RyZen 3000 Boost Issue: What's your take?

[TheoneandonlyMrK]

my point was lower their tier cpus should be getting better bins,it's not my fault you're taking it somewhere else to muddy the waters to the point of saying frequency becomes irrelevant when you've made some ipc progress.

I dissagree about the bins , and fair point I will mearly dissagree next time and let you tell the world where we should be at by now with Amd CPU's , that way we all get to know what to buy.

So, where's your list , since your now the man where is the list of parts prices and specs you Would have released if you were AMD.

Because personally I think your just trolling for an argument ,pm me for one if you want, at least then i could use the words i want.

This is a thread about Amd CPU's not hitting boost clocks , is that something your having an issue with or am i the issue because you started this by quoting me then bemoaning AMD'S non existent headroom for overclocking, ie irrelevant shite since ryzen 2.

 

[Bones]

My take on this is I don't care since I don't use boost.
I set mine to run where I want it to and that's it, none of the yo-yo'ing of CPU speeds boost makes happen.

 

[TheoneandonlyMrK]

My take on this is I don't care since I don't use boost.
I set mine to run where I want it to and that's it, none of the yo-yo'ing of CPU speeds boost makes happen.

Exactly how i would have run but honestly , on a r2600X i beat AMD by 100Mhz , on this 3800X i beat AMD by 50 ATM ,it;s not been worth the effort in that area lately for me, now memory and other bus clocking is a totally different beast .
on the 2600X i got 7200 physics score in 3dmark through this.(stck bench timespy)

 

[Aquinus]

Yeah, so clearly someone is a little sour given the way the poll options are worded so I'm going to refrain from choosing any poll options. You might as well have just put "I'm pissed," into each option.

Honestly, I would just be happy that the algorithm is so well tuned already that you'd be hard pressed to manually overclock it better, so there is that, on top of conditions being right for maximum boost clocks, there is that too. A maximum doesn't mean you'll always get it. My Vega "boosts" to 1630Mhz or so, but it doesn't always hold it. Sometimes it doesn't even get there, but that doesn't mean AMD is being deceitful. Maximum attainable is not the same thing as what you'll always see under whatever workload you're putting on it.

 

[Bones]

Speaking of "Take", some prefer it, some don't.
I have my reasons for setting it to run the way I have it but TBF my reasons are specific to me only.

I can't say it's better that way as a blanket statement, another's use of their machine may well make using boost a better way to run it.
To each his/her own.

 

[Deleted member 185158]

Speaking of "Take", some prefer it, some don't.
I have my reasons for setting it to run the way I have it but TBF my reasons are specific to me only.

I can't say it's better that way as a blanket statement, another's use of their machine may well make using boost a better way to run it.
To each his/her own.

My Zen+ 2700X is at 3800mhz at 1.188v while max P-state is 3.7ghz at 1.212v. I found the chip to be super efficient at this speed and voltage. Runs very cool.

Single core boosts go to 4350mhz with PBO and auto is 4ghz. OC setting 3 is 4.1ghz and OC setting 4 at 4.2ghz and right at 1.5v while I can produce the same at slightly lower voltages, on long crunching durations throttles as far down as 3.9ghz while crossing that 90c t-case.

Neat the chips boost and all that, but seems to me like a lot of energy is actually wasted while people play games with 4-8 core loads and a bunch of idle threads eating epeen.

Personally I prefer the SMT feature disabled at higher clocks. Other than showing off Cinebench scores, I get better performance for daily use.

 

[TheoneandonlyMrK]

My Zen+ 2700X is at 3800mhz at 1.188v while max P-state is 3.7ghz at 1.212v. I found the chip to be super efficient at this speed and voltage. Runs very cool.

Single core boosts go to 4350mhz with PBO and auto is 4ghz. OC setting 3 is 4.1ghz and OC setting 4 at 4.2ghz and right at 1.5v while I can produce the same at slightly lower voltages, on long crunching durations throttles as far down as 3.9ghz while crossing that 90c t-case.

Neat the chips boost and all that, but seems to me like a lot of energy is actually wasted while people play games with 4-8 core loads and a bunch of idle threads eating epeen.

Personally I prefer the SMT feature disabled at higher clocks. Other than showing off Cinebench scores, I get better performance for daily use.

ill try Smt off.

 

[Deleted member 185158]

ill try Smt off.

