Well on the AMD FX platform that particular feuture is called cool n Quiet.
On the intel platform 4790K that feuture is called intel speedstep.
Not sure what the exact naming of the future is on Ryzen.
But there should be a function to set it.
Still it shouldnt really matter much, its just a powersaving feuture.
Yeah, not booting is probably LLC. Voltage droops under load, and LLC compensates for that.
Agree with MisteryAngel that increasing voltage makes sense too. You see people talking about going all the way up to 1.40v on Ryzen, 1.41v on water. I don’t know that I’d be comfortable with that but you can definitely go over 1.36.
You will want AMD’s version of speedstep working. Not now, not while you’re trying to find your max overclock, but for end-use. Unless you get free electricity and live somewhere really cold, I guess.
Also remember to overclock your RAM. Memory speed is a huge deal on Ryzen.
Of course you could indeed also go the easy route and just set 1.40V or something like that,
and not bother with LLC but yeah…
If you know how your particular board handles LLC levels, it can be very nice.
Especially on the Taichi which has a very great vrm.
Yeah I got my RAM overclocked to 2933 using the XMP profile
Corsair - Dominator Platinum 16GB (2 x 8GB) DDR4-3000 Memory
@MisteryAngel @Ruffalo So using your advice I adjusted LLC to 4 and 5 and still couldn’t get it to boot. I’m fine with 3.8Ghz. I had 3.8Ghz working pretty stable t 1.35v, but just for testing purposes I backed it back down to stock and it seems to be working fine. When I’m stress testing what am I looking for in terms of voltage? Like if I stress test and after no crashes etc… then I should be good then?
So because I am encountering that 15.5x multibug the workaround prevents downclocking when idle. Until Asrock releases a new bios update that fixes that issue I’m just going to use Ryzen Master to overclock when I play games or do anything heavy CPU centric and leave the CPU bios OC to stock settings.
If only there was a me_cleaner for AMD’s ASP/PSP (AMD Secure Processor / Platform Security Processor) that would really make the question a no-brainer.
Although if Ryzen gets popular enough, hackers might start poking around and find something similar. Maybe the NSA will start demanding a HAP_disable bit in ASP/PSP?
im personally not too concerned about AMD ASP/PSP. Until there is a confirmed security flaw, I will take ASP/PSP over Intel ME.
09-28-17 - Vulnerability reported to AMD Security Team.
12-07-17 - Fix is ready. Vendor works on a rollout to affected partners.
01-03-18 - Public disclosure due to 90 day disclosure deadline.
Fixed - something Intel has yet to do with ME.
Again, the situation with ASP/PSP is less concerning than Intel’s ME.
Definitely less concerning, but it’s still a shit situation.
but they def fixed the issue compared to Intel. So, I am not really concerned about ASP/
PSP
I’m concerned about it from a “it’s in my system, has control over everything and I can’t remove it.” perspective.
im accustomed to that based on Intel. Ive been stuck with a 4770k since 2012.
Every component that is vital, ever.
I see the push for open source, but open source everything is not going to make the world a better place.
Not everything, true.
The problem is that if we need complex software to initialize the x86 hardware, it should be open source because some people like to be able to trust the platform they’re running on. I totally understand if you don’t care about it, but I think it’s a major problem.
Would like to have this as “have the option to make it open source”. Like you can switch the BIOS chips on some motherboards, have this enabled or disabled by a DIP switch or by replacing an 8-pin microcontroller.
I agree. I’m all for that.
If you bother open sourcing it, why would you maintain a proprietary fork? Just look at what’s happened with IBM and OpenPOWER firmware, most recently with the testing suites; why would you go back to the proprietary leaning tower of crap when you have a choice:
That’s one of the beauties of open source, it forces people to write better code or suffer the humiliation. Unless you are OpenSSL apparently…
While it may work out, there isn´t just suddenly a better open source version.
If it were to happen, it would start off slowly and then ramp up to some margin of the linux user base and the two windows-wierdos.