside of AMD's control which can futher exacerbate users issues.

[Hurricane28]

21 minutes ago, ENTERPRISE said:

I think its also Motherboard vendor related. I have no issues on my mobo. I think its unfair to assume AMD are fully to blame. Motherboard vendors are free to make their boards and BIOS as they see fit, aside from socket and anything driven by AMD AGESA . There is a lot that mobo vendors do that is completely outside of AMD's control which can futher exacerbate users issues.

Well maybe MSI is the way to go on AMD. 

 

I had issues on my Asus ROG board 2 of them and now on Gigabyte. 

 

Guess i have to save for new mobo and see if you a right. 

 

That being said, AGESA is controlling the memory and motherboard vendors have no business with that as they can only implement it. But maybe you are right that MSI has different way of implementing it which works well for AMD. 

 

I have a whole list of things send to Gigabyte, curious as to what they have to say. The positive is that Gigabyte has Rep. in the Netherlands so if there are issues i can talk to them directly. Idk how MSI is in that regard tho. 

 

That being said, after a quick google search i discovered that the entire Elite AX range is plagued with problems on Intel and AMD: 

gigabyte+z790+aorus+elite+ax+issues - Google Search

WWW.GOOGLE.COM

 

Edited November 25, 2023 by Hurricane28
add teskt

[neurotix]

3 hours ago, Hurricane28 said:

Well maybe MSI is the way to go on AMD. 

 

I had issues on my Asus ROG board 2 of them and now on Gigabyte. 

 

Guess i have to save for new mobo and see if you a right. 

 

That being said, AGESA is controlling the memory and motherboard vendors have no business with that as they can only implement it. But maybe you are right that MSI has different way of implementing it which works well for AMD. 

 

I have a whole list of things send to Gigabyte, curious as to what they have to say. The positive is that Gigabyte has Rep. in the Netherlands so if there are issues i can talk to them directly. Idk how MSI is in that regard tho. 

 

That being said, after a quick google search i discovered that the entire Elite AX range is plagued with problems on Intel and AMD: 

gigabyte+z790+aorus+elite+ax+issues - Google Search

WWW.GOOGLE.COM

 

Did you have AM5 Asus motherboards, or boards for a previous gen?

 

I am NOT happy with Asus at the moment, you guys with Gigabyte boards got AGESA 1.0.8.0 BIOS like 3 weeks ago, it still hasn't released for my board but I tried a test version and it gave terrible performance. Like 72ns latency with my RAM OC and 660 pts single thread in CPU-Z bench as opposed to 62ns latency and 720 single thread in CPU-Z. So I rolled back.

 

I check the bios spreadsheet today and see that the latest (testing) BIOS is actually AGESA 1.1.0.0 and they haven't released that either for my board, but a heads up for you guys that that is coming for you although apparently not for my ROG Strix board. In that link there might be test bios for you guys, maybe @J7SC_Orion would be interested.

[ENTERPRISE]

46 minutes ago, neurotix said:

Did you have AM5 Asus motherboards, or boards for a previous gen?

 

I am NOT happy with Asus at the moment, you guys with Gigabyte boards got AGESA 1.0.8.0 BIOS like 3 weeks ago, it still hasn't released for my board but I tried a test version and it gave terrible performance. Like 72ns latency with my RAM OC and 660 pts single thread in CPU-Z bench as opposed to 62ns latency and 720 single thread in CPU-Z. So I rolled back.

 

I check the bios spreadsheet today and see that the latest (testing) BIOS is actually AGESA 1.1.0.0 and they haven't released that either for my board, but a heads up for you guys that that is coming for you although apparently not for my ROG Strix board. In that link there might be test bios for you guys, maybe @J7SC_Orion would be interested.

 

1.1.0.0 has not dropped for my MSI board yet. Though to be honest I am not too fussed as 1.0.8.0 is working out well for me lol.

[Hurricane28]

1 hour ago, neurotix said:

Did you have AM5 Asus motherboards, or boards for a previous gen?

 

I am NOT happy with Asus at the moment, you guys with Gigabyte boards got AGESA 1.0.8.0 BIOS like 3 weeks ago, it still hasn't released for my board but I tried a test version and it gave terrible performance. Like 72ns latency with my RAM OC and 660 pts single thread in CPU-Z bench as opposed to 62ns latency and 720 single thread in CPU-Z. So I rolled back.

