RYZEN 3000 Delidded - Overclocking Expectations and Temperature Scaling
HTML-код
- Опубликовано: 22 авг 2024
- iFixit: ifixit.com/Der8...
Pro Tech Toolkit: bit.ly/2JOFD8f
Essential Electronics Toolkit: bit.ly/2JUUYo7
Manta Precision Bit Set: bit.ly/2JSpkYa
Mahi Precision Bit Set: bit.ly/2JSp1fY
Mako Precision Bit Set: bit.ly/2JOxalw
---------------------------------------------------------
Follow der8auer & Marry:
/ der8auer
/ marryfoxreal
---------------------------------------------------------
Support me on Patreon:
/ der8auer
---------------------------------------------------------
Find my products at Caseking:
Delid Die Mate 2: bit.ly/2Rhv4y7
Delid Die Mate X: bit.ly/2EYLwwG
Skylake-X Direct Die Frame: bit.ly/2GW6yyC
9th Gen OC Frame: bit.ly/2UVSubi
Debug-LED: bit.ly/2QVUEt0
Pretested CPUs: bit.ly/2Aqhf6y
---------------------------------------------------------
My Equipment:
USB- Microscope*: amzn.to/2Vi4dky
My Camera*: amzn.to/2BN4h2O
(*Affiliate Links: If you buy something, I will get a part of the profit)
---------------------------------------------------------
Music / Credits:
Phaura - Departure
/ departure
Outro:
Vendetta Beats - Angelus
(Custom Version)
Again, thank you for the English version. Two languages, double the work...wow!
Only double the recording of speak, and double graphics ;-)
@@Sihodude And double editing, rendering, uploading, description all that stuff.
I would watch the German one in CC if I had to, but this really helps. Thankfully der8auer is much more educated than I and can provide us with this.
Lul wut? He is the English twin
Yes Thank for the English language Version 👍
Another review says that there is a bios issue causing the issues with overclocks and it's affecting both x570 and x470. He said that he downgraded to an older bios version and was able to hit 4.6ghz.
@Advocatus Diaboli
That's going no-where the same person has since reported back that even newer bios don't change anything and that Ryzen just can't hit the advertised boost speeds, which is a shame :(
what is the difference between 4.4 and 4.5-7? any testing done yet?
@@rankcolour8780 i wouldnt jump to conclusions just yet. Everythings brand new and its amd
if the cpu was only going, in the case of the 4900x, to 4.4 ghz for most reviewers 4.7 ghz should be a substantial improvement.
@Advocatus Diaboli don't know why they would fix something that works the way their fine-print advertises, boost clocks are a may-reach-up-to, not guaranteed-speed.
This video saved me from trying to delid the Ryzen 3000. :)
Never had a doubt in my mind that you will be first to de-lid the new RYZEN :)
Is that a VOODOO 2 just chilling in the background there sir?
Does look like it yes.
and i think a 690
Def!
Is that a VRISKA SERKET IN YOUR PROFILE PICTURE, SIR/MA'AM?
@@daviddebroux4708 You are correct sir.
what about direct die cooling? since you don't need the IHS to clamp down the CPU, like an Intel.
back to Athlon XP times :)
Obviously u should get some more degrees, but beware the pressure on corners of die, this thing is not centered any more.
I approve of this message.
@@vivianvaldi7871 I was thinking of direct contact too, but that's a very good point. The center of pressure should be inside the "footprint" of the die, and the die and chip have to be absolutely flat. I think it should work, interesting to see if someone tries it.
My old partner "Zenzzo" was the original direct die waterpimp back in the day of the Athlon XP...wonder if I could get him out of retirement to take a crack at this. Was just thinking about this exact problem (And Zenzzo} when I was at work today...kinda weird, must mean it should happen!
Your english is pretty flawless, grammar is usually the fallback of anyone learning, but it's spot on and even on such a technical topic.
Which BIOS revision did you use? Because The Stilt has this to say:
"In case of the ASUS Crosshair VIII Hero Wi-Fi motherboard, the media was instructed to use 0066 bios build,
which had been vetted and approved by AMD. However, newer bios builds were available and ASUS has also (allegedly) told the media to use those versions. What exactly has transpired here is still under investigation, but regardless of the actual reasons behind it, the consequences might be rather significant. In practical terms, all reviews which were done on ASUS Crosshair VIII Formula or Hero motherboards using other than 0066 bios build must be considered invalid, at least partially. Reviews using other ASUS motherboard models (not provided by AMD) are under suspicion as well.
Few days ago, I noticed certain anomalies, while measuring the power consumption of the different Matisse SKUs. Inspection of the power management parameters revealed no issues, which could have explained those anomalies.
The external power measurements (VRM DCR) revealed that the CPU was consuming significantly more power, than its power management should have allowed it to. I initially suspected that this was AMDs own doing, in an effort trying to boost the performance of the new CPUs even further, but further investigation indicated otherwise.
AMD had no part in it, and the actions by ASUS are the sole reason behind it. The investigation revealed that ASUS is altering one or more power management parameters of the CPU, causing it believe it consumes less power than it actually does. As a result, the frequencies will be higher than the actual power budget would normally allow to. Tricks like this are pretty much a common (mal)practice these days however, there is a good reason why this must be considered worse than the others: this "thing" is completely undetectable without external measurements and rather deep knowledge, but also there is no way to disable it either. Even a person such as myself, who can control most things on these platforms cannot disable this "thing". As you may notice, at the moment I call this issue the "thing", since I'm giving ASUS the benefit of a doubt."
