You do NOT need PBO to hit fMax because the CPU is not power or current limited in lightly threaded workloads. You hit fMax with XFR, which is basically like Thermal Velocity Boost for Intel (though XFR has been around longer.) XFR is just part of the SenseMI suite along with PB2 and PBO and is enabled by default on all processors. It is stock behavior.

@btarunr Unless something fundamental changes for Zen4, this is just the stock global ceiling. Global limit for most Zen 3 parts is already +50MHz or +150MHz above the "max boost" speed advertised on the box (5050MHz for a "4.9GHz" 5950X). You're not guaranteed to see that number on a CPU you buy, but late prod. CPUs generally do.

PBO then takes you up to further +200MHz above that (silicon quality allowing), again unless Zen4 has a new higher boost override limit. Has nothing to do with changing any multipliers.

So 5.8 vs 5.8 ?
We will know it tomorrow

The biggest OC contest since ever: Intel-RL Vs. AMD-ZEN4.
Just get a room and get over it.

Show me some products AMD and Intel

Waiting for leaks that put 13900K at 5.9GHz.....

Given AMD's own claims about 8% ipc and 15% overall single thread uplift this just won't be enough to match Raptor Lake which will boost at least as high (and recent info that it can quite happily go above 6.0 even). So, they won't have single thread win, they wont have the once all important multi thread win (apart from maaaybe 7950x in some purely rendering type workloads) and it'll be a much more expensive platform to boot, at least initially. I wonder how many people will still drone about better value, lol.

HenrySomeoneGiven AMD's own claims about 8% ipc and 15% overall single thread uplift this just won't be enough to match Raptor Lake which will boost at least as high (and recent info that it can quite happily go above 6.0 even). So, they won't have single thread win, they wont have the once all important multi thread win (apart from maaaybe 7950x in some purely rendering type workloads) and it'll be a much more expensive platform to boot, at least initially. I wonder how many people will still drone about better value, lol.

Not sure where you got those numbers from (8% IPC is misleading). I know WCCFTech used that comparison. AMD claimed 15% single threaded performance uplift followed by 31% multi threaded uplift and 25% performance per watt gain comparing to Zen3. Will it be enough to match or surpass Rocket Lake time will tell but you seem to know it wont already. Even though both are not released yet.

ratirtNot sure where you got those numbers from (8% IPC is misleading). I know WCCFTech used that comparison. AMD claimed 15% single threaded performance uplift followed by 31% multi threaded uplift and 25% performance per watt gain comparing to Zen3. Will it be enough to match or surpass Rocket Lake time will tell but you seem to know it wont already. Even though both are not released yet.

Even closer to 20% single thread uplift won't even bring them to parity with Alder Lake, so unless you're saying they are either massively underselling themselves or that Raptor Lake will somehow see significant performance regression, those conclusions are pretty safe to make at this point.

HenrySomeoneGiven AMD's own claims about 8% ipc and 15% overall single thread uplift this just won't be enough to match Raptor Lake which will boost at least as high (and recent info that it can quite happily go above 6.0 even). So, they won't have single thread win, they wont have the once all important multi thread win (apart from maaaybe 7950x in some purely rendering type workloads) and it'll be a much more expensive platform to boot, at least initially. I wonder how many people will still drone about better value, lol.

I see Intel fanboys have upped their copium dosage. Of course, what you forgot to mention about the "6GHz" result for Raptor Lake is that it was achieved using a water chiller and with the E cores disabled, meaning crippled multi-threaded performance. And of course Intel is lacking the same instruction set support, with Zen 4 having AVX-512 and Raptor Lake not, making the choice easy for anyone interested in emulation.

Zen 4 X3D will also be out before the end of the year and that's just going to be a complete bloodbath. It may well take Intel several generations to even match that in terms of gaming performance. Worrying times ahead for the blue team.

AretakZen 4 X3D will also be out before the end of the year and that's just going to be a complete bloodbath. It may well take Intel several generations to even match that in terms of gaming performance. Worrying times ahead for the blue team.

:roll: :roll: :roll:
Oh my, the delusions are strong with this one!

john_Waiting for leaks that put 13900K at 5.9GHz.....

their 13900KS ~6Ghz :D

it is quite interesting to see the AMD final getting something ~5.5+GHz. Some fresh new competition in OC as well. Hope it will good match in OC btwn Blue vs Red :D

HenrySomeoneEven closer to 20% single thread uplift won't even bring them to parity with Alder Lake, so unless you're saying they are either massively underselling themselves or that Raptor Lake will somehow see significant performance regression, those conclusions are pretty safe to make at this point.

Sure that one. IPC is not something that will give you an answer of the performance uplift but only IPC. you may argue how that one is being measured.

ratirtSure that one. IPC is not something that will give you an answer of the performance uplift but only IPC. you may argue how that one is being measured.

