Ryzen Owners Zen Garden

[lexluthermiester]

R3-1200 ordered with a MSI B350M PRO-VDH, and some cheap 2x4GB 2400MHz Crucial RAM. For now I'm recycling my HD7750 and a notebook 1TB HDD, I need a VGA port for my monitor and the APUs are still far away.

That's a good setup. Are you going to stick with Radeon for a GPU or were you thinking of an NVidia option?

4ghz is a mild overclock.

At least we can agree on one thing..

 

[notb]

So again, your experience was bad, mine was overwhelmingly good. I must be wrong. Let's move on.

Just a common-sense note for you. @cdawall is talking about his experience with PCs he has. You're talking about builds you sell. So no offense, but the fact that you had no problems with T4 during few hours of testing is not that convincing.
And I'm afraid you're wrong about Ryzen's power consumption and temperatures. Ryzen 7 has been shown to pull 140-150W easily under load and with some pretty pedestrian OC. And AMD TDP is something different than Intel's TDP.
The 70*C you've mentioned is just way too high for this chip. Again: you can't base your judgement on experience with Intel. Intel's chips tolerate much higher temperatures.

Remember that temperature will increase with time: as both the cooler and TIM will slowly lose their efficiency.
And even before that happens, dust gathering on the cooler can "provide" additional 5*C as well (and even 10*C for top-down coolers, easily).
I presume it's winter where you live at the moment. What happens in summer, when the ambient room temperature increases by 10*C or so?
Because of things like this you ought to be seeing solid temp margin on a new setup: 15-20*C at least, IMO. And you left none.

Not to mention that even if that Ryzen 7 survives, the T4 will be really stressed and the fan will be at max RPM a lot => high noise.
General rule of thumb for cheaper coolers: get those designed for significantly higher power consumption. Personally, I try not to go below 150%, more or less.
This isn't necessarily true for expensive ones, since they're built to operate properly also near the limit.

The first T4 review that google returns:
http://www.dvtests.com/cooler-master-hyper-t4-test-and-review/
Check the graphs at the bottom. And it's on an Athlon X4 620, so TDP=95W.

 

[GoldenX]

That's a good setup. Are you going to stick with Radeon for a GPU or were you thinking of an NVidia option?

I want a Pascal or Volta in the far future, but that means getting a new digital output monitor (this one is 7 years old and still strong), while the Vega IGP offers VGA. Before that I have to get a proper HDD, a M.2 SSD, more better clocked RAM, maybe an UPS, and I'm dry with just that CPU+MoBo+RAM combo.

 

[lexluthermiester]

Just a common-sense note for you. @cdawall is talking about his experience with PCs he has. You're talking about builds you sell. So no offense, but the fact that you had no problems with T4 during few hours of testing is not that convincing.

8 hours under Prime95 load is enough to prove stability for anybody's needs. And I run Furmark for 8 hours after that to test any GPU OC applied, if needed. Your conclution is not "common" sense. Don't care if a few people are not "convinced" by the methodologies used for qualifying a build as stable and ready for customer use. They work for what my clients need and they stay stable.

And it's getting a bit worse when you're just clearly wrong about Ryzen's power consumption and temperatures. Ryzen 7 has been shown to pull 140-150W easily under load and with some pretty pedestrian OC.

I have a wattage measurement unit and have calculated the differences. The most I have witnessed a Ryzen 7 pull through a circuit was 118watts. And that was OC'd. Granted, I haven't measured every Ryzen system built as such an effort would be a waste of time. And for the record, the 65w and 95w specs quoted earlier are from AMD.
https://www.amd.com/en/products/cpu/amd-ryzen-7-1800X#product-specs
https://www.amd.com/en/products/cpu/amd-ryzen-5-1400#product-specs
Oh, and before anyone goes there, I know the difference between TDP and electrical wattage. We are discussing a cooling solution and in that context only the TDP matters.

And AMD TDP is something different than Intel's TDP.

It sure is, isn't it.

The 70*C you've mentioned is just way too high for this chip.

Where are you getting that? The Ryzen line doesn't even start to thermal throttle until it hits 90C and shuts down completely 100c. 70c for brief moments is not a problem on any level.

Again: you can't base your judgement on experience with Intel. Intel's chips tolerate much higher temperatures

Ok, good to know. Thanks for the tip.

Remember that temperature will increase with time: as both the cooler and TIM will slowly lose their efficiency.

