AMD Ryzen Precision Boost Overdrive & AutoOC Benchmarks & Explanation
HTML-код
- Опубликовано: 4 окт 2024
- This deep-dive explains PBO vs. Precision Boost 2, including benchmarks with test data on differences between stock, PBO, and AutoOC.
Support GN via the store! store.gamersne...
Article: www.gamersnexu...
In this video, we're exploring Precision Boost Overdrive, Precision Boost 2 / XFR, and AutoOC features on the AMD Ryzen 3000 series CPUs. The first half of the video explains the differences between PB2 and PBO, talks about what Precision Boost Overdrive is supposed to do, what features are technically part of (and not part of) PBO, and how the feature is theoretically supposed to be used. The second half talks benchmark data, differences, and about whether it's "worth it" to overclock via Precision Boost Overdrive. Testing uses the AMD Ryzen 9 3900X and AMD Ryzen 5 3600, with the MSI X570 Godlike and Gigabyte X570 Master boards used as the primary test platforms.
NEW RYZEN CPU LISTINGS:
AMD Ryzen 5 3600 (Amazon): geni.us/vtPUE
AMD Ryzen 5 3600X (Amazon): geni.us/uxr8
AMD Ryzen 9 3900X (Amazon): geni.us/uPxxADC
AMD Ryzen 7 3800X (Amazon): geni.us/kHxcq
AMD Ryzen 7 3700X (Amazon): geni.us/sdVDHO
AMD Ryzen 5 3400G (Amazon): geni.us/IWN0c
AMD Ryzen 3 3200G (Amazon): geni.us/AxJmDW
Or grab an affordable R7 2700X on Amazon: geni.us/HPeV or R7 2700 geni.us/CHQ0B
Intel's i9-9900K (Amazon): geni.us/eNRPu
Like our content? Please consider becoming our Patron to support us: / gamersnexus
** Please like, comment, and subscribe for more! **
Follow us in these locations for more gaming and hardware updates:
t: / gamersnexus
f: / gamersnexus
w: www.gamersnexus...
Host, Test Lead: Steve Burke
Editorial, Testing: Patrick Lathan
Video: Andrew Coleman, Keegan Gallick
Links to Amazon and Newegg are typically monetized on our channel (affiliate links) and may return a commission of sales to us from the retailer. This is unrelated to the product manufacturer. Any advertisements or sponsorships are disclosed within the video ("this video is brought to you by") and above the fold in the description. We do not ever produce paid content or "sponsored content" (meaning that the content is our idea and is not funded externally aside from whatever ad placement is in the beginning) and we do not ever charge manufacturers for coverage. - Игры
This was a massive research effort undertaken by Patrick on the team. He did a great job putting it all together!
Article: www.gamersnexus.net/guides/3491-explaining-precision-boost-overdrive-benchmarks-auto-oc
Check out our Best Boards for Ryzen 3000 video here: ruclips.net/video/zuyuS04lD4o/видео.html
Hello guys I need help because I have a seasonic focus gold 650 watts and the motherboard I am looking to purchase requires an 8 pin and 4 pin for the cpu.It only includes an 8 pin cpu power cable,but the motherboard that I want to buy requires an additional 4 pin.My problem is that im not sure if I can use 4 pins from a split 8 pin to power the 4 pin on the motherboard because I cannot find any 4 pin to 4 pin cable that is long enough?any suggestions
Gamers Nexus give that man a raise, this was a super confusing topic and I feel informed now
ok, I'm lost here. I understand what you are talking about in this video but in your charts there seemed to be no difference between the liquid cooled results and the air results but you are saying a better cooler will give more performance. Which is it? I want to get a 3600 but do I need a liquid cooler if I'm not going extreme overclocking or will the box cooler be fine? I understand just leave PBO and auto OC alone because it really doesn't do much.
Damn AMD.. Y U make OC complicated!?