So you're conveniently ignoring the line about 15% single thread uplift just below? Are you saying that performance numbers also don't answer anything about, well, performance? :D

When Ryzen 3000 came out, still losing to Intel wasn't enough to stop AMD from gaining market share. Intel losing to 5000 wasn't enough to make people stop buying Intel. Alder Lake was a huge step forward for Intel, in both performance and... marketing, still AMD boosted sales by just introducing more models in the market. Today high CPU performance is not an expensive luxury. With $100-$200 we get last gen 6 cores/12 threads options that are more than enough to almost everybody. Even if Intel is faster by 5-10%, or AMD wins easily with Zen 4 VCache models in games, both companies will keep selling in markets where they are behind the competition.

AretakI see Intel fanboys have upped their copium dosage.

They are greater in numbers too lately.

I'm using Ryzen CPUs for a couple of years now, right now my CPU is 5900X. What significance does the max boost frequency even have?

Remember when Der8auer disclosed with his polls that a lot of Ryzen 5000 CPUs didn't achieve their advertised boost frequency, and then AMD kind of repaired this with new AGESA? (And then nobody checked again since then.) All this had very little impact on benchmark scores or performance in games - even if you took care and used synthetic single thread benchmark.

Ryzen CPUs can't perform any task at their boost frequency, it's just an extremely artificial "we can do this too" jumping to high frequency for a short period of time when the task isn't really that hard.

I don't know if latest Intel CPUs also do this, a while back with sufficient cooling you could just run the single thread task or benchmark indefinitely at the boost frequency.

I'm not saying that AMD CPUs are slower because of that, all the benchmarks and review results were achieved with this behaviour. I'm just saying that it's a very pointless spec which tells us very little. At what frequency can Ryzen 7000 actually perform any task - single, multi-threaded, AVX-512? It's complicated...

HenrySomeoneSo you're conveniently ignoring the line about 15% single thread uplift just below? Are you saying that performance numbers also don't answer anything about, well, performance? :D

There is also the slight little detail the PL1/2 limits are again expanded for Intel. We have yet to see Ryzen guzzle north of 220W to get peak perf. Intel makes a more bursty CPU, up to insane levels. Not always a bad thing, but we are left trying to tame it.

Man i can see this thread turning into a Intel vs AMD shit show. "grabs popcorn"

HenrySomeoneSo you're conveniently ignoring the line about 15% single thread uplift just below? Are you saying that performance numbers also don't answer anything about, well, performance? :D

what about the 15% uplift? How am I ignoring it? I mentioned 15% for single thread and 31% for multi-thread uplift respectively. How is that ignoring anything?

Vayra86There is also the slight little detail the PL1/2 limits are again expanded for Intel. We have yet to see Ryzen guzzle north of 220W to get peak perf. Intel makes a more bursty CPU, up to insane levels. Not always a bad thing, but we are left trying to tame it.

Yeah, but at least as far as single thread (and light to medium threaded workloads for that matter) is concerned, it won't matter even if you set both to 125w or even lower.

ratirtwhat about the 15% uplift? How am I ignoring it? I mentioned 15% for single thread and 31% for multi-thread uplift respectively. How is that ignoring anything?

First, you said this:

ratirtNot sure where you got those numbers from (8% IPC is misleading).

Then, after I pasted the slide with both numbers, you said this:

ratirtSure that one. IPC is not something that will give you an answer of the performance uplift but only IPC. you may argue how that one is being measured.

You literally have everything on one, single slide, directly from AMD and yet you keep trying to wiggle your way out of it. :D

HenrySomeoneThen, after I pasted the slide with both numbers, you said this:

Ok so how is that not acknowledged the 15% Single Thread performance uplift?
IPC uplift is a relative metric that does not give you a lot of information. What does it actually mean IPC uplift for you cayuse I know there's been several different interpretations of what an IPC is and how it is measured.

HenrySomeoneSo you're conveniently ignoring the line about 15% single thread uplift just below? Are you saying that performance numbers also don't answer anything about, well, performance? :D

this is what you said. So again, how am ignoring the 15% IPC uplift?

HenrySomeoneGiven AMD's own claims about 8% ipc and 15% overall single thread uplift this just won't be enough to match Raptor Lake which will boost at least as high (and recent info that it can quite happily go above 6.0 even). So, they won't have single thread win, they wont have the once all important multi thread win (apart from maaaybe 7950x in some purely rendering type workloads) and it'll be a much more expensive platform to boot, at least initially. I wonder how many people will still drone about better value, lol.

I still haven't forgotten how many security issues Intel cpu's had over the years... so I could care less, still going AMD myself. AMD has some too, but not nearly as many.

BwazeI'm using Ryzen CPUs for a couple of years now, right now my CPU is 5900X. What significance does the max boost frequency even have?

Remember when Der8auer disclosed with his polls that a lot of Ryzen 5000 CPUs didn't achieve their advertised boost frequency, and then AMD kind of repaired this with new AGESA? (And then nobody checked again since then.) All this had very little impact on benchmark scores or performance in games - even if you took care and used synthetic single thread benchmark.

