AMD's Ryzen 3000 Boost-Fixing BIOS Leaks out

tangoseal said:
I'm gonna p95 mine later. I'm running 7x120mm radiator lol. I predict that I have excessive cooling but Temps will still hit 80c.
Click to expand...
I bet you will also hit 80c+ and if your 3900X holds 4400Mhz I may go buy one.

sirmonkey1985 said:
just run the normal p95 and select blend test instead of small fft.
Click to expand...
I know how to run Prime95 and I know the clocks one would get with Prine95 Blend test and I know the temperatures also.
 
Riccochet said:
I have the F6a BIOS on my x570 Ultra. It's stable. Boost clocks are definitely higher, but I'm still not hitting 4.4 on any cores. 4374 is the most I've seen on only one core.
Click to expand...
Check your base clock it will be 99.xMHz instead of 100MHz hence the missing gigahertz.
Let it manage your volts.. It knows best. Zen2 updates on win10 in ms range, so it has such fine voltage control you can't always even measure it on software correctly. Just let it do the voltage control and it'll be fine.
In past people only forced voltages if they were just running all core loads and needed to push the speed as high as possible. You might get a few percent over amd internal cpu management in zen+ not sure if that counts for zen2 though. I think the improvements to granularity have made that practically void.
 
so the new bios is out legit for the asus ws x570 motherboard at least. Flashed it... Didn't notice much difference. I still had to manually set the EDC to > 140, as auto would always set it to whatever the cpu current capability setting was set to, which maxes at around 140 amps This is too low for all core regular PB frequencies.

I set the EDC to 200 and instead of pegging it at 100%, I now only reach 81% of that. So it looks like PBO is still not attempting to push up to 200Mhz over the PB2 frequncies. At least for all core. I did for a moment see 4.8Ghz for a split second on one of the golden cores. But sustained speeds were basically the same as before the new bios.
 
Riccochet said:
I have the F6a BIOS on my x570 Ultra. It's stable. Boost clocks are definitely higher, but I'm still not hitting 4.4 on any cores. 4374 is the most I've seen on only one core. The rest are in the 4200 range. CB20 scores are 476 single core and 4796 multi.

I'd love to know why I can't control CPU voltage properly. For instance. If I leave CPU voltage on auto I get nice low idle, around 1.3v all core load and around 1.45v single core load. BUT, if I set my voltage manually to, say, 1.4v, shit gets all fucked up. Idle voltages are fine, but single core load voltage goes to 1.29v and multi core load it goes to 1.1v, which results is completely shit clocks. I just don't get it. Why would my voltage be so low if it's set to 1.4v?

This new Ryzen platform confuses the hell out of me.
Click to expand...

What's your LLC set to?
 
MSI X470 Gaming Plus -BIOS 1.0.0.3AB 7B79vAE1(Beta version) with SMU FW 46.49.00 Injected by The Stilt .Nice helpful person.

Besides the actual SMU FW, these files also contain up to date PSP, PMU (IMC) FWs, bootloaders and decryption keys, depending on their original state (some already had them)

Patched BIOS by the Stilt
https://www.overclock.net/forum/11-...tack-patched-bioses-3rd-gen.html#post28128464

7B79vAE1(Beta version)-CPU Clock 4250Mhz Idle and up to 4250Mhz under light loads. PBO Override 4275Mhz
AE1MFI(Update version)-CPU Clock 3600Mhz Idle and up to 4400Mhz under light loads. PBO Override 4400Mhz+
Desktop-Screenshot-2019-09-15-14-35-06-22-2.png
 
Last edited:
Riccochet said:
Auto. Everything is on auto right now.
Click to expand...

I'm assuming you mean "everything but VCore." Looks like you're suffering some bad droop, so try setting LLC manually and start from there. Start with lower settings and work your way up so you don't overshoot like crazy.
 
BrotherMichigan said:
I'm assuming you mean "everything but VCore." Looks like you're suffering some bad droop, so try setting LLC manually and start from there. Start with lower settings and work your way up so you don't overshoot like crazy.
Click to expand...

I've tried. Actually tried every LLC setting. Oddly, if I set LLC to anything other than auto the vdroop is much worse.
 