He's using a STRIX motherboard besides what could this "thing" even be? Idunno seems super fishy if it's true but needs more confirmation.
I am also curious about his bios settings, as others who have done reviews of the ryzen 3700x enabled a setting that was typically disabled that allowed the processor to run at maximum performance with out manually overclocking it.
Multi Core Enhancement all over again?
I love that you delid everything. Love it. Love watching you pop those lids off, so I don't have to. Great content as always!
HI De8aur from England. I saw a video of you with Steve from GN at computex so thought I would check out your channel and this video on Ryzen 3000 launch day. Excellent content very clear and concise with no waffle. very proffesional, thank you very much for this, I have watched so many videos today, but I got to say apart from OC3D comparrison of X470 to X570 crosshair hero versions, this was the best video Ive seen. great work, and really looking forward to watching more of your videos. P.S did you say you delid every new CPU thats launched? I winced a little when you did this. :-)
Good job, also making it clear why the delid isn't worth the risk for the return.
I am convinced now after seeing the thermals you have shown. Here is some tin-foil/complete guesses I have. Likely the reason Intel has had so many issues with 10nm is because of heat when scaling frequency. Their reference samples and laptop processors are showing great power efficiency and performance at low clocks. I bet they struggle to push out higher than 9th gen cpu performance on lower thread optimized software, and have poor scaling in the desktop CPU space. AMD looks to be matching 9th gen core to core performance, but do not have the thermal headroom to push for higher clocks and performance in lower thread optimized programs, like games. AMD built wide with the 3900x and 3950x, and added more cores, something Intel likely could do too. It just does not seem to be worth the investment for Intel to start production on a line of CPUs that will perform better in power efficiency, have more cores, but can't top their current line-up in gaming.
I bet this is why we have not seen 10nm Intel desktop CPUs, and we likely won't until they solve these issues. I wonder if there really is anywhere to go for Zen 3 at the moment, and if AMD is also fighting these issues. This is fascinating to say the least, as the core to core efficiency on Zen 2 chips seems to suffer from variance, and binning of chips could produce some outliers with better performance.
All speculation, but I wonder how close I am to the reality. Hopefully I am wrong, and the bios fixes for Zen 2 allow it to punch to boost numbers posted in their literature, and not hit thermal limits/throttle. This is going to be a fascinating couple of years in CPU competition, and I can't wait to see what happens.
AMD said themselves that they would see a frequency regression when shrinking past 14nm. We saw the same when Intel went from 32nm to 22nm.
The fact that these CPUs are hitting these clocks is expected. 7nm is brand new so maybe 7nm+ will give us higher clocks.
@@DenGuleBalje I agree. I wonder if the initial die-shrink is just not allowing competitive clocks from current engineering samples at Intel. They probably having been trying to deal with heat, power, and performance to be competitive with 8th-9th gen gaming, they likely have good IPC lift and good performance overall, like the laptop chips show, but haven't figured out scaling for desktop.
Honestly AMD is suffering the same issues, so it might be the case. The PBO numbers for Zen 2 chips are pretty damn toasty at lower clocks. The heat growth is pretty massive all core 4.3 overclocked. So I am not sure what they are going to do for Zen 3, and probably will need more RND time themselves.
I will say this though. Once Intel and AMD figure it out, the competition should be fierce for a while. I am excited. I just hope the Intel 14nm refresh rumor is false, even if they need to buy time. Having super hot, super high tdp chips, that beat or tie Zen 2 seems like a meh option. I was to see real growth from both companies.
@@Zerosixious From what I've heard Intel's 10nm process was fundamentally broken for a really long time. Meltdown and Spectre probably also caused issues because suddenly now people expect 10nm to have hardware mitigations for those vulnerabilities.
Intel threw A LOT of money at the problem and things seem to mostly work.
10nm is probably too little too late though because TSMCs 5nm is about 2-3 years out. I also HIGHLY doubt that the yields on 10nm is high enough to make them anywhere near the same amount of money as 14nm did.
My bet is that they will sell SOME 10nm products along with mostly 14nm+++++ until they jump to 7nm.
The bottom line is: Intel is bleeding money and need to act fast.
*Honest review, not sponsored, so, why buy a motherboard for overclocking if the Ryzen 3000 series is not good for that? your video make me think, im wait your next video about the new chipset, im shure after this, what combination mother + cpu is the best, regards from Argentina*
buying a 500$ mb and OC 50mhz RIP. still think ryzen is awsome just the leaks were way out of hand. 5ghz i wish
@@selmopt Ya, best to always be skeptically hopeful about "leaks" with no direct leak to the actual company itself.
@Dr. Fresh_2k the prices are only high for motherboards lmao
www.reddit.com/r/Amd/comments/cacwf9/psa_ryzen_3000_gaming_performance_is_being_gimped/
@Dr. Fresh_2k the fuck you on about? Ryzen too pricey? tis you whos on drugs moron.
*_I love you, 3000._*
OK very informative.