Ryzen CPUs can't perform any task at their boost frequency, it's just an extremely artificial "we can do this too" jumping to high frequency for a short period of time when the task isn't really that hard.

I don't know if latest Intel CPUs also do this, a while back with sufficient cooling you could just run the single thread task or benchmark indefinitely at the boost frequency.

I'm not saying that AMD CPUs are slower because of that, all the benchmarks and review results were achieved with this behaviour. I'm just saying that it's a very pointless spec which tells us very little. At what frequency can Ryzen 7000 actually perform any task - single, multi-threaded, AVX-512? It's complicated...

Your 5900x comes with a base clock of 3.7Ghz and a UP TO 4.8GHz out of the box.

If its doing 4.5Ghz all core then yeah the boost works as intended. I have a 2700X as well, constantly kept under 60 degrees. It's all core boost just sticks at 4.25Ghz for a good 6 minutes before the water starts to warm up significant. Then it starts to drop.

Intel's z170 upgrade uplift was 300MHz CPU clock speed 0MB cache 0 additional cores and 266MHz better official memory support. That's it just kidding they also provided you with MELTDOWN absolutely free!!! Sorry if I'm off topic, but the thread's pretty derailed already. If you wish feel free to compare the top of the line Gen 1 Ryzen to whatever the original top of the like Gen 1 MB's supported later on for the average effective score/overclock score on UserBenchmarks so it's very reputable and totally not shilling for Intel.

6700K vs 7700K %5 effective/7% average/5% overclock combined average 13% +5% effective less than the IPC uplift Ryzen offers generation over generation alongside cores and/or cache improvements. Go ahead and be a Intel cheerleader though. Let's see what they wrote.

The Core i7-7700K is Intel’s flagship Kaby Lake based CPU which is reported to have the same IPC as its predecessor, Skylake. Comparing the 7700K and 6700K shows that both average effective speed and peak overclocked speed are up by 7%. Most of the increase in average effective speed is explained by the 5% boost in base clocks from 4.0 to 4.2 GHz. The improved peak lab speed is attributable to a combination of better overclocking capacity and improvements in Intel’s speedshift technology which make the 7700K slightly more responsive. Kaby Lake also has marginally better HD 630 integrated graphics. The i7-7700K is priced similarly to the i7-6700K so for top end gaming and workstation builds, the 7700K is the clear choice for 2017. AMD’s Ryzen will release later this year so things could change at that time. [Jan '17 CPUPro]

It would appear they were less harsh on AMD in 2017 when they were still actively strangled by Intel's monopoly and it was entirely fresh on everyone's minds how devasting it was to the PC industry as a whole. I too would like to be a Intel fanboi with infinite dollars to spend on liquid nitrogen.

Owners of any unlocked-K Intel CPU from Sandy Bridge or onwards still have no real reason to upgrade as the performance improvements are largely academic but the i7-6700K will be the CPU of choice for the vast majority of top end PC builds in 2015.

What does that even mean largely academic? 0.1s better Pi 32m results generational uplift!!? Good times back when Intel was busy taking it easy nearly bankrupting it's only competition unless you want to include VIA that's been even less competitive yet also suffered due to Intel one can argue. Now a moment of silence to sit and pray and shill for the Intel fanboi's.

So this is what AMD offered roughly for x370 Ryzen 1800X vs 5950X by UserBenchmarksIntelCheerleaders.com 30% effective speed 41% average 43% overclock combined average 84% + 30% effective. Along with a additional +8 cores/16 thread 4MB L2 cache and 48MB L3.

Now here's a few of the juicier quotes about 5950x by UserBenchcucks.

5950x
16 cores are only suitable for professional use cases that have CPU processing needs which cannot be more efficiently met by a GPU or other dedicated hardware.

1800x
Historically AMD's CPU architecture has had a much lower IPC than Intel's and consequently Intel have dominated the CPU market since 2009.

So apparently by extension historically Intel has been a monopoly since 2009 up thru until 2017. Let's forget about the other part of AMD history with AMD64 just as long as we can historically smear them for Intel. I'm pretty certain also during that time frame from 2009 to 2017 was back when Intel was using compilers to skew things negatively against AMD in their favor as well. Meanwhile now with Alder Lake they need all the help they can get for thread scheduling how ironic is it not.

ratirtNot sure where you got those numbers from (8% IPC is misleading). I know WCCFTech used that comparison. AMD claimed 15% single threaded performance uplift followed by 31% multi threaded uplift and 25% performance per watt gain comparing to Zen3. Will it be enough to match or surpass Rocket Lake time will tell but you seem to know it wont already. Even though both are not released yet.

The 25% perfromance per watt metric refers to same wattage. Dont expect zen 4 to be more efficient than zen 3,you will be dissapointed. They will be less efficient out of the box.