That depends greatly on the TIM being used. Some don't degrade at all. And Arctic Silver 5 actually gets better as it sets. That has proven many times.

Also, at the moment it's winter where you live. What happens in summer, when the ambient room temperature increases by 10*C or so?

Ok, and what about the systems I built this past summer? None of them had or are having any problems.

Not to mention that even if that Ryzen 7 survives, the T4 will be operating near the limits => high noise.

The fans only got noisy under load. Not worried.

General rule of thumb for cheaper coolers: get those designed for significantly higher power consumption. Personally, I try not to go below 150%, more or less.

Again, Thanks for the tip. However, gonna keep going with the methodologies that have served both me and my clients well for as long as I can remember, which is more than 3 decades. Yeah..

So again, some have had experiences that were bad, mine have been good. I must be doing something wrong. Let's move on.

I want a Pascal or Volta in the far future, but that means getting a new digital output monitor (this one is 7 years old and still strong), while the Vega IGP offers VGA. Before that I have to get a proper HDD, a M.2 SSD, more better clocked RAM, maybe an UPS, and I'm dry with just that CPU+MoBo+RAM combo.

Given your choice of CPU, you're likely to get the best bang for buck out of a GTX 1060 6GB. With a lot of games that CPU is likely to bottleneck the GPU and a 1070 and 1080 will see diminishing returns. That's not to say that the 1070 or 1080 would not give great performance on games that are GPU bound. Yet giving your CPU a modest OC will help a lot in games that are more CPU dependent. And the money you save going with a 1060 would allow you to get a good sized SSD to fit into the M2 slot on your new mobo.

 

[notb]

8 hours under Prime95 load is enough to prove stability for anybody's needs.

No, it's not.
Putting aside how relevant torture testing a new PC is for stability during next 3+ years, why not test for at least 24h straight? That used to be the common recommendation back in a day when I was still having fun with OC. And back then AMD CPUs were called XP - like experience. I presume FX has some connection with fireworks. Not sure about Zen, but you clearly need a lot of patience to setup this thing properly. ;-)

I have a wattage measurement unit and have calculated the differences. The most I have witnessed a Ryzen 7 pull through a circuit was 118watts.

I can't argue with that. But others have experienced higher values. So hey... maybe the Ryzen's your clients are using will also reach that 140W one day? What then?

And for the record, the 65w and 95w specs quoted earlier are from AMD.

Of course they are. But just the fact AMD says 95W doesn't mean that the CPU won't pull more.

That depends greatly on the TIM being used. Some don't degrade at all. And Arctic Silver 5 actually gets better as it sets. That has proven many times.

All matter degrades. All TIM lose efficiency, all cooler fans become louder with time. No way around it.

Ok, and what about the systems I built this past summer? None of them had or are having any problems.

I don't know. Luck? Relatively cool season?
What about the degradation? Will they work well next summer?

Again, Thanks for the tip. However, gonna keep going with the methodologies that have served both me and my clients well for as long as I can remember, which is more than 3 decades. Yeah..

I don't see how your experience from 30 years ago could be useful with modern PCs. They're totally different. But what you should have learned during those 3 decades is not to believe in TDP on consumer CPUs.
And BTW: based on earlier discussions I did fall under impression that you're much younger (like mid 20s, max 30).

So again, some have had experiences that were bad, mine have been good. I must be doing something wrong. Let's move on.

Well.. we're in fact talking about cooling. So if I were you, I'd base my decisions on the bad experience and just be happy that they aren't mine.
I'm in the business of looking for worst case scenario, so at least for me this seems pretty natural.

 

[cdawall]

don't see how your experience from 30 years ago could be useful with modern PCs. They're totally different. But what you should have learned during those 3 decades is not to believe in TDP on consumer CPUs.

Hey now I am sure his experience in punch cards and mainframes applies very well into a ryzen build. The "I have been doing this since WW nam" is a joke to me.

 

[R-T-B]

Why is there no Zen in my Zen garden?

I come here to share builds, not fight over their eccentricities. This is not pleasing.

Since there is no Zen, here is my Ryzen pain today:

Built a new ASRock build for myself, it wouldn't clock above 2.2Ghz. It turns out if you set a fixed voltage, this is a known issue on the latest BIOS of the ASRock Taichi. Nice one. A downgrading we go...

 

[Mussels]

you guys arguing over the temps and clocks: dont forget the 20C offset between X and non X chips.

Without specifying the amount of cores, and X or not every time you make these arguments you're comparing teslas with squirrels and driving me nuts.