@10:10 Important frames to look at for the 3600 : here u see where the chipsets are located, relative to the special squared corner of the CPU.
When u are facing the VRM, the big chip is on the mid right side.
The small chip is on the bottom left side.
When pasting, this is where it matters. Better past all the CPU with a card, but if not, don't miss this 2 locations.
I have to hand it to you. You have a ton of guts just getting in there and risking high dollar equipment like that. This info helps all of us. Thank you!
its still under NDA your around 40miniutes too early be careful!
@Israel Salgado Timmy Joe fucked up too
It's all good.
Timmy Joe posted a video 2 hours ago but he removed it.
nice so i can take a walk with my doggo and then have a coffee while reading reviews thanks for the info.
@uP-Andre can you point me to the results where they get "stomp"?
After the disappointment with Ryzen 2000, this was not entirely unexpected, you ask me. But I totally agree that the way to market their products is misleading and somewhat idiotic. Ryzen 3000 = 5ghz, It is mostly what one can find online regarding Ryzen 3000 if you go a few weeks back in time, so that this realse is a disappointment to most including me, is very understandable.
On launch day you've done everything we want to see in 1 video 🙏🙏🙏🙏
der8auer you deserve millions of subscribers, absolute mad lad! love your videos.
der8auer
, which AGESA do you use? Because 1usmus wrote 06.07.2019 (google-translate from russian):
"For example, only today the first release of BIOS for Crosshair8 with AM4 Combo PI 1.0.0.3AB was released
.
My hint: tons of Excel tests that will be far from reality"
Isn't the embargo lift in about 40Min? LOL
Fail :D
He is special
German Privilege:)
Done goofed
David Pérez this is der8auer what do you expect,
I'm sure he did this last time too lol, I ain't complaining, his vids are different to the rest anyway
This really helped me in delidding my 5950x. You are a legend mate!
You can actually bite through a slug of Indium with only moderate force. Cody's Lab did a video on it. Pretty neat metal.
Yey Der8auer brings First AMD 3000 video on the tube :)
there was one 3 hours ago by someone else but they deleted it
@@topendtrucker It was this one ruclips.net/video/6C1Jr5GMnqI/видео.html
Where's 3700x? literally all reviews are either missing 3700x or 3800x..what is the actual difference between those 2? except the tdp
Paul's hardware.
Timmy Joe did the 3700x
The 3800x is actually the lower end one due the extra tdp. If it was a binned chip it would hit higher clocks at a lower or the same tdp but it doesn't so just go with the 3700x
@@Traumatree More like $70 less. And I disagree. 100 MHz isn't worth $70.
@Advocatus Diaboli IKR?
Great work as usual. It is good to see RYZEN 3000 stock quality doesn't disappoint either.
This is very disappointing overclocking results. Any explanation as to why they can't hit higher clocks even with massive voltage increase?
www.reddit.com/r/Amd/comments/cacwf9/psa_ryzen_3000_gaming_performance_is_being_gimped/
That was expected considering low 7nm process. This happens because of a higher efficiency in thermals and power consumption. The same scenario happens when you try to overclock Maxwell/Pascall/Turing GPUs and even VEGA cards, they wont OC much even if you put lots of volts, thats only making them "thermal throttling" faster.. In the end a waterblock or maybe a chiller is not enough over the stock cooler, yeah they will have lower temps but are not enough to increase clocks. Adding too much voltage to your CPU will degrade it over time, thats why Roman is right recommending 1.35-1.40 max. Also over 1.4v you wont scale much and your temps will be very high, making CPU more unstable to run anything.
This is exactly the kind of information I've been waiting for. Thank you for your hard work and honesty. Those chips are definitely impressive and I welcome them wholeheartedly but most people will not understand the why and how when they look at clocks. I will definitely reference this video!
hello i am here, waiting still 12 hours to be able to buy ryzen in australia :)
u can get it from shopping express within the next hour
ɐıʃɐɹʇsn∀ uı uǝzʎᴚ ʎnq oʇ ǝʃqɐ ǝq oʇ sɹnoɥ ᄅ⇂ ʃʃıʇs ƃuıʇıɐʍ 'ǝɹǝɥ ɯɐ I oʃʃǝɥ
Fixed it
Ryzen cpus sold put like hotcakes! It launched here in the U.S and BAM, gone. I have to wait for more cpu's.
They available to buy on pccasegear or mwave right now... Unless u wanna buy in person
I like the new way of advertising true max theoretical, as it gives the true information of what the processor can handle- but they definitely need to explain conditions somewhere, even if its on their website.
Dont know how I missed this video. Was pointed here from a forum. Best scaling video I have seen from past few days. It seems AMD should revisit temperature scalling in future bios updates to maybe not be such a harsh scale at 50 degress celcius. Subscribed.
AWESOME! Der8aur has already put up a video regarding overclocking the 3000 series!!!
In short, if someone ever tries to sell you an i3 7350K, call 911 immediately.
Especially if the price is same as ryzen 5 or even 7...seen that and its very sad realy
I really appreciate how well you did in exploring delidding this CPU
thank you for saving many people making that mistake kudos mr sacrificial lamb
People didn't realize from Zen1 that Amd cpu just cant overclock all cores as high as Intel. Is it bad no, did they mislead people no, because they said max speed would be something but not on all cores. No matter what the Amd Chips are good just if you want Overclocking abilities you go Intel. Fan boys on both sides man just be glad customers have options than one company having a monopoly which is bad for everyone in any market...
