ABBA!

Mamma Mia - here we go again ...

Their versioning is confusing... They have 4 different digits, yet every release is just adding As and Bs to the end. Anyway, it'll be interesting to see if this does actually solidify the results. What happened to the whole thing about preserving the lifespan of the CPU?

schmidtbag:

Their versioning is confusing... They have 4 different digits, yet every release is just adding As and Bs to the end. Anyway, it'll be interesting to see if this does actually solidify the results. What happened to the whole thing about preserving the lifespan of the CPU?

https://twitter.com/1usmus/status/1171123725133070337 This may actually lower the voltage & overall average clockspeeds - too early to say for certain now though, need to wait for the final bios's to come out and whatnot.

Denial:

https://twitter.com/1usmus/status/1171123725133070337 This may actually lower the voltage & overall average clockspeeds - too early to say for certain now though, need to wait for the final bios's to come out and whatnot.

Yeah, I'm going to wait for official benchmarks. Seems a bit weird that it'd be going in the opposite direction.

MSI and Gigabyte already have the beta BIOS available albeit not officially. MSI: https://forum-en.msi.com/index.php?topic=323855.0 Gigabyte: https://forums.tweaktown.com/gigabyte/28441-gigabyte-beta-bios.html The best core on my 3800X now reaches 4520Mhz with the Aorus Master. Seems to be fine so far. [URL='https://forums.tweaktown.com/gigabyte/28441-gigabyte-beta-bios.html?utm_source=share&utm_medium=ios_app&utm_name=iossmf'] [/URL]

schmidtbag:

Yeah, I'm going to wait for official benchmarks. Seems a bit weird that it'd be going in the opposite direction.

It's kind of hard to say what direction it's going in because the higher boost states on more cores that this bios enables does improve performance on the majority of workloads - the preliminary tests on beta's are showing that and even 1usmus (the dev of the tweet I linked) confirms that. That being said they are tweaking the voltage tables downward and have been since the initial bios so there does seem to be some concern about reliability. What does that mean though? Are they concerned the processors will only last 10 years instead of 15 with a failure rate of 2%? Without knowing what AMD is actually aiming for it's kind of pointless to speculate.

Hilbert Hagedoorn:

Mamma Mia - here we go again ...

You can't have drama without new AMD product being involved.

Denial:

That being said they are tweaking the voltage tables downward and have been since the initial bios so there does seem to be some concern about reliability. What does that mean though?

Do people think an ASUS employee would risk their position by telling a lie that the reason AMD did this was because of long term reliability. I don't think so. As we all know heat and too much voltage is what kills CPUs. I totally believe ASUS on this. AMD shouldn't have over promised and now they look like fools. They couldn't have just had a good competitive product, they had to do stupid bull marketing on their products boxes. This totally misleading video is still online, btw: [youtube=prAaADB9Kck] I'm now waiting for the influx of AMD fanboy's who say this isn't a big deal and they are happy with their falsely advertised product because performance to them is just fine and it doesn't matter if AMD is in the wrong. Intel is the devil, blah blah.

Denial:

It's kind of hard to say what direction it's going in because the higher boost states on more cores that this bios enables does improve performance on the majority of workloads - the preliminary tests on beta's are showing that and even 1usmus (the dev of the tweet I linked) confirms that. That being said they are tweaking the voltage tables downward and have been since the initial bios so there does seem to be some concern about reliability. What does that mean though? Are they concerned the processors will only last 10 years instead of 15 with a failure rate of 2%? Without knowing what AMD is actually aiming for it's kind of pointless to speculate.