Did you talk to Buildzoid about PBO? On Zen+ he found, that with PBO the processor is "overvolted" by the MoBo. Setting a negative Voltage Offset and Adjusting the Loadline Calibration gave much better results than auto. ruclips.net/video/YQ8zdprzEjI/видео.html
Timestamps:
00:15 Snowflake knocked an Intel CPU down
10:47 Snowflake playing with an wire held by Steve
The important stuff
The real reason I watch these videos.
Thanks Keith, I made sure I watched the Important parts of the video,
apparently there was something about AMD CPU stuff Blah Blah AMD Blah Blah CPU Blah Blah PBO Blah Blah OOh! SnowFlake 😺😍 Blah Blah more AMD
This is why filters are handy.
stupid cats.
The feeling when a cat gets a R9 3900x before you do
**Senior AMD Analyst Snowflake
I got one but I'm waiting for a mobo now 😑
Pre order on Best buy or I think shop blt was getting some in stock on the 22nd and were taking pre orders. Shop blt had a few thousand coming on that shipment since they had a page showing future availability and how many were reserved. I think Best buy is getting more in a similar time frame.
猫に小判
There's no stock anywhere on the planet, cat doesn't realise how lucky he is to see one.
Pretty sure I'm even more confused about Ryzen features after watching the video. Lol
enable everything, and you might as well get 1-4% performance. Sometimes.
Don't touch anything. Just overclock your memory
@@lifemocker85 touch everything!!! get afucking headache benchmarking, touch everything again. get an annurismy cos your points fell, reset Cmos, THEN just oc memory* that's the enthusiast way
been there, done that....
Just get the best from your memory and leave the rest alone is my take.
@@TheHighborn you couldnt be more right hahah
He talks so systematically and concisely and my brain still overflows. I have to watch this like 10 times.
Because gamers nexus is NOT a good channel.
Call your doctor if your boost clock speed last longer then 4 hours.
I did, he said “lucky bas***d, no need to brag”.
I heard walking up stairs helps...
He'll give you a prescription to run prime95 for two days and call him in the morning.
@@1blackice1 charge you $80 co-pay
Boost cock speed
PBO is dependant on the alignment of the planets, lol.
lmfao thats a great comment :)
now that the planets have realigned since testing, time to test again!!
Ah shit we missed the planets alignment guys, time to wait 1000000y to test it
the fortnite game and its RNG bullshit challenges depend on that too
It's the dawning of the age of Aquarius
AMD: _"We'll do all the overclocking for you so you don't even have to touch anything. We're keeping it simple - just play straight out the box!"_
Overclockers: *_"WE HAVE TO BE SURE."_*
Liquid sunshine is fucking dead
@@mrlolas5114 wait what
@@liquidsunshine697 ion know man edited comment and 1 year old
@@mrlolas5114 just random it said I was dead
So happy to see you tackle this. I need a trusted source to inform me on this.
Still wondering why Infinity Fabric OCing is not being discussed. If set on manual at 1:1 (1800 IF / 3600 MEM FREQ) there's an apparent jump in performance compared to OCing with PBO/Auto OC. I hope to see you guys tackle this!
Are you talking about just overclocking the infinity fabric and memory or also overclocking the cpu after the infinity fabric and memory are overclocked as well?
@@Vortex1988 I'm just talking about setting the infinity fabric to 1800 manually (in auto it fluctuates and isn't set to 1:1 which is the best setting for Ryzon 3xxx) )and setting your memory speed to 3600 Mhz. This apparently shows more gains than doing PBO+Auto OC.
What if the Infinity Fabric clock is higher than the memory clock? Say 1800 IF with 3200 Mhz memory
@@pkpnyt4711 I could see that. I'm just curious if that would have any additional effect on cpu overclocking if you combined all three.
@@Lemard77 Then the ratio is not 1:1 and the system performs slower. AMD even said that the best ratio is 1:1. I would probably set the IF to 1600 if RAM is 3200. That would probably get better performance than leaving it in auto. Though the fastest performance would be in 1800/3600.
Just found this channel. I did 30% of this on my own and suffered a great deal and just settled with Mem OC and sometimes enabled PBO on certain days/use cases. This guy is amazing. Very detailed and precise. I can now make even more educated decisions in my motherboard/CPU/Perf adventures.