I dare say that it must be crazy to do the coding for the bios etc if simple things cause world of grief for the end user let alone the "makers" that code them so they work at all.

I do put blame at everyone feet who use "turbo" mode, when they not specific what this means i.e amount of cores/threads increase in speed for X duration and so forth.

not really got much of those answers at all, other than fairly basic "we looking into, suffice to say, it IS hitting it's rated speeds, even though may not report as it is"

Having 10 bajillions new toggle/adjustments are awesome to have, but only when those adjustments actually do something worthwhile.

not things like LLC doing not much of anything, PBO being "broke" or at least if want to use X with Y (which should be able) prevents from work proper....that not good thing, not at all.

is VERY early on for something pretty "bleeding edge" so ofc there will be teething issues.

my .02c.....if a feature no MATTER what it might be, requries some fancy code to work right, segregate the code some and/or build in the required base line drivers stack needed....nothing like having something with a bunch of fancy blinking/noisy BS that you cannot adjust with a reply feedback as "we are aware of the issue, stay tuned"

Not good enough.....buy with featrues they should be 98% functional from the get go, otherwise....not ship it and expect full price.....these are not video games and DLC, these are products that should work, and correctly at that, beyond some driver quirks (out of the maker in question control)

^.^
 
One intriguing thing i have noticed is that my 0.1 undervolt is no longer stable with the new bios for the Asus WS x570.

I wonder if they built in a more aggressive low voltage to PB/PBO to extend the clocks at higher frequencies. I haven't tested much since borderlands is more of a priority at the moment than a 1-2% change in performance.
 
Last edited:
I'm still on ABB but if I clock my RAM to 3600 it kills all boost on my 3600 and at best I get 4150 single core. Running my RAM at 3200 c14 I boost to 4300. Anyone missing some clock speed should try lowering memory clocks and see what happens.
 
tom_ozahoski said:
I'm still on ABB but if I clock my RAM to 3600 it kills all boost on my 3600 and at best I get 4150 single core. Running my RAM at 3200 c14 I boost to 4300. Anyone missing some clock speed should try lowering memory clocks and see what happens.
Click to expand...

To be honest you'll get better performance overall with the higher clocked ram than another intermittent 100 Mhz on the CPU.
 
Riccochet said:
To be honest you'll get better performance overall with the higher clocked ram than another intermittent 100 Mhz on the CPU.
Click to expand...

What he said...I am running my ram @ 3533 14-14-14-28-1T and is super snappy vs slower speed/timings.
 
ccityinstaller said:
What he said...I am running my ram @ 3533 14-14-14-28-1T and is super snappy vs slower speed/timings.
Click to expand...

I have no doubt the sustained faster memory clocks (which also speed up infinity fabric) are more beneficial than a momentary spike in CPU frequency, but i'm going to say outside of placebo, there's no way you notice a "snappier"system when we are talking about latency difference of a few nanoseconds. For a sustained workload over the course of a meaningful amount of time, sure, but that's not typically when people use the adjective "snappy"
 
Darth Ender said:
One intriguing thing i have noticed is that my 0.1 undervolt is no longer stable with the new bios for the Asus WS x570.

I wonder if they built in a more aggressive low voltage to PB/PBO to extend the clocks at higher frequencies. I haven't tested much since borderlands is more of a priority at the moment than a 1-2% change in performance.
Click to expand...

Glad you mentioned the undervolt. I've been running -0.1 from the get go, but haven't upgraded to the new BIOS yet. I am still dialing in RAM timings, when I have the time.
 
RamonGTP said:
I have no doubt the sustained faster memory clocks (which also speed up infinity fabric) are more beneficial than a momentary spike in CPU frequency, but i'm going to say outside of placebo, there's no way you notice a "snappier"system when we are talking about latency difference of a few nanoseconds. For a sustained workload over the course of a meaningful amount of time, sure, but that's not typically when people use the adjective "snappy"
Click to expand...

