Is PBO worthless? Ryzen 5 2600x

[Suspect3d]

So I wanted to get some more juice out of my 2600x. When I built it 6 months ago i didn't touch bios at all, I just set my rig and played. When I started to look at OC, i saw about PBO. well when I looked at my system it looked like my clock a 4.2ghz give or take a few. SO I assumed PBO was already enabled.  Out of curiosity, I downloaded Auto suite for Ryzen and did its things. 3.8ghz acreoss all cores. Did all my benchmarks and I saw a 1-4% gain on everything I ran. This was all using high performance powerplan. 

 

Why am I seeing better performance on auto OC 3.8 vs the PBO which i understood to be higher clocks? I'm looking at getting a better Cooler and manually OC now.

[spartaman64]

4.2ghz is the boost clock which is not pbo and i think its only on a few cores not all the cores

[5x5]

PBO is going to be far better than a manual OC when you try to go for less heat and noise since it dynamically shifts clockspeeds where needed. It's usually also better for gaming.

[Suspect3d]

16 hours ago, 5x5 said:

PBO is going to be far better than a manual OC when you try to go for less heat and noise since it dynamically shifts clockspeeds where needed. It's usually also better for gaming.

Hey @5x5 so I watched some videos and figured out how to turn on the PBO on my system.  I ran some tests and just idle and I notice as it adjusts that it bounces over 1.4v pretty often. I was told by a few people that seems a bit high. I've never done OC before so this is all new for me.  Also during all the benchmarks I did on PBO, they were lower than my Benchmarks with the auto OC at 3.8. So i'm confused how to this is better?

[paulmohr]

1.4 sounds about right to me. Just remember it only applies that to the cores that need it, not all the cores are ramping up to 1.4 volts. It also isn't steady, it jumps up and down faster than your monitoring software can keep up. And the software might not even be accurate in reporting the actual voltage. It is also regulating current, not just voltage so it can hit a higher voltage without doing anything bad.

 

As for benchmarks they are what they are, all depends on what benchmark you are running and what it is testing. Most things you actually do on a computer are not going to load up every core that hard. If you are running a benchmark that tests multi core under full load then yes a manual overclock will probably score higher. If you run one that tests single core performance then PBO will probably give you the better score.

 

PBO will boost individual cores at different times much higher than you will ever get with a manual overclock. It is what it was designed to do. Which do you want? Only way to know is test it with what you actually do on your computer.

 

If you game run it stock and see what kind of fps you get in different games. Then turn PBO and check the same games to if it increased and by how much. Then do the same thing with your manual overlcock. Which ever one gave you a considerably higher fps then go with that. If it had no or little effect just run it stock. Like say you got 200 fps stock and overclocking got you 210 fps. I would just run it stock since that 5 or 10 fps isn't really doing anything for you but generating more heat. Now if you were trying to run at 60 hz or whatever and that extra 10 fps keeps you from dropping below 60 fps then yes it would help.

 

Bench marks are nifty and all, but they don't always relate to real world performance. This why reviewers run many different benchmarks, tests and games to evaluate a product. I use benchmarks to see if a new component is significantly better than another one and to stress test.

[Suspect3d]

2 hours ago, paulmohr said:

1.4 sounds about right to me. Just remember it only applies that to the cores that need it, not all the cores are ramping up to 1.4 volts. It also isn't steady, it jumps up and down faster than your monitoring software can keep up. And the software might not even be accurate in reporting the actual voltage. It is also regulating current, not just voltage so it can hit a higher voltage without doing anything bad.

 

As for benchmarks they are what they are, all depends on what benchmark you are running and what it is testing. Most things you actually do on a computer are not going to load up every core that hard. If you are running a benchmark that tests multi core under full load then yes a manual overclock will probably score higher. If you run one that tests single core performance then PBO will probably give you the better score.

 

PBO will boost individual cores at different times much higher than you will ever get with a manual overclock. It is what it was designed to do. Which do you want? Only way to know is test it with what you actually do on your computer.

 

If you game run it stock and see what kind of fps you get in different games. Then turn PBO and check the same games to if it increased and by how much. Then do the same thing with your manual overlcock. Which ever one gave you a considerably higher fps then go with that. If it had no or little effect just run it stock. Like say you got 200 fps stock and overclocking got you 210 fps. I would just run it stock since that 5 or 10 fps isn't really doing anything for you but generating more heat. Now if you were trying to run at 60 hz or whatever and that extra 10 fps keeps you from dropping below 60 fps then yes it would help.

 

Bench marks are nifty and all, but they don't always relate to real world performance. This why reviewers run many different benchmarks, tests and games to evaluate a product. I use benchmarks to see if a new component is significantly better than another one and to stress test.

Hey, thanks for the reply. what is funny I don't think I had PBO enabled. I watched some vidoes and messed with my BIOS and found the "PBO" options. which before I think I just turned on performance enhance mode. I also downloaded  Ryzen master to see more info and I ordered a new Air CPU cooler today that will be here tomorrow. Anything else I need to know about this?