Thank you so much for spending this huge amount of time and resources to reveal the bluff!
Tried watching this three times!
One day I'll make it to the end and understand.
Thanks for uploading.
Steve, man. This video is amazing! extremely clear, I love how deep you go with some stuff. Congrats to the team of this channel, keep this great content
Now I'm really curious what ACTUALLY limits Ryzen 3000 from running an all-core overclock to the rated single core boost clock high end overclock situations? Obviously all the transistors are capable of running at 4.5, 4.6 GHz for at least short periods. But have we got to a point with 7nm that leakage current just runs away? Is the node so small that micro-heating of small parts of the chip exceed silicon's thermal conductivity to the point that only time (milliseconds of boost and then clock dithering) or extreme cooling (LN2) can allow the switching frequency to maintain top clocks? I'm not knocking AMD for having a product that clocks in a different manner than what enthusiasts are used to. We can be a stubborn bunch.
I'm not a physicist or an electro-chemist, but my theories above about "micro" thermals are based on the huge jumps in power consumption on LN2 with Ryzen 3000. Tech reviewers and semiconductor analysts probably need to start looking at Watts/sq mm more closely when looking at products like Zen 2 and Navi to get an understanding of how manufacturers are handling clocks and power and thermal management.
Interesting... Fact- temp certainly varies effecting something of a certain size and density and material;
Therefore I find your statement inherently, with respect to what I just stated, is most certainly correct and therefore 100% applicable- and in more than just the CPU.
HMMmm...perhaps a relative test on another earlier CPU that has a size difference but same materials would likely tell us how then the respective sizes react to equal testing thus revealing the answer to size affecting the temperature.
Good eh? Yes I'm an engineer- [edit; another test on earlier CPU that used a DIFFERENT material would result in definitive resistance data with respect to the materials used in todays chips however I believe most people can, by simply deduction and extrapolation, come up w with many answers and/or ways to get them... Testing ones hypothesis. And that's exactly a lot of what GN and some others constantly do.
Better for GAMERS NEXUS to simply say it's basically next to nil then sugar coat his concluding statement for amd.(In NOT SUGAR COATING, Steve would have also shown mindful strength, and so deserved[as Steve is proving he will go further now into what the hell comoanies are trying to sell versus what they claim their product is and can do!]:
- AMD should know better than to release such garbage to the public. Whoever allowed that to pass I would shit-can immediately, and release a statement in fact and well delivered with meaningful "props" to regain some face. Otherwise it seems to me by AMD letting that go, thinks it's customers are UNEDUCATED, AND STUPID. And frankly that should bother everyone. If you pay attention; you'll see some disdain in Steve's face talking about that clip by AMD and some clown pointing at nonsense- that's desi?however NOT having the scrotal fortitude to NOT SUGAR COAT his response to that simply weakens himself, and the fact is; Steve is more influential than he thinks. Everyone knows who GAMERS NEXUS is... and his face and that HAIR- forget about it. What he says can cause thousands upon thousands to not buy a product, or buy it with great haste too.
Very interesting and on depth testing to reveal the bozo in that amd promotion or whatever pathetic thing one wishes to call it- I knew by simply watching the manner plus sorry scribbling, with verbage that what the bozo was saying didn't add up to rally ANYTHING...wth is/was THAT? And they actually paid to show that crap. Most annoying.
That all said, Again-good job GN
and let ZERO GO even off the table with parts- everyone sees this and it simply adds to your professionalism.
There's NOTHING unprofessional calling things out AND PEOPLE when anything exists to call.
PERIOD. AMEN.
6:39 As an AMD enthusiast, I'm not too familiar with Intel's tech, but it's interesting to see T.J. Maxx stepping into the CPU market. With their strong track record for delivering great value and performance, they have the potential to shake up the industry and bring some significant benefits to consumers. Exciting times ahead for the tech world!
This video is a Rosetta Stone for Ryzen, good job.