Right but my point is if this ABBA version is supposed to more reliably reach advertised boost clocks, wouldn't that imply higher voltages? I know, it doesn't necessarily mean that, because Ryzen's performance is heavily dependent on temperature. What I'm thinking is if all cores can have their voltages lowered, that could give them the headroom to cool off just enough to reach the single-core boost clocks at a very high voltage, even if it isn't totally sustainable. As for lifespan, in general, I personally don't think there's much to concern over. First of all, nobody buys a CPU like this expecting to run one instance of a single-threaded workload for an extended period of time. Frankly, the fact that anyone cares about single-threaded performance on this thing is kind of stupid IMO (I get that it's disappointing if your CPU isn't reaching the max advertised speed, but at that point it's more about principle than practicality). To my recollection, the CPU lowers its voltage as more cores are used (probably for the sake of maintaining thermals), so, in a normal workload that you'd expect this CPU to regularly work in, it should still remain in a safe voltage range. But also, it's not boosting the same core every time. Over the course of its life, each core will be heavily boosted, which helps spread out the potential damage. This is important, because each core slowly and evenly degrading over time is much better than having all but one or two of your cores fully functional, with the outliers being totally unreliable (at least in the perspective of the average consumer, who might not know how to lock out the defective cores).

not higher, but better use of voltages. these single core turbo's are supposed to be hit only when cores are literally sleeping.

jwb1:

This totally misleading video is still online, btw:

I should publish a "Guru3D - The Drinking Game" where rule #1 is "whenever jwb1 posts that video, take a sip".

schmidtbag:

I should publish a "Guru3D - The Drinking Game" where rule #1 is "whenever jwb1 posts that video, take a sip".

We should actually be taking bets on how many days it will take for AMD to remove it. If they even do.

jwb1:

We should actually be taking bets on how many days it will take for AMD to remove it. If they even do.

If they haven't done so now, they probably won't ever.

schmidtbag:

If they haven't done so now, they probably won't ever.

Sad and probably true. Unless.... lawsuit comes.

I haven't tested the auto overclocking but as of this BIOS that video is accurate. PBO successfully brings multiple cores to the advertised boost clock frequency. Whether or not AutoOC can take you over that, no idea - but overclocking was never a guaranteed thing to begin with. Edit: Yeah computebase shows a 3800x hitting 4.575 - that's punching above the boost clock. So that video is entirely accurate as of this beta bios.

Can you hear the trolls fernando?

schmidtbag:

I should publish a "Guru3D - The Drinking Game" where rule #1 is "whenever jwb1 posts that video, take a sip".

We're gonna need more alcohol then. I've never seen someone repost a particular video so much, LOL. He probably has it saved in his favorites and posts it whenever he gets a chance - watch, the next thread about testing results and he'll post it again, along with the exact same message about false advertising and AMD fanboys.

Denial:

I haven't tested the auto overclocking but as of this BIOS that video is accurate. PBO successfully brings multiple cores to the advertised boost clock frequency. Whether or not AutoOC can take you over that, no idea - but overclocking was never a guaranteed thing to begin with. Edit: Yeah computebase shows a 3800x hitting 4.575 - that's punching above the boost clock. So that video is entirely accurate as of this beta bios.

I suspect the real issue was that cores weren't being slept properly on the previous SMU which brought the power limit into play and kept the chip from boosting properly based on voltage limits. by keeping cores asleep properly, boost can sustain, ramp and power usage would lower overall.

jwb1:

Sad and probably true. Unless.... lawsuit comes.

JWB1, why are you such a fanboy my dude. I have never purchased an AMD CPU in my life, but people like you are the ones who make me feel bad for buying Intel. Seriously, I would suggest getting a real job and acting your age, as a start. Things happen with early gens, I can remember my SATA port problems with Sandybridge. So, just relax. And as far as we know Intel hasn't produced a new gen since Sandybridge. EDIT: typos

craycray:

JWB1, why are you such a fanboy my dude. I have never purchased an AMD CPU in my life, but people like you are the ones who make me feel bad for buying Intel. Seriously, I would suggest getting a real job and acting your age, as a start. Things happen with early gens, I can remember my SATA port problems with Sandybridge. So, just relax. And as far as we know Intel hasn't produced a new gen since Sandybridge. EDIT: typos

Owned AMD before. Building an AMD system for Plex. Main rig is Intel. I call out AMD and Intel on BS. It just so happens AMD has more BS and their fanboys.