For reference, a stock 9900k scores 501 in that test. A 5ghz overclocked 9900k scores 522.
Apparently PBO is +200mhz and works on all Ryzen 3000 models, down to the 3600, and keep in mind that outside of these PBO slides, AMD has *not* been using the automatic overclocking feature for any of their performance numbers for Ryzen 3000 so far.
>Ryzen 5 3600 with PBO enabled bets a stock 9900k in Cinebench R20 Single Threaded
delid dis
It will be interesting to see how the 3800x and 3900x perform with PBO turned on and upgraded cooling to allow them to use the +200mhz. As AMD's own slides show, their stock coolers don't give enough thermal headroom to make full use of that +200mhz available.
Soooo, any idea if the 3600X will be more likely to overclock farther than the 3600? I'm currently deciding between the two at the moment.
As in, if it's the case that the farther you move up the 3000 series CPU stack the better the silicon quality will be due to binning. It seems like AMD has been using the best silicon to get the lowest power usage on their highest core count chips, right?
Possibly 100-200 Mhz higher. Just like the way it is today with 1600/1600X and 2600/2600X.
9900k gets ~522 @ 5ghz and ~532 @ 5.1ghz.
how dumb would it be to use the stock cooler with stock clocks and then buy an aftermarket cooler later when I have more money?
haven't those basically overclocked the same?
are there any stats?
its fine, just don't try to OC until you get the aftermarket cooler
Why would that be dumb? It's not like it costs you extra. And aren't the ryzen stock coolers supposed to be pretty decent?
>poozen has to overclock to beat intel at stock
AHAHAHAHAHA AMDJEETS BTFO
but won't PBO or whatever work with it still?
that's what I've heard but most Jow Forumsentoomen still recommend getting an aftermarket one so that's what im basing it off of
>haven't those basically overclocked the same?
Not from all the reports I've read throughout the years.
>are there any stats?
Here's one that compares 1700 vs 1700X vs 1800X. The same data can be extrapolate with 1600/X too if the data was available. But we see a ~100 Mhz higher clock stability across the these chips.
Is that with or without all the castration patches it requires in order to be secure?
>mitigation has to be disabled for Intelaviv to keep up
Dammit, here's the statistics from silicon lottery
Sure PBO works with the default cooler just like it works with any other.
PBO is basically limited by heat. You'll get more out of the CPU with aftermarket coolers compared to the default option.
The 9900k costs a little under 500$ on amazon. The 3600 is what, 250$? So your Intel loses to an AMD that costs half and your only point is "yeah well but you have to OC"?
Intel should really hire better shills. The Apple ones for example are good.
I figured as much, I've never owned ryzen before so I don't know much about that kind of stuff
AMD has already admited that they disable MCE and lock Intel CPUs to their TDPs in all Intel benchmarks. These results don't reflect real word performance.
spoonfeed me links, daddy
cope
The word overclock means "higher than what the manufacturer is pushing the chip".
And intel is pushing their chips VERY hard at the moment, which is why they're little furnaces, while AMD is using much more conservative voltages and clocks to sell em on the "it's a very cold chip" marketing.
But it's the state any chip company gets when they're being beaten with a superior technology.
Just wait until intel comes up with a new superior to Zen 2 arch, and watch the watts go up on the AMD side.
This been the thing since 1990.
>just wait
AMD also did not use the speculative execution mitigations, so you might actually get worse performance in reality, especially if you want to go the absolute secure route and disable HT on the 9900K.
If you enable MCE and unlock TDP, then Intel's 105 TDP is not real world specification. Intel's CPU should then be specified as 200+W monsters vs 65 watt AMD chips. The problem is of standardization of tests.
Except that is not needed
[x] Security don't matter
Probably not dumb at all, assuming your case won't make the procedure too annoying. Just be careful not to pull out the CPU with the old cooler.
Find me one case where someone got affected by these meme threats
Pro tip: there is none
That's what you do when you're on the losing side of something that balances like a pendulum.
>y-you d-don't need security!
Pathetic.
>shortcut speedhacks sacrificing security are not important when a multi-billion dollar company decides on their serverfarm expenditures
>9900k gets ~522 @ 5ghz and ~532 @ 5.1ghz.
Yes, I'm aware, however this is an R5 3600 we are discussing. It's normal clock range is 3.6ghz - 4.2ghz, with PBO allowing +200mhz to 4.4ghz. Comparing to a *stock* 9900k in Cinebench R20 single threaded was just to give people an easy reference point to contextualize how good a 506 score is.
That said, if you give a 3800x sufficient cooling to actually be able to achieve +200mhz over its 4.5ghz boost clock, it should obviously score better than 506.
>Ha ha you have to overclock your $199 chip to beat a $488 chip
(lol
We are talking about home users faggot.
>199
Show me where the 3600 is being sold?
>AMD has already admited that they disable MCE and lock Intel CPUs to their TDPs in all Intel benchmarks. These results don't reflect real word performance.