Came in, just jumped to the conclusion, then before leaving decided to watch it all because I know how much effort you guys put into making these videos, and I know the fidelity of your words. Love from Portugal
Never forget the alignment of the planets in your overclocking
You guys went above and beyond with the testing. Very informative. I might just leave my 3600x @ stock.
The cat: I'm smelling some marketing BS on this processor lol!
Thanks for the clarification, Tech Jesus.
Gamers Nexus Steve, the kind of guy to have easily discovered the cure for cancer had he chosen the medical field instead of hardware
Like seriously we need to applaud this man and his insane research efforts (and the other ppl at Gamers Nexus too!)
TDC: top dead center, good knowledge for when you need to retime your engine when tuning it, as to much advanced timing is key to cause of detonation, not good.
Lol!!
Changed a timing belt on my car and the amount of times I worried about TDC
However in 1964 the correct timing was 4 degrees before top dead center.
That's why I run 105 octane fuel in my engine! Got to stay away from detonation! Ping ping ping
It is also interesting to note that when you turn on PBO on that the 1% and 0.1% lows get worse then when you leave it off.
testing it right now and I'm getting much better performance using manual OC than PBO
I have done hours and hours of testing with this myself. Using a 3700x, 3900x and 3950x on my Asus X570-e. I would like to chat with you on this. I watch all you videos. I have a background in electronics and love your vids. Keep up the good work.
I watched the entire 36:19 of this video and all I have to say is.... Ive no clue what anything you said, meant little kid.. but rest assured, those 36minutes and 19seconds touched my heart.
I would love to see a revisit of PBO given the recent AGESA updates
These new 3000 chips really get my Precision Boost going
Just what I've been looking for. Appreciate the time taken! Basically leave it the hell alone and enjoy your 3900x at 4.2ghz when the marketing heavily pointed towards a 4.6 boost :)
So in a nutshell.
1. Get a good cooler
2. Make sure VRMs don't explode (on 3900X)
@Anjellica Kraychek no 5, why ??
@Anjellica Kraychek Saturn? Lol
Omg thank you, I’m a noob and this is exactly what I needed to know before getting my 3900x.
just overclock your memory
lol
just do not play with CPU voltage or your 3900x will die if your noob
warlock TV yea, exactly why I’m excited about using PB/PBO. Would like to not fry anything 😂
For most people the test on a $200 board would be interesting.
Depends. Most Gigabyte X570s have the same VRMs, which are fairly powerful and capable.
It's nice and appreciative to see you doing all this work to find out just how you can push the CPU further. It's a shame that AMD have obfuscated the names and explanations because it only causes confusion and misunderstanding and then AMD takes a hit in the aftermath
Kudos for doing this extremely time consuming but very informative test.
Modern hardware is getting so much more tricky to nail down firmly with all these flexible boost goalpoals that can show so much variance. Hopefully there is still some room for AMD firmware in the months ahead, but at least for now we know for the CPU end it is a "Assemble rig with best cooling u can afford and you're good" without the hours of OC tuning and stability testing.
Zero is a valid result, the effort is appreciated.
I would really like to see a video on DDR4 comparisons with Ryzen 3000. Like, how much is the difference between 3200c16, 3200c14, 3600c16, 3733c16 and 4000mhz+. I know there is the whole infinity fabric downclocking at 3733mhz thing but how much does it affect performance and are there scenarios where running something like 4400mhz DDR4 is beneficial even with infinity fabric speed halved. Love the videos Steve, keep it up.
3200-3333-3466 CL16... CL14 and 3600 kits are damn expensive
4000+ MHz kits are a joke. Shit timings and stupidly high cost. Just get some cheap B-die and overclock it to 3600 MHz CL14
@@Amusia727 bdies aren't cheap
@@lifemocker85 They are if you find a good deal. Look in the used market
We need some memory comparison is there much difference. Found some review which said its only few fps 3000 vs 3600
I've found some reddit post where one of the users posted a question for Robert Hallock about boost clocks and he asked if voltage setting is on auto or normal and if on auto to set it on normal. User posted afterwards that his cpu now reaches higher clocks. Nice video and explanation, thanks. The board in question was Gigabyte X570 Aorus Pro, the cpu was R9 3900x
Actively messing with this rn now and I understand everything your saying but could you please do a new video for Ryzen 5000? I have a 5800x and found the best performance with PBO enabled +XMP.