2:06 We need to talk about that, that is fked up.
yeah, and so unnecessary. amd doesn't have to bullshit anyone to sell these cpu's, the product speaks for itself. why would they ruin all that by bullshitting about the clocks?
So, i wrote with Linus, from Linus tech Tipps, it seems like there was an emergency Update, and yesterday IT came out, which is why Linus and some other youtubers had to Remake their benchmarks. Seems like it was fixed. Upvote this so people can See.
@@VoidBass6412 Does that now mean that they now can hit the Advertised Clocks ?
@@snowgentleman1937 i quote linus here : "We got an agesa update literally yesterday (had to retest everything...) that dramatically altered boost performance".
@@VoidBass6412 Well Intel has done the same for years. some of their CPUs they can't even reach boost clocks.
I'm kinda disappointed that AMD is doing the same, but "boost clock" doesn't mean "all core".
Apologies for The Longwidedness.
Your honesty/”frustration” is enlightening, to say the least. Regardless of one’s loyalty to, “The-Master”, perhaps asking, “What Happens When A-Thief is Rewarded for Stealing?”, might break “The-BeDazzlement”. And it is NOT that difficult to “Gamble-Responsibly” with The-Answer either because whenever said “Thief” then promises to, “Stop-Thieving”, that is, perhaps, due-to, “More Money Than Sense”, being realised/”instituted”. You know, when Being-Addicted-To Frivolity & Nonsense has become the reality.
You and ALL “intellectuals” know, of course, how “The-Master” keeps growing “Annually”. How Others/”Dem-Smelly-Lot” are kept waiting for “Technology-to-Intrude”. You know, kept waiting for when, “Energy is Free[/d]”, so-to-speak. And what about, “The-Frustration” of Others”? More so when “Others” are your Real Brothers & Sisters.
Unfortunately [for The “Biased”/Mesmerised], there is NO Real-Difference between Dem-Blue-Boyz/”Past”/Fear und Dem-Reddites /”Kommies”/”Future”/Greed. This is because BOTH “respond similarly to, “More Money Than Sense”, Running-Amok. You know, when Fear-Greed are Running-Amok.
So whether ,“self is some Statistician Running-Amok”, or merely a “Scientifico” who “most-sincerely” believes in, “The-Goodness-of-ManKind” [beecuz “Virtual-Reality Rulez, AOK”], whenever Energy-Transformers are able-to become, Energy-Creators [of sorts], Dem-Blue-Boyz and their Red-Rivals are merely, “The Different Faces of The Same Coin”.
This, then, means that The-Vast-Majority have been “materialised”. You know, when “Materialism” is The Be-All, End-All. You know, when, “Materialism or Bust”/WhipSawing is, “How-the-Game-is-Played”. When ALL “Born”, have to, somehow but don’t know how, “Die”, that is akin-to being forced-to, “Face-Reality”. When, “The Freedom of Thought-Processing" [TP], is taken for granted.
ALL, without exception, were “Thoughtless”/The-Embodiment-of-Innocence/”Children” once. The Era when there is NO Past-Future other than existing “Presently”/Joyfully. Like “Animals”/The-Innocents do. But once The-Genie that is TP/”The-Anti-Christ” had been “Acquired”/Institutionalised, that Genie cannot be “Bottled-Again”. NOT-by The “3-Dimensionally”/Materialism Limited Humanity who are able-to perceive ONLY “The Right-Side”. That of “Somaticism”/Physicality-Intellectualism.
“Oil” came originally/”invisibly” from The-Sun, said Sun’s Energy Output to The-Earth for “One Day” being enough to “Supply-The-Energy-Requirement” of The-Earth for approximately One-Year. But what of the excess energy for the remainder 364 days? No wonder “Global-Warming” is about promoting, “More Money Than Sense”. That of promoting The-Genie called Fear-Greed to Run-Amok. You know, when, “Technology-or-Bust”/Hypocrisy Rulez, AOK.
Evolution exists because we were ALL once, “Lifeless-Rocks”. Like it or not, “Thoughtfully-Processing Humans”, are NOT The Be-ALL, End-All to Evolution. This is because Final-Evolution is, “Returning”/Reverting To-Thoughtlessness. A State Attainable after, “One’s Very Own Reality”, had been Realised/”Faced”.
It might, “Benefit-Absolutely” to finally realise that without Admitting-To/”Facing”, “One’s Very Own Reality”, there is NO Reality. You know, when Reality is actually “Relativity Running-Amok”. No wonder “Democracy” is actually “DemonKrazy” Running-Amok. To-Promoting Fear-Greed Running-Amok. You know, when, “Booming & Busting”/Re-Cycling, is The-Rationale for Existing.
It will also “Benefit-Absolutely” to be able-to realise that “3-Dimensionally”/Materially Limited-Humanity gravitates from Cunningness/Licentiousness/Abandonment/”Invisibility” to Cleverness/Ritualism/Control-Freakishness/”Tangibility”. The-Moment when, “Aggression”, is seen-as ”Being-Intelligent”. But NOT when, “There Is Always Something After-The-Fact”. You know, when Knowledge is merely Of-Relativism, The ModernDay-Affliction being The-InAbility to absolutely separate Fact/”Benevolence” from Fiction/”Malevolence”.