[jxh5760]

I have 2600x and was just doing some overclocking. turn XFR and PBO to Enable I get the PBO OC, which in my case yields same benchmark scores as my manual OC as long as they are the same all core clock speed. during cinebench r15 single core test, PBO gives me 4.2Ghz and a score of 169. When I manual OC to 4.2Ghz, doing cinebench r15 single core again, i got 171. SO basically no difference. But PBO runs it at 1.35V and my manual runs it at 1.3375V, temp at load is the same but idle temp is better with manual OC. 

 

Overall for me, my manual and PBO OCs have the same benchmark results. But manual OC gives me lower temps. So I'm going with manual.

 

Edit: also, multi core cb15 gives identical scores.

[Suspect3d]

6 minutes ago, jxh5760 said:

I have 2600x and was just doing some overclocking. turn XFR and PBO to Enable I get the PBO OC, which in my case yields same benchmark scores as my manual OC as long as they are the same all core clock speed. during cinebench r15 single core test, PBO gives me 4.2Ghz and a score of 169. When I manual OC to 4.2Ghz, doing cinebench r15 single core again, i got 171. SO basically no difference. But PBO runs it at 1.35V and my manual runs it at 1.3375V, temp at load is the same but idle temp is better with manual OC. 

 

Overall for me, my manual and PBO OCs have the same benchmark results. But manual OC gives me lower temps. So I'm going with manual.

 

Edit: also, multi core cb15 gives identical scores.

ya PBO right now is running 1.375 to 1.42. whats funny is I did the auto Suite 3 over clock which no one seems to like. it only capped me at 3.8 on all cored but 1.3v  and it was 1% slower than PBO. My new cooler comes tomorrow. I'll be messing with manual

[jxh5760]

10 minutes ago, Suspect3d said:

ya PBO right now is running 1.375 to 1.42. whats funny is I did the auto Suite 3 over clock which no one seems to like. it only capped me at 3.8 on all cored but 1.3v  and it was 1% slower than PBO. My new cooler comes tomorrow. I'll be messing with manual

What cooler did you get? I got a NZXT X52 and it's crap. I'm at 80C max in aida64 with 1.337V and that's 51C over ambient. 

[Suspect3d]

1 minute ago, jxh5760 said:

What cooler did you get? I got a NZXT X52 and it's crap. I'm at 80C max in aida64 with 1.337V and that's 51C over ambient. 

Cooler Master MAP-T6PN-218PC-R1

[paulmohr]

13 hours ago, Suspect3d said:

Anything else I need to know about this?

I don't actually have that cpu or that motherboard so I really can't help you on how it all works and what the options are. They are all different for some stupid reason. I think bios settings and naming should be standardized, just a pet peeve on mine.

 

I would only reiterate that you should test it on things you actually do on your computer and not rely solely on benchmark results, or even bother with them honestly. It either works better for your use case or it doesn't, that is the bottom line. There is no point in putting extra stress and heat on your parts if it gives you little to no gain with what you intend to use it for. If I overclock my cpu, memory and video card (which I did) yes my benchmark scores go up. R15 showed a slight improvement and gaming benchmarks like time spy and fire strike went up as well. So did my voltage, wattage and heat output. However when I fired up a couple of games that I actually play it made no difference at all. Like Project Cars 2, it made no difference at all. I run this game at 4k on a 60 hz 55 inch tv. Most of my setting are ultra, except the effects I don't like and I turn those off. I use Vsync that locks my fps 60 fps because I hate screen tear. My video card only uses between 1200-1400 mhz while playing and it never comes close to dipping below 60fps, even before I overclocked everything. So for me overclocking makes zero sense. Now if I get a new game or something and it has trouble holding 60fps then yes overclocking might be an advantage. Or I could just tweak a few settings in the game possibly.

 

 

[Suspect3d]

22 hours ago, paulmohr said:

I don't actually have that cpu or that motherboard so I really can't help you on how it all works and what the options are. They are all different for some stupid reason. I think bios settings and naming should be standardized, just a pet peeve on mine.

 

I would only reiterate that you should test it on things you actually do on your computer and not rely solely on benchmark results, or even bother with them honestly. It either works better for your use case or it doesn't, that is the bottom line. There is no point in putting extra stress and heat on your parts if it gives you little to no gain with what you intend to use it for. If I overclock my cpu, memory and video card (which I did) yes my benchmark scores go up. R15 showed a slight improvement and gaming benchmarks like time spy and fire strike went up as well. So did my voltage, wattage and heat output. However when I fired up a couple of games that I actually play it made no difference at all. Like Project Cars 2, it made no difference at all. I run this game at 4k on a 60 hz 55 inch tv. Most of my setting are ultra, except the effects I don't like and I turn those off. I use Vsync that locks my fps 60 fps because I hate screen tear. My video card only uses between 1200-1400 mhz while playing and it never comes close to dipping below 60fps, even before I overclocked everything. So for me overclocking makes zero sense. Now if I get a new game or something and it has trouble holding 60fps then yes overclocking might be an advantage. Or I could just tweak a few settings in the game possibly.