Where exactly did I indicate that a 501 score for the stock 9900k came from AMD? It did not. I searched independent tests of the 9900k in Cinebench R20 to find that score.
You are retarded if you think 3600 will get more than 4.2 all core even with liquid cooling.
2700X can already do 4.2 Ghz all core. Why wouldn't 3600 be able to do so when stock specification states 4.2 ghz.
You can bet your ass that 100% of the financial sector or any corporate environment that deals with money or sensitive customer data is going to enable every single mitigation imaginable.
Same goes for home users who are dealing with wealth on their computers.
>Inb4 we can only talk about gaymers.
One market for these super high clocking chips are day traders. Security holes are pretty damn relevant among them.
200
>You are retarded if you think 3600 will get more than 4.2 all core even with liquid cooling.
I don't have to "think" anything. AMD literally announced that the 3600 can be automatically overclocked to +200mhz over its stock boost clock of 4.2ghz using PBO, and it gains 21 points in Cinebench R20 single-threaded from doing so. It's literally in the presentation slide posted in the OP.
is that Vcore voltage, of CPU voltage?
Max boost is 2 cores
t. retard
There is no 2 core boost. The whole "single core boost" or "2 core boost" is Intel specified limit. AMD can do all cores if thermals/voltage requirements are being kept in check.
You clearly don't understand how PBO works. It does not have any limitations on the number of cores it will boost, its rules are based on frequency limits, TDP limits, and voltage limits. If have the cooling, turned off the TDP limit, and it doesn't require too much voltage, PBO will boost as many cores as it can to its clock ceiling, 4.4ghz in the case of automatic overclocking on the 3600.
>source: my ass
Ye just how zen+ needs 1.5+ volts to oc by 100mhz
Nice non-sequitur.
wtf, I always felt like my own 1700x was a potato since it couldn't do 4ghz stable on 1.4v
it's doing 3.9 at 1.35v (standard vid)+2 ticks offset (.0125?)
No. You are talking about home users. 'We' are talking about all users.
no I care what is best for me not the companies. I don't own any stocks unlike you so I'm not trying to shill and I don't care how a capitalist is affected.
Well, Intel most certainly thinks that their hardware vulnerabilities are bad enough to warrant pushing automated updates to all their home users and decreasing their performance, but I guess you know better than Intel.
Picture, thousand words, etc.
holy fuck you're dumb
I guess this means PBO will remove the TDP 65 shit the 3700x has?
>single thread
>r20
OH NO NO NO NO NO!
I'm honestly not sure if the TDP even means what it used to considering they've got the 16 core with a 3.5ghz base and 4.7ghz boost a 105w TDP
And an 8 core 3.9ghz base with 4.5ghz boost at the same 105w TDP.
I can't wait for July 7 to happen and the Intel Fanboy (he's not a shill, shills have scripted response, this person is just an idiot) to be BTFO'd ultimately.
He'll probably say that the independent viewers are now bought out.
Yeah I can't wait until 3rd party benchmarks are out when you won't have to:
use 5133mhz ram
lock intel's tdp and run them stock
use the best silicon chip you have
amd already got exposed by gamersnexus using fake benchmarks for their streaming slide.
depending on your bin, you might not even need to change the stock cooler, unless you're buying the 3600 (the wraith stealth is terrible).
>use 5133mhz ram
AMD did not use it. It was gigabyte's own test for their motherboard. It was for an on-air overclock.
>Inturd literally thinks that it's the speed, not the timings
in one year the 3600 is gonna be selling for $100 and gonna be beating the $500 9900k and the next incel 10core housefire in single thread
jesus christ it's a slaughterfest. incel is finished and done
Either one works to reduce latency. Difference is usually people forget to tweak subtimings which can make a big difference sometimes
Gamer's Nexus is also a really big retard and purposefully changes benchmarks to better represent Intel.
I'm not a fanboy either direction, I just want the best product possible, but Nexus changed their Watchdogs 2 benchmark from driving around the first mission to going down an alleyway and walking back the alley, just so Intel can compete with Ryzen in the FPS.
He also purposefully uses slow as fuck RAM (2666) to make ryzen look worse.
I'm okay with admitting Ryzen 2000's are obviously weaker but god damn if you have to go out of your way to do that, you're just a huge faggot.
He's a faggot who can't wait to find any tiny flaw in AMD's shit and to blow it out of proportion but when it's about Intel flaws and bribes and lies he's all "yeah this happened whatever moving on lol"
i overclock manually anyway.
intel btfo
remember when he said nothing but x570 was gonna be revealed at computex? pepperige farm remebers
>He also purposefully uses slow as fuck RAM (2666) to make ryzen look wors
checked here
gamersnexus.net
says: All desktop platforms: 2x 8GB 3200MHz Corsair Vengeance LPX 16-18-18-36 DDR4
where did he use 2666 on ryzen?
Some of those intels to my knowledge don't run above 2666, and that's just the RAM he's using, it doesn't mean he clocks it at 3200.