No, i am totally lost with the PBO, cant even tell if it is working in my rig, just manually overclocked my 3900X to 4.2G lol
Steve it was great. The video was well done, informative, and everything else. But amd's confusion on this will make me have to watch this video at least 3 times before i get a firm grasp as the what your saying. Anyhow, keep up the good work. Your shit lately is setting industry standards on how reviews and overviews should be done. And it is much appreciated. And you coverage of this ryzen release has been award winning, or at least it should be if awards were given for such things, and i think your peers as in fellow techtuberers would prob agree in that respect.
Thank you, Patrick, for putting so much care and effort into this topic.
Great testing, interesting results. Would help some of us if a visual cue, relationship, maybe graphically what PB is and then PBO, PBO+. Very confusing indeed but looks like PB is the way to go. Awesome channel digging into these things and saving the multitude OCers the wasted time, arguing etc. over the same thing. We know what to look for now. COOLING!
About time, you let your cat be part of the show. Really shows that you're ready to go to the next level.
In msi bios u can set the cpu boost override to +500mhz, my friends 3600 boosts to 4,55ghz single core my own goes to 4,45
JC AIC msi gaming pro carbon b350 en b450
how are your voltages? mine was hitting 1.5v+ at 4.4ghz
Andrew Puleston 1.4-1.5
@@Grena567 Hello, sorry for barging in, but is that mean that 1.4V - 1.5V is normal for 3600? I have mine went to 1.47V and I freaked out. My last Cinebench R20 test made my CPU thermals went up to 104C with a Corsair H100iV2, but I was too panicked to see what's the actual clock at that temp.
Andrew Puleston yea for me too
This was actually still a bit confusing. summed up most of the important parts after, that helped a lot
Hmm yes, yes... I have learned nothing about how to activate this.
All the other tech channels who didn't get this before need to watch
Depends on the SKU. On the 3700X PBO is needed because on stock it runs into Power-Limits immediately.
Yes I found this as well. That might be the only CPU in the entire lineup that it actually helps with.
Thanks for the video, I recently built a R5 3600/ASUS Tuf Gaming X570 -Plus system and thought I was doing something wrong because I wasn't seeing any difference with PBO. Thanks for all the hard work on something that makes little to no difference Patrick!
I get lower clock speeds with PBO on than with it off. Also my 3900X refuses to accept a higher PPT limit than the default 142 W even though my board's limit is 700 W. No matter what I set in BIOS or Ryzen Master, the Ryzen Master home screen still shows 142 W PPT, even though the increases for TDC and EDC show.
I watch the whole vid and found it very informing and I don't even buy AMD but It's good to see how far they have came in the last few years. Thx for the run down and debunk some of the confusion.
That is a lot of research to confirm that something is pointless, Through. Thank you, GN.
Thanks for the great effort in explaining these as I have a 2700X with an AIO that idles at 60 and other videos telling me that PB or PBO would help and trying to figure it all out on my own in bios and then realizing that none of it really matters and I should just get a bigger more open case for better airflow would solve the issue just saved me days of research. Much appreciated.
Nice informative test as always :) Im dying to see how a 3900x would OC when SMT is off compared to SMT ON. PBO2 / manual . I assume we could reach even higher clock speeds at lower voltage. And 12pure core is just perfect for gaming.
Tru that would be an interesting test
Curious aswell
Pls!make a video on unlocked power tables of Navi!!