When there is NO Reality to,“Knowing-So”, it is because Reality is Beyond knowing how-to, “Bake A-Loaf-of-Bread”/Fool-The-Fool-ables. When, “Fooling-The-Fool-able”, is merely Cunnning Running-Amok, it is because Reality is when there is NO “Other”/Cunningness-Cleverness.
Like it or NOT, A-Moment will arrive when one’s MisDeeds will-be,”Called-To-Account”. “Just because Another is “Blind-As-A-Bat”, does NOT mean that those who are able to-perceive should be, “Similarly-Flawed”.
DO NOT “play”/indulge-in “The-Blame-Game” of Feeling Guilty-Angry. Just Don’t “Do-It”/Be-Baited/”Act-ReAct-Relatively”. Just because “Dem-Reddites” are having “A Whale-of-A-Time”, does NOT mean that Dem-Blue-WunderKind did NOT had theirs. More so when Fear is merely The “Past”/Source of Greed.
There is NO Real Difference between The-Blue-Boyz und Dem-Reddites for the simple reason that BOTH react similarly when, “More Money Than Sense”, is Running-Amok. You know, when Sell,Sell,Sell is for promoting Buy,Buy,Buy - and Lttle Else. Let The Blind-Blinded/”Religious” Ones See What They Like. So long as self is able-to see Reality, that is all that matters for promoting Reality.
Always the Gold Standard of Reviews! Sir!
So thats why they were so quiet about overclocking at computex and E3
Wow really, did your 2600x boost to 4.2 all core? No, then why the fuck did you expect this to get to 4.4?
my 2600 does 4,1 all core easily. with better cooling and that x-chip 4.3 should be realistic
@@DerKek89 People were expecting 4.5ghz because of that idiot Adored yet it barely makes 4.3-4.4
I'd wait for Hardware Unboxed review before jumping to conclusions.
@@m.streicher8286 I'd expect it to hit single core max frequency they advertise consistently. This is saying none of the chips even on a single core hit their max advertised speed.
Which AGESA did your UEFI have when testing though? 1.0.0.3AB is needed for proper overclocking.
he is on x570
LG Nilsson yeah, but maybe he has it early or maybe he doesn’t and he done goofed, I will wait for deeper reviews
They've been saying forever that x570 has been a big headache and will take some time to get ironed out
And exactly, why should Jay be a clown? Just because he's der8auer it doesn't mean he can't be wrong.
@@UptownNYCso true👍
For those watching this after 7/9, turns out this guy was wrong. Its a bios issue, check GamersNexus to see them hitting the advertised clock speeds and SUSTAINING the boost speeds as advertised once the bios teething issues were worked out. *Meanwhile this guy assumed day 1 that the chips were defective and that none of them would ever hit their advertised speeds. He was completely wrong* Really sloppy not taking this video down after its been clarified....
Awesome video, it confirms everything I noticed with my 3700X, and it saved me from the hassle of delidding.
Thermal dispersion on this chips is awful, when CPU runs over 90°C in stress the cooler stays almost cold, it is a problem mentioned by hundreds of people.
I ll avoid delidding but I may have to lap the IHS cause it doesn't look absolutely straight, there seems to be a hump on the left side and thermal paste is not spread well on the other.
I'm a simple man:I see der8auer's new video,I watch it immediately.
Isn't there some kind of bios limitation that stops or at least hinder's clock speeds?
btw boost and all core OC are not the same thing. there are a few reviews that has 3900x hitting 4.6ghz on 2 cores at a time which switches to a lower frequency and another core will boost to that speed while the others rest when you are just letting it run stock.
i manually overclocked mine to 4.525 on all cores with an aio an max temps while playing games is around 60 cel
I am so glad that you and the other tech tubers are here to dispel myths about this stuff. I have to say I would have been pretty disappointed when I get my 3000 series chip and it doesn’t boost as high as the box said it would.
Thank you for this and your continued gifts to the community!
This video is sponsored by Coca-Cola! Buy one today. :p
Will do! Anything for my boy Roman!
Who doesn't have coca-cola in their house lol
@@sanjiyoung1433 me
I have a sudden urge to drink Coca-Cola Zero ©® now!
@@sanjiyoung1433 I really like Sprite but not that Coca-Cola.
honestly i like that automatic oc shit cus i am not that big on oc, except when i get a big difference
These were exactly the questions I had. All hail der8auer.
Der Bauer: Chief Professor of Silicon Surgery
*AMD: "wait, that's illegal"*
The overclock headroom is so disappointing.
And that with 7nm
ryzen 4000 will be the real killer i guess
@@MyAsdfqwe Can't keep kicking the can down the road haha. At this point, have to accept that Ryzen simply gain a ton of clock per iteration, but that's why there was so much emphasis on the "IPC uplift". The 3600X clocking lower than the 8 and 12-cores is a bit disappointing though, as I never spend more than $199 on a CPU.
@@terp02Andrew yeah it seems r5 3600 is best for buck. still, this means 4.3ghz ryzen 3000 is same as 9900k at 4.6ghz. the gap is getting pretty close.
