Overclock.net banner

46801 - 46820 of 46925 Posts

·
Registered
Joined
·
33 Posts
take it from here
use bootable usb flash with DOS and afudos tool to crossflash bios
it would be great to have results from CTR 2.0 CPU OC (frequency, voltage, CPU temp, benchmark results) and DRAM OC capabilities
manual CPU OC is interesting too (via multiplier+voltage)
I disagree. I don't believe in software that pushes unsafe voltages into the CPU.
 

·
Registered
Joined
·
277 Posts
I disagree. I don't believe in software that pushes unsafe voltages into the CPU.
no, it's not, voltages are more safe than with original C6H, at stock they are same, with OC voltages are lower specified settings with certain offset, moreover to say original bios also have such offset, just less than asrock
it's easy to check real voltage via "on CPU" SVI2 TFN sensor or via probelt points on mobo, but asrock push less voltage than specified by user, and CTR doesn't care at all, it runs with a stress test by using real feedback from mobo (Asrock have only ON CPU IO working, no crap data from SuperIO)
 

·
Registered
Joined
·
50 Posts
no, it's not, voltages are more safe than with original C6H, at stock they are same, with OC voltages are lower specified settings with certain offset, moreover to say original bios also have such offset, just less than asrock
He is talking about CTR, not ASRock BIOS. We truly believe that this BIOS is safe, or at least, it is safer that CTR xd
 

·
Registered
Joined
·
277 Posts
He is talking about CTR
with CTR you can set parameters or stress with certain profiles and settings, it doesn't matter for software what numbers it have inside (CTR have no idea about board VRM, +LLC Auto recommended, and only CPU core voltage access is needed for the software), it uses lowest possible value. We don't need to know that value as well, but only data from onCPU sensor or probelt points (just for reference vs other boards which also have voltage offset, they all have).
 

·
Registered
Joined
·
50 Posts
with CTR you can set parameters or stress with certain profiles and settings, it doesn't matter for software what numbers it have inside (CTR have no idea about board VRM, +LLC Auto recommended, and only CPU core voltage access is needed for the software), it uses lowest possible value. We don't need to know that value as well, but only data from onCPU sensor or probelt points (just for reference vs other boards which also have voltage offset, they all have).
This software isn`t good written and isn`t safe. It is not guaranteed that it will do what you expect from it. And there are conditions where CTR will punch 1.55V (exactly that) into CPU VID under workload no matter what you set in its settings.

Also, since you said that on Biostar BIOS probelt doesn`t work, I may make a conclusion that it is not true voltage, but an emulated one. From that point, it is possible that it may not represent real values.

You can measure same stuff on the other side of the socket, or from the chokes, it won`t be too much off from the real on-die values.
 

·
Registered
Joined
·
50 Posts
It is better to use ZenStates or RyzenMaster for OCing cores. At least, they are known to work as expected. Also, TurboV might still work with ASRock bios, but requires detailed hardware monitoring during testing.
 

·
Registered
Joined
·
277 Posts
This software isn`t good written and isn`t safe. It is not guaranteed that it will do what you expect from it.
we can control that during OC process

It is better to use ZenStates or RyzenMaster for OCing cores
CTR using similar methods to Ryzen Master, 1.0 version even come with condition "Ryzen Master installed nedeed", now it's just have that functions internally

And there are conditions where CTR will punch 1.55V (exactly that) into CPU VID under workload no matter what you set in its settings.
it can, but CPU VID is fake value, real value is lower, especially with Asrock bios. And for Zen3 it's safe even with highest possible value

Also, TurboV might still work with ASRock bios, but requires detailed hardware monitoring during testing.
i'd like to have PBO2\CO, but at this moment we don't have agesa 1200 for asrock
ZenStates good with Zen and Zen+, but with Zen2\3 it doesn't provide much performance (according to official website it's not even compatible with Zen3)

Also, since you said that on Biostar BIOS probelt doesn`t work, I may make a conclusion that it is not true voltage, but an emulated one. From that point, it is poss
we are talking about Asrock only, Biostar is no go for CPU OC, that bios is blocking voltage control for ANY software, it's always at stock CPU value. SVI2 TFN sensor gives me correct values with any bios
 

·
Registered
Joined
·
277 Posts
tried Zenstates with Asrock bios, now it's tie with original bios in terms of my ryzen 2600 OC, both crushes at 4100mhz in CB20, both passed well @ 4050mhz with a similar results, Asus bios have better auto controls while OC, with Zenstates i see no difference (well, that's same VRM after all)
offset it still there though, just keep in mind that Asrock need more CPU VID to achieve real target voltage under workload
 

·
Registered
Joined
·
50 Posts
we can control that during OC process
True, it is possible. Not convenient and/or simple, but possible.

CTR using similar methods to Ryzen Master, 1.0 version even come with condition "Ryzen Master installed nedeed", now it's just have that functions internally
From how stable 1.0 versions were, and how its team "decided" to completely rewrite it, I am not excited of using, or suggesting anyone, neither version.

it can, but CPU VID is fake value, real value is lower, especially with Asrock bios. And for Zen3 it's safe even with highest possible value
It is not "fake". By AMD datasheet, it is voltage that CPU requests from VRM via SVI2 TFN interface. That is why I am talking about this thing serious. Ofc it may be higher that actual on-die, coz we have Vdroop in place.
My point was that CTR may f* up with this voltage, causing VRM to spike high


And from provided info, ASRock bios might have droopier LLC settings.

we are talking about Asrock only, Biostar
I mentioned it only coz I had a point about probeIt

that bios is blocking voltage control for ANY software, it's always at stock CPU value.
SMU overrides using ZenStates/Master too? Hmmm, sounds like in giga`s low end b450 boards, with broken SVI2 TFN 🤔
 

·
Registered
Joined
·
277 Posts
From how stable 1.0 versions were, and how its team "decided" to completely rewrite it, I am not excited of using, or suggesting anyone, neither version.
latest Ryzen Master is from November 2020, guess no PBO2 and Curve Optimizer there. Nothing is ideal, but i don't think that CTR is dangerous for C6H+Asrock bios using Zen3 CPU (they have higher stock voltages and bios max is 1.55V which is about 1.45V of real voltage)

By AMD datasheet, it is voltage that CPU requests from VRM via SVI2 TFN interface
exactly, but requests and receives are not the same values, voltage drops under workload, even with original bios, if i want to have stable 1.27V (to make C6H stable under stress test) i have to set 1.31V with original bios and 1.35 with Asrock, and i have almost the same voltage on SVI TFN2 in the end, if i put 1.31 on Asrock then i get less than 1.25V on SVI2 TFN and board can't pass stress test.

SMU overrides using ZenStates/Master too? Hmmm, sounds like in giga`s low end b450 boards, with broken SVI2 TFN 🤔
not sure what you mean but with Biostar bios i always have 1.1375V in SVI2 TFN, always. And i can't go higher than 3800mhz (board crashes under stress test), software\bios changes doesn't apply to voltage value. CPU VID shows as 0.56V in BIOS and CPU-Z, SVI2 TFN shows stock voltage, and i can't do anything with that voltage
 
46801 - 46820 of 46925 Posts
Top