I've seen a lot of questions regarding the temperature and voltage readings for these 3rd-gen chips from monitoring software, and the fact that these readings are often incorrect. I've seen reports stating that temps are actually 15-20 degrees Celsius cooler than what monitoring software is showing, as well as the fact that the voltage/frequency is affected by the monitoring software. I've also seen that CPU-Z should report actual voltages, but in the case of my Ryzen 3600, it seems to be showing 50 degrees at idle with a 4.2GHz frequency at all times. I'd love to see a video on this and how to properly monitor how the chips are performing. BTW - I've disabled other monitoring software like iCue and CAM when running CPU-Z.
Gamers Nexus to the rescue!!! Ive been beating my head off this for days trying to increase my 2600x's performance.
I'm just going to turn off PBO and let the processor do its own thing. With a better cooler. Because I bought an NZXT H500... At least my graphics card stays cool.
how have you been beating your head? I have a 2600x and its running like a champ if you need help with settings
PBO makes a slight difference on my 2700X. Just not enough to warrant the extra heat
PBO isn't officially supported by 2000-series AM4 processors. It was introduced with 2000-series Threadrippers only.
PBO increased my clockspeeds on my 2600x around 15 percent in cinebench but sadly my new bios update broke the feature
@Dat Strange Boi that sounds like something is wrong, I'm not 100% sure with the 2600 since i think you dont have all the settings the the 2600x has if i remember correctly.
If you really care about getting that extra gaming performance 3600mhz memory with good low and tight timings will get you far more FPS than getting a few more hundred MHz out of your ryzen 3000 cpu.
I really appreciate this info. I can only imagine how many hours where put into this research and for this I am grateful. Also this helps me understand so much and I will have to watch this a few times just so I can take all this in. You and Patrick are one hell of a team.
Thanks for the news on PBO Steve! Great content as always!
Jesus! This man hasn't slept! He's been dropping videos like mad these last couple days
Ryzenmaster feels like bloatware...but it actually works.
Yup, just like Intel's XTU... I'll never complain about first party tools. People like to crap on Wattman in the Radeon Software control panel, but AMD is doing A LOT more innovation with significantly less software development resources than it's competition.
@@carbonsx3 Now if only XTU wouldnt reset everything every time the computer crashes for ANY reason
Hearing PBO over and over before I had my coffee. Now I want a Peanut Butter and Orange marmalade sandwich.
Snowflake always makes me smile.
Thank you for this video!
Still getting to know my 3900X. I think it´s strange that it sometimes shows using 1.45V will basically idle in Win10..? Although it suddenly drops to 0.7V. Turned off PBO now since it seems kind of useless..
You are great at explaning in detail and I love that! Maybe an addintional summary + chart on screen could help make things more understandable. Like "PB is: *keywords* PBO is: Auto OC is: / PBO is NOT: - Auto OC is NOT: - just a thought :)
any plans for undervolting tests?
Nicely done on the explanation of the features and how they work.
2020: AMD attempts to rid itself of confusing name schemes with its new Ryzen 7 4790X
4790kX :P
@@Jimster481 4790kX:P07
@@bangyahead1 lmao
AMD 2020: We have a new cpu called 5700x, not to be confused with our gpu 5700xt
@@user-gl1bt1ns6z Meanwhile Intel has the i7-1165G7, because who needs names that aren't confusing in product lineups lol
Thank you for this extensive piece, great work!
I have a asrock b450m steel legend with a 5600 in it was wanting to do PBO but have no idea what I'm doing I seen videos where they do PBO inside the Bios and no videos on how to do it in ryzen masters step by step video type thing. So, think imma keep my 5600 at default settings
I would just undervolt it, this helps with power and stability + heat/temps ;)
Great summary of the AMD marketing. It seems like it would have been better for AMD to speak about automatic "turbo" or "boost" clocks instead of "Precision Boost", and "overboost" or "overclocking" instead of PBO.
Snowflake is better at handling components better than the Verge.
Seems there are so many variables when considering the clock that the cpu will run at? Why couldn't it just be as simple as my 9700k....set all cores to 4.9GHz, manually adjust the clock until it's stable then forget about it
Not that you asked, but a request (if anyone sees this) would be to compare the Skylake-based i7 6700K with maximum sustainable overclock to the R7 3700X with the same conditions to see how they compare. I've so far not really seen anything (other than additional threads) that makes the 3700X at 4.3GHz competitive with the 6700K at 4.6GHz. I know you keep the 7700K in the charts but it's at the three year point for upgrades and there are many of us with 6700K CPUs.