Yes there absolutely is a noticeable difference. Gaming makes a huge difference. I can notice between 14 andn16 ns at 3600mhz. My games feel smoother and snappier across the board. It is a tangible effect to the senses. You.have to be a gamer that has years of constant gaming experience to have senses tuned enough to notice and trust me it's there.
 
tangoseal said:
Yes there absolutely is a noticeable difference. Gaming makes a huge difference. I can notice between 14 andn16 ns at 3600mhz. My games feel smoother and snappier across the board. It is a tangible effect to the senses. You.have to be a gamer that has years of constant gaming experience to have senses tuned enough to notice and trust me it's there.
Click to expand...

Fuck I been gaming on PC forever and I can not tell a difference between 14ns compared to 16ns. Hell I can not tell a difference between running ram at 2133Mhz or 4200Mhz .I feel stupid now,think you could tell the difference between CL10 2133Mhz -CL16 4200Mhz
Memory-Chart.png
 
tangoseal said:
Yes there absolutely is a noticeable difference. Gaming makes a huge difference. I can notice between 14 andn16 ns at 3600mhz. My games feel smoother and snappier across the board. It is a tangible effect to the senses. You.have to be a gamer that has years of constant gaming experience to have senses tuned enough to notice and trust me it's there.
Click to expand...

Gaming is a sustained load over a period of time. Which I said you could notice a difference there. I’ve never heard anyone refer to games as “snappy” until I read your post just now and I’ve been gaming since commander keen
 
gerardfraser said:
Fuck I been gaming on PC forever and I can not tell a difference between 14ns compared to 16ns. Hell I can not tell a difference between running ram at 2133Mhz or 4200Mhz .I feel stupid now,think you could tell the difference between CL10 2133Mhz -CL16 4200Mhz
View attachment 187586
Click to expand...

Humans are not computers
We do not process information digitally. We do not reference benchmarks as our qualitative experience.

We do not look at a game and our brains process a spreadsheet of benchmark data.

We barely have displays that can approach even average visual acuity. We so not see or hear or taste or touch in digital data.

You can post all the hardware benches you want and it has no factor in the quantum senses our bodies possess.

I can easily tell the difference between 14 and 16ns or 100 and 240hz etc... because human senses are adaptive and superior to anything a computer can mimic or render.

Stop trying to justify the limitations of human senses based on some mumbo jumbo quantified spreadsheets.

Human and biological sensations are far superior to digital because we are superior to machines in every way.
 
RamonGTP said:
Gaming is a sustained load over a period of time. Which I said you could notice a difference there. I’ve never heard anyone refer to games as “snappy” until I read your post just now and I’ve been gaming since commander keen
Click to expand...

We are all different..I have suffered from Chronic migraine daily for over 15 years now. I am extremely sensitive to input lag (even on the desktop), studdering (FreeSync was a miracle for me) and motion. I can set my system to run at 2133, and it is noticbilty slower...Benchmarks prove it, but I am talking about the experience from the time you press the power on button, to loading apps etc.

Would I go out and spend hundreds more on a higher rated kit? Not really. I paid $45ish more for my 32GB of B die vs the cheaper Micron E die based Crucial kit I had. The B die is running much faster (3533c14 vs 3200c16) and the system seems much responsive.

I do not think it is placebo, but it could VERY WELL be related to BIOS coding etc....Ignoring this, the guy that I quoted said that faster ram and IF fabric makes the system faster then ~100Mhz of peak boost will, and I agreed with him. I believe I was the one that used the term "snappier", and it is just that. Would you feel the same? I dunno but it is an improvement for me!
 
if that were true, We'd have to treat audiophiles as something other than huge placebo victims with a laughable hatred of having money.

No sir, I will not look at audiophiles with respect.
 
RamonGTP said:
Gaming is a sustained load over a period of time. Which I said you could notice a difference there. I’ve never heard anyone refer to games as “snappy” until I read your post just now and I’ve been gaming since commander keen
Click to expand...

Snappy, quicker, load faster, render faster, delay less

Not sure how many ways something can define snappy but whatever. You want to have a battle of semantics. Look past this reply I dont give a rat's behind what your ram speed is or what your senses perceive. Those are yours. Not mine. My reality is mine and mine only. You cant place my perceptions in a box with a million others. After all I think therefore I am
 
Darth Ender said:
if that were true, We'd have to treat audiophiles as something other than huge placebo victims with a laughable hatred of having money.