@@MyAsdfqwe That's what I'm thinking too. Ryzen 4000 will probably be to the 3000 series what 2000 was to the 1000 series.
THANK YOU for this video and all the hard work you do with TWO languages!
Roman, amazing work as usual friend !!
Will we get Direct Die frame for Ryzen 3000?
yeh i want to see him over clock with with a rough made direct die frame and see what it does to temps
@TPBR11 this is why you shouldn't make overclocking videos before BIOS's are mature enough… Like Linus, JayzTwoCents and so on...
When will we see a Der8auer OC frame for Ryzen 3000?
not worth it unfortunately :D
@@der8auer dielapp and lm :D?
Noooo please de8auer I need this oc frame for my 3900x
@@der8auer Please, we need Ryzen 3000 Direct Die Frame
Max single core clock. I agree it's dodgy advertising but the max single core clocks are going to be higher than all core clocks due to the temperature. Which in a way makes sense. Why would you want all 16 cores at 5ghz anyway? You only need the cores you are using for your game to be as high as possible. The other 10 cores can go slower. Honestly if you want more speed, wait for Ryzen 4000 next year.
Im wondering about max s1-2 core clock also....
Top-class content from Roman, as always.
Rumor: Lisa Su is a God.
NDA has been lifted.
Fact: she's only a Demi God.
der8auer
s video turned out to be invalid because of vendor bios bugs which didn't allow the boards to boost properly. He should really delete the videos as they are flat out misinformation at this point.
@@UKgamer87 it's not completely incorrect. He can always just add an annotation or upload another video correcting that small part.
I love being in the east. Get to see the videos first!
Never, .... ever..., overpromise and underdeliver. Do it the other way around and customers will love you. Marketing depts never seem to get this.
Basic info ? This is great ! All the AIO mfg are cringing that extreme cooling only is making a minimal delta with the new Ryzen 3000 release . Also you rock to lead that AMD was actually documenting boost clock that isnt being acheived in testing .. so we had good CONTENT even before you shared the delidded results
9700k,,,under watercooling...king of gaming......temps....fps...facts......im saying gaming only....$364 king....!!!
5.2 Ghz on air! a beast!
amazing video !!! thank you. There is any way to overclock only a few cores? Is the Ryzen 3900x ECC compatible?
For the temp offset this is where something like your graphite sheet TIM would really help as it should be great at laterally transferring heat!
BTW, to remove the Indium layer on the cpu, you can use liquid metal(like galinstan metal, or conductonaut if you're rich) to soften it up, then it's a lot easier to get rid of them.
And don't spill liquid metal on solder joints, they can be soften up too.
@der8auer, no direct die tests? Also, it would be interesting to see carbonaut installed on die.
They are soldered arent they?
@@fullyverified7491 Who?
@@rusTORK The Ryzen CPUs. And yes they are all soldered, except the 2000 series APUs I think.
@@DenGuleBalje I know it. We saw it in this video.
Can you OC 3000 CPUs with power limit method? Keeping idle clocks and increasing boost clocks
It doesn't work like that, Ryzens throttle clock based on temperature, they almost never hit the power limit and definitely never if PBO is enabled.
Now, the interesting bit would be how much undervolting is possible - and if it actually helps performance by lowering temperatures.
@@AstralS7orm He means overclocking by P-states where you set your "boost" clock and lower state clocks yourself and the core will automatically throttle up/down based on load the way a non-overclocked CPU will, so you're not permanently running the chip at max the way you would using other methods of BIOS overclocking. Worked great on Ryzen 1, haven't seen anyone doing this/talking about it for the 3k series yet. Also all BIOS versions are very new/immature on these new chips so I really wouldn't worry about what results people are getting at this point. Wait a few months and things will be much more stable.
Fascinating. So, why use a lid at all? Use a cooler (probably make a cooler) that makes direct contact with the two chiplets. Remove material from the face of a waterblock leaving slightly oversized rectangular bumps from the un-machined areas of the face. Or, use a lid that has its solid face removed (mill out the face from the pocket side) and use this as a standoff for the face of the cooling block. Or make a standoff similar to the lid but with no center, made from non-conducting material, like Macor. Your findings make my head swim. AAAnd late to the party again.
Proud owner of a 1600af @ 4.35Ghz -3466Mhz Crucial Ballistix. For 100 bucks, it's the cheapest and fastest CPU I've ever bought myself!
Pepsi > Coke
So the chances are that the 9900k will still beat these chips at purely gaming benchmarks at enthusiast level.. Shame as i was thinking about upgrading my 9700k to 3900x. Not long to wait for gaming benchmarks now!
if you plan to game on 1080p then yes, 9900k is still better. But think about how much more you get with 3900x. It's 50 percent more powerful with less power consumption.
Nobody really expected that!!! Intel led both on speed and the ipc before this launch.. Haven't you looked at any 9900k vs 2700 benchmarks? Not in the same league at all. The only thing 3000's series got going for it is core count and cheaper than Intel..
8700k and even the 8600k will probably still be better in gaming
@@sparklauerei1 yeah, it's a fail if they can't even beat chips 2 generations old.........
U were wrong
Excellent content. Excellent presentation. A real pro.
what an absolutely beautiful machine. Great job AMD .
so is a disappointment i was at least 4.5 on all cores im sad now
Like always, AMD lost again.