I would also be interested to hear GN's perspective on AMD's claim that overclocking voids warranty, which is against consumer protection laws in many countries (e.g. UK). AMD has made sure to include it on all slides I've seen as presentations while Intel officially (confirmed via phone call) states overclocking will not invalidate warranty.
I learned this best when I got a 1700X that would get up to 3.5GHz in games by default, and the best overclock I was able to get out of it was 3.7GHz, not ever getting up to the 3.8GHz boost clock, even on a single thread job, AMD claimed for the part. I replaced it with a 8700k that overclocks to 5GHz on all cores with no trouble.
"TDP = Power consumption" Only if power is spent and no work is done. (TDP is the THERMAL Design Power)
If some of the power consumed is used to work, then less power will transform into heat.
If work is done then there's a gap between them called efficiency, the bigger the gap the bigger the efficiency (do not mistake it with effectiveness
).
So your WORKING CPU will always consume more power than it's TPD. Some of it turns into work (good), some turns into heat(TDP).
So i came here to see if activating some kind of function in my Bios would make my PC go BRRRRR faster, i still don't know.
Thank you Gamers Nexus, I stair at these setting everyday and I wondered what they meant and would anything change if I fiddled with them. Thank you Patrick.
Love how GN are now including Astrology in there testing methodology, that’s just awesome 😂
The alignment of the planets is a scientifically observable variable. That makes it astronomy as well.
I know I am late posting, but thank you so much for this video!
Nice to see you slackers actually doing something, have Patrick handcuffed to the test bench in the back room for the past 2 weeks.
Informative video, just what I need..
Thank you
Amazing video. So long story short - PBO disabled and buy an awesome cooler for your Ryzen 3rd gen.
This was a great video. Still confused but my take away based on GN's testing was it does not do that much in the long run, especially in games. So I changed PB Overdrive in BIOS from Auto to Disabled. I am sure for everyday, casual use I will not notice or miss anything.
GN can you make a video showing performance between x370 vs x470 and x570 with the same CPU like 3600x or 3700x ( i have a 370x and don't know if its worth upgrading without upgrading my MB)
Clement Kirton other sites has tested this and no... there is not difference. Motherboardowerclock site did find out that max memory speed in older board was 3200 when x570 did run fine at 3600... but there was not any speed difference in test based on that ram speed difference. The cpu itself did run the same speed in x570 and older board. Some older board has so poor power delivery that overclocking is fouing to put vrm temperatures really high, so no owerclocking unles you vrm is good.
I don know if anything has significantly changed since this video was made, but if I enable PBO on my Asus Strix x370 Gaming-F I get 0,05ghz more (40.8 vs 40.3 multiplier) on all cores when doing a Cinebench r15 run and there is definitely a difference in scores. With that said I ran my custom water cooling setup (single 360 for Cpu only) on 100% on both runs and the difference in Cpu thermals was significant, as well as an increased power consumption. Temps went up from around 65c to 73c with PBO enabled. And the single-thread run was one point worse with PBO enabled.
For me the higher thermals and power consumption make it not worth is as I like to run my system as quiet as possible.
I'm watching this while rumdrunk and have absolutely no idea what's going on.
Opens Ryzen Master, clicks on Precision Boost Overdrive tab, clicks on Auto Overclocking, scratches chin, closes Ryzen master. looks for How to use Ryzen Master for Dummies.
Can you guys please do some benchmarks on gaming performance with the 3900 X with hyper-threading disabled?