 

I check the bios spreadsheet today and see that the latest (testing) BIOS is actually AGESA 1.1.0.0 and they haven't released that either for my board, but a heads up for you guys that that is coming for you although apparently not for my ROG Strix board. In that link there might be test bios for you guys, maybe @J7SC_Orion would be interested.

 

That was with x370 and x470 ROG boards. 

 

I don't even have AGESA 1.0.8.0 for this board, they went from 1.0.0.7 to 1.1.0.0 but that bios crashed for me too lol. 

 

Its not the motherboard vendors alone, its combo of AGESA and the implementation of it. Been like this since Ryzen launch. I talked to Elmor and The Stilt about it and they all said that. Not in these words ofcos but they made the point clear. 

 

AMD will figure it out but takes so much time for them. This is also a reason why they support the AM5 socket for so long cos they need that time to fix it. 

 

I really really would like to know how this works as one day i am stable as a rock and all of a sudden, out of the blue it becomes erratic. This is so weird to me and i have to know it. 

[neurotix]

10 minutes ago, Hurricane28 said:

 

That was with x370 and x470 ROG boards. 

 

I don't even have AGESA 1.0.8.0 for this board, they went from 1.0.0.7 to 1.1.0.0 but that bios crashed for me too lol. 

 

Its not the motherboard vendors alone, its combo of AGESA and the implementation of it. Been like this since Ryzen launch. I talked to Elmor and The Stilt about it and they all said that. Not in these words ofcos but they made the point clear. 

 

AMD will figure it out but takes so much time for them. This is also a reason why they support the AM5 socket for so long cos they need that time to fix it. 

 

I really really would like to know how this works as one day i am stable as a rock and all of a sudden, out of the blue it becomes erratic. This is so weird to me and i have to know it. 

Yes, sorry for all your frustrations with it. I have my own too, being unable to do competitive benchmarking because my 2nd CCD won't boost to 5650 under load in 3dmark anymore. Tried so many things to fix it and nothing works.

 

Have you tried lowering your RAM to 6200MHz yet?

[Hurricane28]

2 hours ago, neurotix said:

Yes, sorry for all your frustrations with it. I have my own too, being unable to do competitive benchmarking because my 2nd CCD won't boost to 5650 under load in 3dmark anymore. Tried so many things to fix it and nothing works.

 

Have you tried lowering your RAM to 6200MHz yet?

 

I am curious what GIgabyte has to say about this man. I am on top of it and i wanna know if its Gigabyte or AMD, If its Gigabyte they need to step up their BIOS game if its AMD i am sending them some feedback on the problems and go from there. 

 

Most CPU's cannot boost much higher than that tbh. I tried Ryzenmaster software but that makes things even worse lol. After an "stable" curved optimizer in Ryzenmaster, the PC wouldn't even boot anymore and lock up lol. 

 

Without sounding too negative, its just embarrassingly how bad this mess is. 

I have several clips from how bad it really is and what i am doing etc. etc. Soon i will upload it to YouTube. Lot of my vids get a lot of fews and most of them commenting having the same problems i see too so maybe we can help each other out.  

[J7SC_Orion]

9 hours ago, ENTERPRISE said:

 

1.1.0.0 has not dropped for my MSI board yet. Though to be honest I am not too fussed as 1.0.8.0 is working out well for me lol.

 

Greetings fellow bios luddite ! I downloaded the 1.1.0.0 bios for my 670E Aorus Master the day it came out, but I am still on F13C and my 5950 / ROG Dark Hero is still on 3801

[Hurricane28]

5 hours ago, J7SC_Orion said:

 

Greetings fellow bios luddite ! I downloaded the 1.1.0.0 bios for my 670E Aorus Master the day it came out, but I am still on F13C and my 5950 / ROG Dark Hero is still on 3801

 

You run your memory at 8KMHz 24/7 stable?

[ENTERPRISE]

Hey guys,

 

Those of you using Curve Optimizer, what sort of negative values were you able to achieve per core ? Just interested to see. I am currently working on my CO, fairly involved process but getting there. 

 

I find it odd sometime that when running CoreCycler at times it will tell me I have an unstable core, on a core where no negative or positive value has been applied which is a little strange, happened a couple of times now...on different cores mind you. 

 

Thanks guys!

[neurotix]

36 minutes ago, ENTERPRISE said:

Hey guys,

 

Those of you using Curve Optimizer, what sort of negative values were you able to achieve per core ? Just interested to see. I am currently working on my CO, fairly involved process but getting there. 