Interested in the temp difference. It will speak a lot on how much heat the SMT produces virtually for nothing.
Indeed turn it on when needed, but meh a few percent increase is a likable thing too

 

[oxrufiioxo]

Interested in the temp difference. It will speak a lot on how much heat the SMT produces virtually for nothing.
Indeed turn it on when needed, but meh a few percent increase is a likable thing too

I honestly think it's only going to cause people to go threadripper instead. Especially if its easier to cool with it's much larger ihs. The 16 core AM4 is going to be harder to cool than a 9900k I'm sure.

My 3900X already behaves similarly stock vs stock (all limits disabled) to my 9900k.

 

[Deleted member 185158]

Threadripper....

Epic Rome 7502 32 core 3.5ghz.

For FFXV lol...

 

[Nordic]

I have tested with smt on and off. I have not seen a difference in single threaded benchmark scores or heat.

 

[cucker tarlson]

I dissagree about the bins , and fair point I will mearly dissagree next time and let you tell the world where we should be at by now with Amd CPU's , that way we all get to know what to buy.

So, where's your list , since your now the man where is the list of parts prices and specs you Would have released if you were AMD.

Because personally I think your just trolling for an argument ,pm me for one if you want, at least then i could use the words i want.

This is a thread about Amd CPU's not hitting boost clocks , is that something your having an issue with or am i the issue because you started this by quoting me then bemoaning AMD'S non existent headroom for overclocking, ie irrelevant shite since ryzen 2.

some oc is bettter than no oc,not gonna argue about that since it's pretty obvious for someone that doesn't look at this through red glasses.I never said 6ghz.I said manual 4.5 over stock 4.2.you're being so defensive that all you can say is that I'm trolling.do you realize that is trolling ?

 

[thesmokingman]

Yeah, so clearly someone is a little sour given the way the poll options are worded so I'm going to refrain from choosing any poll options. You might as well have just put "I'm pissed," into each option.

Honestly, I would just be happy that the algorithm is so well tuned already that you'd be hard pressed to manually overclock it better, so there is that, on top of conditions being right for maximum boost clocks, there is that too. A maximum doesn't mean you'll always get it. My Vega "boosts" to 1630Mhz or so, but it doesn't always hold it. Sometimes it doesn't even get there, but that doesn't mean AMD is being deceitful. Maximum attainable is not the same thing as what you'll always see under whatever workload you're putting on it.

I noticed the same type threads on other forums also had salty as hell wording as well.

 

[biffzinker]

I have tested with smt on and off. I have not seen a difference in single threaded benchmark scores or heat.

Ryzen 5 3600 with SMT On vs Off (Stock clocks +100 MHz over the rumored 3500X.)

Geekbench 5
SMT On

SMT Off

AMD Ryzen 5 3500X CPU Listed

I woudn't be suprised at all if one of these turns out to be similar to what the i5's have been for gaming vs the i7's. Related to a bang for the buck investment, I don't see these as being a bad - Would probrably be an excellent value in the end. However it's still too early to really know...

www.techpowerup.com

 

[TheoneandonlyMrK]

some oc is bettter than no oc,not gonna argue about that since it's pretty obvious for someone that doesn't look at this through red glasses.I never said 6ghz.I said manual 4.5 over stock 4.2.you're being so defensive that all you can say is that I'm trolling.do you realize that is trolling ?

I've explained adequately why i dissagree , this still is not a thread about overclocking, we dissagree ,so what.

Red glasses , so now im a fanboi , not red glasses man, but the glasses of time Oc went west a while back, keep up with the news, moan when it might make sense like years ago possibly ryzen gen two Would have been the time if you're blue glasses fogged up.
Not gen 3.

 

[Aquinus]

Ryzen 5 3600 with SMT On vs Off (Stock clocks +100 MHz over the rumored 3500X.)

Geekbench 5
SMT On

SMT Off

AMD Ryzen 5 3500X CPU Listed

I woudn't be suprised at all if one of these turns out to be similar to what the i5's have been for gaming vs the i7's. Related to a bang for the buck investment, I don't see these as being a bad - Would probrably be an excellent value in the end. However it's still too early to really know...

www.techpowerup.com

1.9% single threaded increase for an 18% decrease in multi-core throughput. Doesn't seem like a worthwhile tradeoff. I would run it another 2 time for each because you might find that the numbers are within error on the single-thread side. A only a few points more or less on a second run would be close enough to say that it's with error to be honest.

 

[TheoneandonlyMrK]

1.9% single threaded increase for an 18% decrease in multi-core throughput. Doesn't seem like a worthwhile tradeoff. I would run it another 2 time for each because you might find that the numbers are within error on the single-thread side. A only a few points more or less on a second run would be close enough to say that it's with error to be honest.