Also, I can confirm I ran some of the similar benchmarks and got way better FPS than he did on my 1080ti (he uses a 2080 TI for benchmarking btw) on my 2600x.
What's worse is that when he OC's Ryzens, he doesn't use PBO, he uses an unstable max boost clock OC which can wildly affect FPS, when most Ryzen users should actually be using PBO for better performance.
Lastly he always breezes through frame time graphs where it actually shows Ryzen does consistently better than Intel.
Again I'm not saying Ryzen isn't still weaker or I'm a shill or a fan boy, I just want objective facts and he's too much of a faggot to be taken seriously when it comes to these metrics.
>mfw buying the 3900x
>mfw buying the 3950x and selling the 3900x
>gamersnexus
kek, imagine unironically referencing this hack
>remember when he said nothing but x570 was gonna be revealed at computex? pepperige farm remebers
He also said Ryzen is almost as vulnerable to most of the flaws that affected Intel like a retarded nigger.
>got way better FPS than he did on my 1080ti (he uses a 2080 TI for benchmarking btw) on my 2600x.
that's really bad then
>he uses an unstable max boost clock OC which can wildly affect FPS
all core 4.2 is more unstable than pbo? how?
kek oh noh noh
Ryzens were not developed with all core boosts in mind, and their stocks are shit to prioritize keeping it cool over performance.
PBO is basically a dynamic overclock, it draws what it needs and pushes past the TDP to give better performance than OC'ing it yourself and makes it run hotter. If you use a Ryzen and have anything over the stock cooler, you should be running PBO.
This isn't some hidden information or anything, it's the literal draw of the X series over the stock (2600 cannot PBO, but the 2600x can)
People like GN should very much know it exists, and when its your job to know shit like this, especially that Ryzen's have shit manual OC's, it's just being disingenuous to lie to people or deceive the audience.
That's why I hate GN, they pretend to know everything but then willingly lie to it's audience, like why the fuck do you have to be like that?
>That's why I hate GN, they pretend to know everything but then willingly lie to it's audience, like why the fuck do you have to be like that?
Worst part is when they get called out and refuse to acknowledge that they were wrong. Fuck those guys, there are practically no good people in the "tech YouTubers".
>Meltdown : N/A
>Foreshadow: N/A
>Spectre V3A: N/A
>Lazy FPU: N/A
>Spoiler: N/A
>MDS: N/A
that was one hell of a low key jab at Intel. The bants are impossible.
You probably know more about CPUs than gaymernexusfag
Is Intel even trying? This is embarrassing
Every time they get called out they don't even apologize, they're just fucking fags honestly.
I really don't think you can have a PC tech channel and not know this, it's not rocket science, it's just malice.
GN are a bunch of fags honestly, and I hate how many people listen to their word as gospel.
still not understanding u are saying a manual 4.2ghz all core is more unstable than letting it just boost?
say a 2700x hitting 4.2 all core boost under load and running a 2700 at 4.2 locked all core
a 3600 should be able to hit 4.5 all core, right?
Yes that is exactly what I'm saying.
They cannot hold an all core force boost really well, they're a bitch to stabilize.
PBO just pushes past the TDP and lets it boost with more head room, and it'll beat any Manual OC on a ryzen unless you REALLY get lucky on the silicon lottery.
What the hell do you mean, a 2600 can utilize pbo
>GN is an intel/nvidia shill
>tomshardware is an intel/nvidia shill
>techpowerup is an intel/nvidia shill
amdrones are hilarious
Intel would NEVER bribe anyone, right?
>techpowerup is an intel/nvidia shill
Does anyone really say that? I've never seen anything that would suggest it. I can kind of see it for GN though and I don't even pay attention to tomshardware so who knows(but I guess it's hard to argue the 'stop kvetching and just buy RTX')
Like I said earlier in the thread, I get much better performance on inferior hardware than those retards do without doing anything special other than PBO.
It's their job to test shit and they don't test it right on purpose.
>Some of those intels to my knowledge don't run above 2666
They'll all run above 2666. 2666mhz is Coffeelake's JEDEC rated memory speed, the same way Zen+ is 2933mhz, and anything further is technically overclocking. Only thing with Intel platforms is they just require a flagship Z370/Z390 (if talking Coffeelake) chipset to run memory faster than 2666mhz, where as AMD isn't as draconian.
>They'll all
I'll correct myself; all unlocked
SIR PLEASE STOP I HAVE A FAMILY OF 27 KIDS TO FEED
yes sirs ples buy intel
THIS CAN'T BE HAPPENING
>They cannot hold an all core force boost really well, they're a bitch to stabilize.
well if they don't crash they are stable are they not. not any different than running the same allcore with same volts? seems memey
How's gamersnexus steve taking the news?
thinking of doing this but with the 3700x and then 3950x
I've been doing this shit for far too long, it's just AMD being dumb and me not being able to wait 2-3 months.
How much will I be able to dump my 2700x for on eGay?