 

I find it odd sometime that when running CoreCycler at times it will tell me I have an unstable core, on a core where no negative or positive value has been applied which is a little strange, happened a couple of times now...on different cores mind you. 

 

Thanks guys!

I OC with PBO... @J7SC_Orion is the one to help you with CO overclocking.

[Hurricane28]

1 hour ago, ENTERPRISE said:

Hey guys,

 

Those of you using Curve Optimizer, what sort of negative values were you able to achieve per core ? Just interested to see. I am currently working on my CO, fairly involved process but getting there. 

 

I find it odd sometime that when running CoreCycler at times it will tell me I have an unstable core, on a core where no negative or positive value has been applied which is a little strange, happened a couple of times now...on different cores mind you. 

 

Thanks guys!

 

 

I use negative -20 so almost all my cores can reach 5.750GHz. Sometimes i tried -22 but saw no difference so i stayed at -20, never had issues with it. 

 

Yesterday i tried Ryzen master but that resulted in bootloop so i had to uninstall it and never use it again as i think its not an good program for my board. 

[neurotix]

Ryzen Master is trash. Agreed 100%. When I dealt with AMD regarding my CPU not boosting, they said both Rivatuner and hwinfo64 showed wrong values (true to some extent) HOWEVER they insisted I use Ryzen Master to get core frequency values. I did, and it was way off on its clock speed values, saying cores were asleep or only at 4500MHz running all-core tests, where hwinfo64 showed the first CCD around 5000MHz and 2nd CCD at 4800MHz, I think I was running Cinebench. And the score reflected the values shown by hwinfo64 whereas Ryzen Master was outright wrong.

 

Rivatuner and hwinfo64 have BOTH been around for like 20 freaking years. They give a good general idea of what your clocks are at and they are used by some of the best professional overclockers. I remember seeing a video on Wizerty and he was running Fire Strike as well as Rivatuner (Afterburner OSD) while benching using LN2.

 

I also have tried "overclocking" using Ryzen Master with my bios at default settings for PBO and CO and it also resulted in a non-bootable system. Trash utility.

Edited November 26, 2023 by neurotix

[Hurricane28]

22 minutes ago, neurotix said:

Ryzen Master is trash. Agreed 100%. When I dealt with AMD regarding my CPU not boosting, they said both Rivatuner and hwinfo64 showed wrong values (true to some extent) HOWEVER they insisted I use Ryzen Master to get core frequency values. I did, and it was way off on its clock speed values, saying cores were asleep or only at 4500MHz running all-core tests, where hwinfo64 showed the first CCD around 5000MHz and 2nd CCD at 4800MHz, I think I was running Cinebench. And the score reflected the values shown by hwinfo64 whereas Ryzen Master was outright wrong.

 

Rivatuner and hwinfo64 have BOTH been around for like 20 freaking years. They give a good general idea of what your clocks are at and they are used by some of the best professional overclockers. I remember seeing a video on Wizerty and he was running Fire Strike as well as Rivatuner (Afterburner OSD) while benching using LN2.

 

I also have tried "overclocking" using Ryzen Master with my bios at default settings for PBO and CO and it also resulted in a non-bootable system. Trash utility.

 

Ye man. 

 

I had problems on my Ryzen DDR4 platform too, must be from the same makers that made Asur ROG armory Crate lmao. 

 

Seriously, its the worst indeed man. 

 

HWINFO64 is indeed here forever and he is hands down the best software engineer i ever met! IF there is an issue he solve it faster than any one i know in the tech world. 

And IF there is an issue its due to motherboard vendors not openly talking to him and giving him info he needs to make it better. I talked to him when i still rocked my x370/x470 motherboard and he is really straightforward and very smart. So to say its HWINFO64 or Aida64 that is at fault here is bs. 

I also talked to Aida64 rep on their forum and he claimed the same thing Mumak did, (Mumak is the maker of HWINFO64.)

 

Idk why, but to me it seems that they, the motherboard vendors and AMD themselves, put less efford in making things stable than Intel. I mean, AMD performance is outstanding and i have no complaints about my 7700x performance and stability, its the software that makes it borked and buggy. 

 

Anyway, i am back at F11A "Beta BIOS" but its better than the "non-beta" F11 *. Its been stable again and no such issues i had on the other BIOS it seems. 

Contacted Gigabyte and the Rep. send an angry letter to Gigabyte for * it up so many times and by this degree lol. This cannot happen on an "stable" version lol. 

 

Curious as to what Gigabyte has to say about this.