SMT
@@BrunoRodrigoPintodaSilva well technically you're right but it's still two different names for the same damn thing LOL
ive seen benchmarks with 3600/x and 3700x with smt disabled and the performance differnec is nothing, within margin of error i believe. also not much of a difference in overclocking, it doesnt make it better even if you score and extra 100mhz on a single core or whatever. so theres really no need to test it he could prolly just say it in 2sec or so lol
I can't remember now if it was GN or Hardware Unboxed, but one of the 3600 reviews had a few tests with SMT off, and it improved overclocking (4.4Ghz all core instead of 4.3), but SMT on/off by itself made no difference. Based on this, and also TheGoodOldGamer doing IPC testing on multiple generations of older chips, it appears AMD's SMT is superior to Intel HyperThreading up through Skylake/CoffeeLake. Likely AMD's larger caches can handle the extra thread, or Intel's limited decode/dispatch throughput causes pipeline bubbles on the second thread that hurt performance in some workloads.
Kenneth Shores thought they did for TWW already?
Can't wait to plop a 3900X into my X370 board and just let it do it's thing. ;-)
Thanks for the massive investigation job.
Gamers nexus should have over a 1million subscribers
Damn that's all so confusing...
I just want the best settings for my Ryzen 5 3600. With the 1600 it was so simple. Just ad a bit of voltage and up the clock speed and you were sorted.
He showed that to accomplish that you have to do nothing at all. Using PBO and AutoOC (He got +200MHz, I couldn't even boot at +100 with my 3700X) yields a fraction of a frame in games. Manual OC is for the most part for adjusting the efficiency.
@@krozareq PBO doesnt change anything on my 3600.
If i turn it off or no, nothing changes, because at full load PB stays in 4000-4075Mhz with 82-86 *C
And PBO would only increase speeds after 4200Mhz. So in my case its useless, Since i dont have big 6 heat pipe cooler.
Only option for me is how i see it is to turn of PBO and PB and lock my cpu to 3.8Ghz to get more stable temps.
I get 76-80 *C in Metro Exodus, omg is that ok? I have arctic freezer 33 plus.
Anatolij Sergejev those are some spicy temps. I got rid of PB and manually overclocked my 3600x 4.3gzh at 1.25v. My temps don’t go over 70°C under load and my case fans don’t ramp up and down all the damn time anymore lol.
@@32biit I did 4.2Ghz at 1.225V. about 80 deg aida64. about 70-76 in games. I use really silent mode for gpu, cpu and case fans. Cpu is only 1000rmp at 70 deg and goes max 1350rmp at 80 deg.
70 in load thats very interesting. what cooler do you have?
@@Kern1909 I think you might have just lost the silicon lottery if your chip is getting that hot :/ My CPU fan doesn't go past 50% under load and my case fans stay at 40% which are inaudible. Do you live in a very hot place? That also might be why you're seeing those temps.
I dunno what it is, but I trust y'all more than the other channels.
This is a totally useless feature. I'm really sorry you spent so much time putting this together, but it is appreciated nonetheless.
PBO is good on paper, but useless in real world situations... It's sad given the potential of Ryzen 2. I guess that AMD doesn't want to implement features that would potentially damage the product. But in the end customers are left with a useless feature and a confusing name scheme 🤦♂️. Anyway good job guys, we appreciate the time and effort 👍
You guys should have more subs... Shame you never get in the recommended. Correct me if I'm wrong?
You're not wrong. Maybe they should change their channel name to Steve Tech Tips
I never saw them in the recommended until after I watched one video, because Luke recommended it on WAN Show.
GN videos were recommended to me... its how I found the channel in the first place.
Fuck, man. I lasted halfway through this. I'll try to watch the rest after I take a nap...
No offense to you, Steve. You're killing it.
I'm definitely grabbing one of these 3900X's. 16c is too much, but the 12 is just fine.
I was getting buyers remorse when I got the 3600X and not the 3600 but after testing on my rig its hitting 4.3 regularly w/only 1.28 vcore on my x370 board boost core enabled, no manual OC. Mainly game on this build.
Could this mean since the processors A: Don't respond to increased voltages and B: Given the contents of this video, could you undervolt it? Would that increase boost frequency?
the main limiting factor seems to be heat.
Undervolting would decrease heat but would also make it possibly unstable at the higher boost it would try to get with its newly gained temperature headroom.