 

 

Thanks, I'm running 1440p now myself.  ya I agree the MOBO need to be the same. It's funny how all the ASUS boards are slightly different settings across the board. 

 

So what I did was this 2 nights ago.   Reset all settings on my MOBO. so performance enhancer etc is all set to auto. this had me bouncing between 3.9-4.0GHz. I testing R15 and some games to get some ideas.  I noticed my volts about mostly be 1.375 range, but idle over 1.4. My heat in R15 was about 82 by the end of it.

 

 

Then I enabled PBO- same tests. again ranged from 3.9ish -4.1ghz a few times. Heat for R15 hit 92c at the end. Score was 60 points higher than me leaving at stock settings/profiles. Games I test/benchmarks were almost identical, some slightly better/worse.

 

AI Suite III (yes the Auto OC that everyone hates). Only boosted me to 3.8GHZ on all cores at like 1.28v. R15 was bout 140 points lower than PBO but never hit more than 75c. Now gaming benchmarks were beter by a few FPS and I seen a % boost on 3d mark for every test I ran.. This to me seemed very odd. BUT for some reason no matter what I do for settings the Suite always overrides my power plan to balance and puts my PC to sleep. had few lock ups, so I uninstalled it.

 

 

Lastly, Last night I got my new cooler installed. I ran r15 again all stock MOBO settings. My score was the same as before within a few points but my Temps never got past 61c. So it dropped 20c over the stock ryzen cooler.  My goal was to not be gimped by my CPU in gaming at all. I'm not sure I am in most cases but I do notice alot that when I run in game benchmarks etc my GPU never goes over 95%.

 

Sorry long reply

[paulmohr]

I think, and let me stress "think" becuase I am not an avid gamer and I am only going by what I have seen in vidoes on youtube, that at 1440p you might be mostly gpu bound. It might depend on the game and what graphics settings you are using though.

 

I am glad the new cooler dropped your temps by that much, that is awesome. I would try doing the same tests again with new cooler and see what happens. If you get a decent increase with pbo stick with that I guess. If you get a decent increase with an all core overclock then the games you playing might actually like that and maybe look into doing a manual overclock. IF, those increases actually make a difference in your game play. If it is only by a few fps or something and it makes no difference in how the game plays then honestly I wouldn't mess with it, but that is me personally.

 

Something I noticed on my system, if I unlock my fps in project cars at 4k it will hit my gpu pertty hard trying to get as many fps as it can. Boosting to whatever my card will go to ( I think its like 1900mhz or something?) and it will heat up pretty good. Once I turn vsync on  and it locks my frame rate to 60 fps it runs much lower. Most of the time its around 1200 mhz but will hit peaks of 1400 mhz. And the temp only hits like 55c.

[Suspect3d]

2 hours ago, paulmohr said:

I think, and let me stress "think" becuase I am not an avid gamer and I am only going by what I have seen in vidoes on youtube, that at 1440p you might be mostly gpu bound. It might depend on the game and what graphics settings you are using though.

 

I am glad the new cooler dropped your temps by that much, that is awesome. I would try doing the same tests again with new cooler and see what happens. If you get a decent increase with pbo stick with that I guess. If you get a decent increase with an all core overclock then the games you playing might actually like that and maybe look into doing a manual overclock. IF, those increases actually make a difference in your game play. If it is only by a few fps or something and it makes no difference in how the game plays then honestly I wouldn't mess with it, but that is me personally.

 

Something I noticed on my system, if I unlock my fps in project cars at 4k it will hit my gpu pertty hard trying to get as many fps as it can. Boosting to whatever my card will go to ( I think its like 1900mhz or something?) and it will heat up pretty good. Once I turn vsync on  and it locks my frame rate to 60 fps it runs much lower. Most of the time its around 1200 mhz but will hit peaks of 1400 mhz. And the temp only hits like 55c.

Thanks, I'll look into it more tonight. I'm trying to understand the Ryzen 2nd gen better as far as XRF and PBO goes. I might be missing settings as well.  I typically play pay a mix of high/ultra settings on 1440p 144hz. I have a Aorus RTX 2080. I guess what I really need to do is wipe out my benchmarks and start from scratch and record them all again. I agree that if I see minimal gains 1-3% but running more power/temps. It's not going to be worth . Thanks the the reply's, I appreciate it.

[Buggerlugz]

On 8/8/2019 at 4:42 PM, paulmohr said:

"I think, and let me stress "think" becuase I am not an avid gamer and I am only going by what I have seen in vidoes on youtube, that at 1440p you might be mostly gpu bound. It might depend on the game and what graphics settings you are using though."

Actually its the opposite, the lower the resolution the more GPU bound you are, the higher the resolution the more emphasis is placed on the GPU. Lowering the resolution of a computer game or software program increases the effect on a CPU. As the resolution decreases, less strain is placed on the graphics card because there are fewer pixels.