4.3ghz on all core for ryzen 3000 means 4.3 + 4.3x15% = 4.975ghz on ryzen 2700x. Which is pretty solid. If you can overclock to 4.2ghz then still it's a great CPU. Especially, considering the price.
@@madd5 But ryzen 2 is slower you can't really compare it that way. Ryzen 2 is about 10-15 percent slower than the 9900k
Who cares about mhz in 2000k19? You need to look at ipc first. Its doest matter that your cpu cant reach 5ghz when it has better perfomance :)
I don't care at all on a conscious level, but even if I had any regret in my subconscious buying a 2700X ahead of time it completely and utterly evaporated into absolute nothingness watching this video.
Herzlichsten Dank!
Just bought 2600x cant wait for it to arrive. Always buying the newest part is not a very cost effective way to buy tech product.
Im running a 3900X on an
asus Prime X570-P. Leaving them stock with a good cooler works out now. The recent bios update out for mother boards allows my chip to auto boost to 4.6Ghz under load(cinebench r15 208 sc , 3254 mc, 177.48 fps)( cpu-z bench 548.6 sc , 8515.9 mc)
10 core OC @ 4.3Ghz w/2 cores (C01 & C04)@ 4.575Ghz running 1.3v
(cpuz 565.2 sc , 8419.3 mc)
Thank you very much
I won't say that AMD is letting down any customer, just because some idiot cannot think.
They are not false advertising, and clock is not even all the story in a CPU performance gain.
No but you cant gain anymore performance than advertised basically. While 9900k 5ghz for most which does increase benchmarks and tests quite a bit
@@dnsean03 So?
Anyway, you are not guaranteed on Intel either.
@@IIARROWS moral of the story if you want guaranteed ocs go get your CPU from siliconlottery.com
@@dnsean03 So? Do you use your CPU for reaching maximum clock, or to make any kind of work?
Because data show performance are higher even with lower clock.
The clock doesn't measure performance across different architectures.
@@IIARROWS mix of both best stable overclock at decent temps for performance. Always try and get most out of your purchase. Theres more overhead than AMD wont go over 4.3 mhz simple as that
So you removed indium and replaced it by indium+galium+stannum alias GALINSTAN. So interesting is that galinstan do better heat conductivity than indium alone. :)
Indium gallium zinc oxide
@@tazboy1934 Thak you,. good to know this.
Darth Fate it is actually the thickness of liquid metal. It is a lot thinner than indium solder.
@@mtunayucer Indeed,.. Surface tension of liquid metal is that most important thing.
I can always look to der8auer to give it to me straight.
How have I not been subbed before. Thanks for this.
So basically Ryzen 3600x is the exactly same as 3600? Only 100 MHz higher for $50? Wtf
Really that's it?
@@magzire yes, it was even advertise in similar frequencies by amd, the difference in bench marks is the thread count differences with the extra 100mhz. as well as the extra cache it is getting. thus why the 3900x which is only advertised 200 mhz higher then for 170 dollar price difference. it gets 8 more threads and 32mb more cache size. thus why on bench marks even though the frequencies are very similar that the 3900x is dominating on benchmarks. same with the 3700x vs the 3600x , the 3700x is only base 3.6 ghz compared to the 3600x base of 3.8 ghz, but with matching 4.4ghz frequency, same frequencies with a 80 dollar price point difference. BUT....3700x gets 4 more threads, and 2 more cores, 1mb more of l2 cache and 30watts lower tdp, so for 80 bucks more for the 3700x you are getting more performance in games with the more cores, getting more real life scenario performance with multi tasking with more threads thus why it matches the 9900k on cinebench. and the biggest selling point is the lower tdp 65 watts. for better performance then the 3600x in gaming, and similar performance as the 9900k in multitasking scenarios, so if you want to game but also do other things like stream, watch movies, do video editing, the 3700x is price perfectly for 160 dollars less then the 9900k. If all you care about is raw single core frequencies or max frequencies go intell, but if you want a chip for the long hall that will do better once companies like msi release better boards in 6months to a year for better chip cooling on board. then 3rd gen ryzen is the way to go, these are maxed boost clocks, not over clocks, you are seeing in his video. drivers and bios are also an issue still like when they first launched ryzen in the first place. you will see in a couple months of stable 4.8 overclocks 3700x.
@@magzire Well yeah since it's stuck at 4.3 while 3600 non-X gets to 4.2; however I'd hope der8auer will answer my question so everything is clear
Just like the 2600x vs 2600
@@Uria-jz1gu ruclips.net/video/7AbNeht4tAE/видео.html theres your full answer in the first minute of the review.
When will AMD learn to underpromise and overdeliver?
Same thing happened when AMD launched R9 290 cards. They claimed max 1 GHz core clock witch never happened (later custom cards achieved this) and people were disappointed. While nVidia used boost clock witch was usually supposed while gaming.
Same thing now with these xfr boost frequencies.
Good point in a marketing perspective, it's bad for business to do what they do. The products themselves seems to be brilliant though, I'm probably getting a 3700X myself.
What? My reference board 290X did 1200mhz on the core and 1500mhz on the mem, with a Raijintek Morpheus on it, I saw quite alot of people getting their cards to 1300+...if your 290 couldn't do 1ghz on the core, it was a dud card.