Had a go, I have to say it did seam even smoother with regards to gaming , no big change , close to naught but not nothing and game benchmarks nudged up if gpu centric.

Crosshair hero VII bios 2801 just dropped, back in a bit. .

 

[cucker tarlson]

I've explained adequately why i dissagree , this still is not a thread about overclocking, we dissagree ,so what.

Red glasses , so now im a fanboi , not red glasses man, but the glasses of time Oc went west a while back, keep up with the news, moan when it might make sense like years ago possibly ryzen gen two Would have been the time if you're blue glasses fogged up.
Not gen 3.

no but you seems awfully defensive about the point I made as far as binning and clocks,and now I just discovered why looking in your specs.
I hope the extra investment is either paying off or gonna in the future cause that's a 20% premium over 3700x.

 

[TheoneandonlyMrK]

no but you seems awfully defensive about the point I made as far as binning and clocks,and now I just discovered why looking in your specs.
I hope the extra investment is either paying off or gonna in the future cause that's a 20% premium over 3700x.

Im not That skint , and im not bothered about your point regarding binning , it's made soley from a consumer Pov ,im an engineer though and was talking about it from a technical pov no red lenses just engineering common sense, no one Has to buy it (lower teir ryzen)and you won't find me on here trying to push people to buy anything i just pointed out that despite the frequency concerns and despite your points a r5 3600 is within spitting distance of a 9900K so THAT is why your points are irrelevant, that and the off topic nature, IMHO, chill im not defensive in any way i like a debate and I am short of time and direct.

On topic , bios 2801 is showing lower volts at any given clock and very ,very marginal boost improvement, nothing to write home about really hitting 4.55 on a single core ,4.5 on three others ,none won't do 4.75, at 18 ambient 56 on tdie.

 

[cucker tarlson]

I am short of time and direct.

I noticed that over the years

 

[Deleted member 185158]

no but you seems awfully defensive about the point I made as far as binning and clocks,and now I just discovered why looking in your specs.
I hope the extra investment is either paying off or gonna in the future cause that's a 20% premium over 3700x.

Always gotta pay more for a better binned chip.
But they dont bin to user overclocks.
Basically the 3800x is going to always have that high base clock stability no user intervention needed. So going to a 3800 from 3700 is performance increase right from the box

Ryzen 5 3600 with SMT On vs Off (Stock clocks +100 MHz over the rumored 3500X.)

Geekbench 5
SMT On

SMT Off

AMD Ryzen 5 3500X CPU Listed

I woudn't be suprised at all if one of these turns out to be similar to what the i5's have been for gaming vs the i7's. Related to a bang for the buck investment, I don't see these as being a bad - Would probrably be an excellent value in the end. However it's still too early to really know...

www.techpowerup.com

Nice. I got just about the same increase and decrease as you did percentage wise. Granted mines a zen plus, but all my point was looking at the per core increase.

As small as it is, it would be similar percentage to overclocking memory for that extra %.

Good stuff thank you

 

[TheoneandonlyMrK]

I noticed that over the years

So stop wasting it then.

You didn't talk back to my point in several posts way back that this method of chip production , to boost to it's max possible has been AMDs way since ryzen 2000 and arguably the first gen with an X and Xfr.

But now it's an issue.

Or that intel and Nvidia pioneered this method and still use it.

 

[cucker tarlson]

So stop wasting it then.

You didn't talk back to my point in several posts way back that this method of chip production , to boost to it's max possible has been AMDs way since ryzen 2000 and arguably the first gen with an X and Xfr.

But now it's an issue.

Or that intel and Nvidia pioneered this method and still use it.

lol,there is no law in the universe that forces you to respond to me in a thread like this,where it's opinion vs opinion.

 

[Nordic]

Ryzen 5 3600 with SMT On vs Off (Stock clocks +100 MHz over the rumored 3500X.)

Geekbench 5
SMT On

SMT Off

AMD Ryzen 5 3500X CPU Listed

I woudn't be suprised at all if one of these turns out to be similar to what the i5's have been for gaming vs the i7's. Related to a bang for the buck investment, I don't see these as being a bad - Would probrably be an excellent value in the end. However it's still too early to really know...

www.techpowerup.com

I know techpoweup did a review with smt off. I did my own testing. I found no difference in average single threaded benchmark scores. For example Cinebench r20 averaged 500 points with and without smt. Multithreaded performance suffered without smt though.

 

[TheoneandonlyMrK]

lol,there is no law in the universe that forces you to respond to me in a thread like this,where it's opinion vs opinion.

And no law that limits the dodging of valid points lucky for you since they make your opinion dubiously flawed.
Bye