No sir, I will not look at audiophiles with respect.
Click to expand...

There is a HUGE difference between an "audiophile" and noticing a system that has a higher amount of input lag etc. One can actually be measured...I know you were not quoting me directly but I still think the logic is sound. And no, we do not want to look at audiophiles with respect. It cracks me up that someone would rather have a weak GPu, say a 2060 super coupled with a pair of $1300 'cans vs the guy with a 2080TI and a $200 set of 'cans....The second is always going to have a better experience.
 
ccityinstaller said:
There is a HUGE difference between an "audiophile" and noticing a system that has a higher amount of input lag etc. One can actually be measured...I know you were not quoting me directly but I still think the logic is sound. And no, we do not want to look at audiophiles with respect. It cracks me up that someone would rather have a weak GPu, say a 2060 super coupled with a pair of $1300 'cans vs the guy with a 2080TI and a $200 set of 'cans....The second is always going to have a better experience.
Click to expand...

true. There is also significant value in being able to communicate perception in a way that can be measured and compared in order for your statements and implications thereof to not be just discounted as the ravings of a lunatic who thinks they define their own reality however they see fit.

Can you tell the difference of context switching between 3200 and 3600 ram? Can you tell the difference between the frame rate offered by those two ram speeds? Maybe...but if it ends up being ridiculously impossible to measure or so small that it isn't believable, you might as well be telling us you can see ghosts and speak to the dead too. Everyone has different limits to their own perception, but once you start meandering into the physical limits of anyone, the old saying goes.. extraordinary claims require extraordinary evidence. You'll never get very far trying to avoid that. ...well, you could become president ...if you're from a rich family. But most other things, no.
 
tangoseal said:
Snappy, quicker, load faster, render faster, delay less

Not sure how many ways something can define snappy but whatever. You want to have a battle of semantics. Look past this reply I dont give a rat's behind what your ram speed is or what your senses perceive. Those are yours. Not mine. My reality is mine and mine only. You cant place my perceptions in a box with a million others. After all I think therefore I am
Click to expand...

There’s no chance you noticed it load faster. What does “delay less” mean? What gaming delays do you notice between say 3200mhz c18 and 3600mhz c16? I’d love to know because it sounds a LOT like BS. You’ll marginally higher FPS which I never argued but younwont notice your system being “snappier” sorry you just won’t. But don’t get side tracked. Please explain these gaming delays you speak of.
 
tangoseal said:
You can post all the hardware benches you want and it has no factor in the quantum senses our bodies possess.

I can easily tell the difference between 14 and 16ns
Click to expand...

tangoseal said:
My reality is mine and mine only. You cant place my perceptions in a box with a million others. After all I think therefore I am
Click to expand...

tangoseal *is* the new benchmark! No, wait...

Intel said you are not representative of real world application usage and AMD said you your workload perception isn't able to fully distinguish and stay accurate beyond 8 cores. I think UserBenchmark might have a job for you though.

Sorry dude.
 
ccityinstaller said:
We are all different..I have suffered from Chronic migraine daily for over 15 years now. I am extremely sensitive to input lag (even on the desktop), studdering (FreeSync was a miracle for me) and motion. I can set my system to run at 2133, and it is noticbilty slower...Benchmarks prove it, but I am talking about the experience from the time you press the power on button, to loading apps etc.

Would I go out and spend hundreds more on a higher rated kit? Not really. I paid $45ish more for my 32GB of B die vs the cheaper Micron E die based Crucial kit I had. The B die is running much faster (3533c14 vs 3200c16) and the system seems much responsive.

I do not think it is placebo, but it could VERY WELL be related to BIOS coding etc....Ignoring this, the guy that I quoted said that faster ram and IF fabric makes the system faster then ~100Mhz of peak boost will, and I agreed with him. I believe I was the one that used the term "snappier", and it is just that. Would you feel the same? I dunno but it is an improvement for me!
Click to expand...