@@glynkatkin I was talking about advertised 1 GHz core frequency of a reference R9 290X, witch never happened in real world. You are giving an example of a modified and overclocked 290X, witch is completely irrelevant to my original point.
I wouldn't say irrelevant - we don't rock stock coolers on CPUs if we want to push them to the limit, nor should we expect to use stock GPU coolers if we want to push them to the limit. :-)
But I do agree that it is much better to under-promise and over-deliver rather than the opposite.
The most recent example is the 2060 SUPER SAIYAN HD XTREME which promises a boost clock of 1.7 Ghz but actually just clocks itself up to 2 Ghz for shits and giggles.
nvidia, good at marketing, whodda thunk it?
The line-up is awesome. Literally everyone has a choice now, on every budget, regardless of what they do. Their marketting strategy is what's ruining it, sadly. They always did this and always. ALWAYS. failed to meet their own promises.
More honesty would bring them way further, imo. Say it like it is! "Workstation CPUs that also game". They already rule the multi-threading/cost/performance market as intel production CPUs cost a shitload by comparison. Why lie about gaming when you know it ain't true? I believe this is why they set-up the embargo on benchmarks, cause it contradicts their marketting "new gaming king" bullshit by a large margin.
They basically pulled an Intel on us, by giving marginal improvement over the former series(which intel gets constantly bashed for), at higher prices, while promising total dominance in aspects where they know they fail.
Also, what's the purpose of expensive motherboards with lots of VRMs if they can barely OC? 😔
I'm sad for AMD. This kind of tactics are never good. Intel and nvidia can afford it, not AMD. Hell, if both stopped selling any products they could still function for like 30 more years from the revenue they got in the last 5 years. If AMD stops selling they'll close shop in 2 years max.
Makes me REALLY proud of the 3600x I got. I had it to 4,4ghz all core with a 1.25v undervikt an a TUF x570. Sold it to my friend, but it's running strong with him.
Might I recomend something instead of that steal blade? there is a company that makes ceramic blades that are edull enough to not cut yourself but still sharp enough to cut through softer stuff like tape and would be good for your application
lol thanks for the early video roman , trust 'ze germans' to be efficiant and prompt to bring us news
Ahh yes, die Deutscher shaffen das.
unsettledroell schaffen* have to do that as a german grammar nazi
unsettledroell schaffen* have to do that as a german grammar nazi
@@checkcommentsfirst3335 ok, ben
Gott schütze den Kaiser !
Great video Der8aur! :)
I actually thought Ryzen would really be a game changer in pure gaming and OC, but doesn’t seem worth switching out my 370 And 8700k now...
Yes, indium has better heat transfer than LM. Would a direct-die solution been better? I'm thinkin of delidding my 3950X and going direct-die route with liquid metal.
I've just delidded my 5950x for direct die. Modified the plastic latch cover by sanding it so my waterblock makes perfect contact. I believe direct die will be an improvement since the block will directly be sitting on the dies
Your starting report, and conclusion were excellent. I'll still get a 3rd Gen Ryzen, but with realistic expectations 😎👍
Well that was unexpected, if I though the CPU would just clock like 2nd gen up to advertised speeds from time to time when needed. But if they really can't even hit 4.7, 4.6ghz whats the point of the advert.
That's not oc he said that's the boost, overclock said 4.3 max
You can start memeing about how they borrowed Intel's 28 CORE 5 GHZ !11!!1 chiller to determine their max boost clocks. ;^)
ZeddBloaxigan kinda true, that sub-zero scaling kinda makes me want one! (I run sub-zero phase for daily machine, I really want a massive one for TR 64 core, unlimited power! Mwahaha!)
The outrageous thing is that the 64 core would probably run fine on high-end x570 board VRMs - as long as you stay not too far above the power sweetspot - as opposed to the fucking 1.5k dollaroo monstrosity boards the infamous 28-core wants for supper.
But hey, the whispers of Zen 3 being even more power efficient are already going around.
Ryzen 4000 will do 5.0 Ghz.
Doubt it
@@filiphedman4392 are you saying amd still cant do 5g ?
Even if this wasn't a joke, it still made laugh!
@@lukevega77 But sad that Intel 5 ghz cpus are not better than 3900x on stock LMAO
@@cemoruc8810 amd already did 5ghz with fx
Again, thank you for the English version. It realy helps! Greetings from Austria
This is called driver maturity
AMD is going to get sued again over falsely advertising clock speed. Just like they did with cores on the FX series.
Haha thank you I made a random bet saying it wouldnt go over 4.3ghz and I was correct
I was really skeptical about overclocking potential since the early 5ghz hype days but I thought it would do 4.4ghz all cores easily LOL
@@madd5 I mean still decent cpus I'd think but Intel got more overhead which helps bench more
i got trollled badly when i predicted 4.4 to be the max OC in the days of 5ghz hype
You may want to hold on to that bet. Several youtubers have had this same issue. Much like radeon 7 not being overclockable days after release , then patch fixed it . Now , it overlocks predictably.
If your assumption is correct about the hot spot, then, what if you can use a air cooler and measure temp difference between the heat pipes and see what you can find out.
that max overclock is for single core boost only. says it on their website. I thought everyone knew how boost on cpu's works now a days.