2133MHz is an extreme example. You're literally underclocking the memory controller and infinity fabric at that point. I'm talking about running at the rated speed of 3200Mhz with "standard" PC3200 dimms with average latencies vs the more halo DIMMs. You will not notice quicker loading games, or "snappier" performance from clicking on chrome or a web page loading. You'll see a very slight FPS increase, but even that won't be noticeable outside of looking at an on screen display in the vast majority of cases, if not all of them.

tangoseal said:
When your done trolling we can continue the conversation, until then your just acting like another reddit bonehead.
Click to expand...

No need to continue the conversation. You've already established you cannot be taken seriously when you claimed you can discern a memory speed difference of 2ns. A difference that even a memory benchmark cannot discern conclusively because it's within the margin of error, so in essence, you really are the new benchmark like Azrak pointed out.
 
RamonGTP said:
No need to continue the conversation. You've already established you cannot be taken seriously when you claimed you can discern a memory speed difference of 2ns. A difference that even a memory benchmark cannot discern conclusively because it's within the margin of error, so in essence, you really are the new benchmark like Azrak pointed out.
Click to expand...

2ns is really unbelievable ,the dude has turned into a botnet.I still can not tell the difference ,oh wait I already said that shit.nevermind
 
There is more than cas 16 to 14. There are about 30 sub timings that get adjusted.


Now at this point you've just started personally attacking me rather than having a conversation. And you have done this to countless people. I could easily recall at least 5 threads where you have attacked others.

We have a serious case of Dunning Kruger effect on these forums.
 
Last edited:
tangoseal said:
There is more than cas 16 to 14. There are about 30 sub timings that get adjusted.


Now at this point you've just started personally attacking me rather than having a conversation. And you have done this to countless people. I could easily recall at least 5 threads where you have attacked others.
Click to expand...

I don't know who this post is directed at, but in case it's me, my saying I cannot take you seriously isn't a personal insult, it's just the way it is. I can't take seriously something I know to be impossible, much like I cannot take seriously politically motivated sexual assault accusations that come 40 years later. It's just the way it is... My reality, i'm sure you understand.
 
tangoseal said:
There is more than cas 16 to 14. There are about 30 sub timings that get adjusted.


Now at this point you've just started personally attacking me rather than having a conversation. And you have done this to countless people. I could easily recall at least 5 threads where you have attacked others.

We have a serious case of Dunning Kruger effect on these forums.
Click to expand...

This is weird,your saying people are attacking you and then you call every one of us in the thread an idiot,that's cool I guess ,I agree Dunning Kruger effect seems appropriate.
 
gerardfraser said:
This is weird,your saying people are attacking you and then you call every one of us in the thread an idiot,that's cool I guess ,I agree Dunning Kruger effect seems appropriate.
Click to expand...

Nope, I didnt say that about anyone in particular. Your self reflection is your self reflection. But I did not call a single person an idiot. There is zero evidence of that.
 
Seems like my post struck a nerve with some people, LOL. Anyway To my brain and a couple of limited benchmarks my computer seems happier at 3200c14 and boosting to 4300mhz than it does at 3600c18 and boosting to 4150. but that's just me, YMMV! I'm looking forward to ABBA and running my CPU faster.
 
tom_ozahoski said:
Seems like my post struck a nerve with some people, LOL. Anyway To my brain and a couple of limited benchmarks my computer seems happier at 3200c14 and boosting to 4300mhz than it does at 3600c18 and boosting to 4150. but that's just me, YMMV! I'm looking forward to ABBA and running my CPU faster.
Click to expand...

I agree. I just tested my new ABBA bios (sig system) and booted it wtih 3200 c14-14-14-34-1T timings at first....Ran CB20, and saw roughly 3% (I know its 3% but that is the difference) in MT testing vs 3533 c14-15-14-28-1T....I reloaded my old ram profile and it set those timings, so I am unsure if I have been running those the entire time or if I was running the 14-14-14-28-1T I said I was!:wideyed::wideyed:


I am now scoring 5053/503 in CB20 with PBO set to enabled/10x/200mhz offset...


Headed to do some TimeSpy testing...
 
You must log in or register to reply here.
Back
Top