 

[lexluthermiester]

No, it's not.
Putting aside how relevant torture testing a new PC is for stability during next 3+ years, why not test for at least 24h straight? That used to be the common recommendation back in a day when I was still having fun with OC. And back then AMD CPUs were called XP - like experience. I presume FX has some connection with fireworks. Not sure about Zen, but you clearly need a lot of patience to setup this thing properly. ;-)

3, 4, 5 years. The systems I build rarely fail and when they do it's the power supply most of the time. There have been OC's have have failed after a few years, but it's usually the extreme OC's and by then most people are good for an upgrade anyway.

I can't argue with that. But others have experienced higher values. So hey... maybe the Ryzen's your clients are using will also reach that 140W one day? What then?

Are you saying that CPU's use more power over time? Never seen this. Ever. It just doesn't happen. Neither Intel nor AMD CPU's will ever do that. Otherwise 486's, Pentiums, K6's and XP's would all be non-functional because they'd draw to much power. Sorry, the laws of physics don't work that way.

Of course they are. But just the fact AMD says 95W doesn't mean that the CPU won't pull more.

Of course it'll draw more power when overclocked, but 50% to 110% more in a modest 10% to 30% OC state? Sorry, never seen that either. It just doesn't happen.

All matter degrades. All TIM lose efficiency, all cooler fans become louder with time. No way around it.

Because of the chemistry involved, it would take decades for currently available quality TIM's to degrade that much. Older stuff, ok it does and has, but it take more than 10 years. Ever taken apart an old Pentium 1 system and looked at the TIM? It's hard and crusty, but it's still doing the job it was intended to do.

I don't know. Luck? Relatively cool season? What about the degradation? Will they work well next summer?

Luck? Most of my clients are repeat customers or referrals from same. If the equipment being sold was failing left and right as is implied it would, there would not be nearly as much business as I do.

I don't see how your experience from 30 years ago could be useful with modern PCs. They're totally different.

You don't understand that experience gained is valuable? Or is it that you don't understand how it is relevant to the Ryzen platform? Overclocking has evolved over time but is essentullly achieved by the same methodology; Change the clock speed. If unstable try a higher voltage. If CPU gets hot, cool it. Repeat until maximum stable OC is reached.

But what you should have learned during those 3 decades is not to believe in TDP on consumer CPUs.

Why not? When a CPU manufacturer tests and retests the products they make and state specifications, they're not just slapping a random value in the spec sheet willy-nilly. It is because they've tested those values and have stated them for reference so all of their potential customers understand what to expect.

And BTW: based on earlier discussions I did fall under impression that you're much younger (like mid 20s, max 30).

LOL! I'm young at heart and I'll admit to being the occasional jackass. But yeah, I'm a grandfather if that helps you out.(Saying that just made me a little sad.. But I love my grandkids!)

Well.. we're in fact talking about cooling. So if I were you, I'd base my decisions on the bad experience and just be happy that they aren't mine.
I'm in the business of looking for worst case scenario, so at least for me this seems pretty natural.

Right. And all I'm trying to point out with this issue(and at this point I'm sorry I brought up the Hyper T4) is that someone was looking for a good cooler and I recommended one that had given good performance. Sure there are better coolers out there, but they come at a higher price. The T4 is a lower profile cooler which will fit into smaller cases, which is the impression that I got reading about the goal of the system build. So there we are.

 

[notb]

Hey now I am sure his experience in punch cards and mainframes applies very well into a ryzen build. The "I have been doing this since WW nam" is a joke to me.

Well... I wouldn't be that hard on him. Electronics work pretty much the same, so I find it comforting that he has a wattmeter and all that.
But as far as CPUs and overclocking it's a vastly different reality... Even compared to a decade ago, when some really massive performance jumps were possible. Simple fact is: we got much closer to silicon limits and since so many CPUs will OC on their own (boost, auto tuning etc), you really need a good reason for manual tinkering. And if you're going hard on your CPU, you have to be careful.

So apart from the general discussion about temperatures and coolers, I'm pretty amazed that he's doing heavy overclocked PCs for other people - with no guarantee whatsoever that they'll take care of dust, that they know how to read PC temperatures and so on...

Why is there no Zen in my Zen garden?

There isn't much Zen in Zen anyway.
https://trends.google.com/trends/explore?date=today 5-y&q=ryzen problem,skylake problem

I come here to share builds, not fight over their eccentricities. This is not pleasing.

But do you share them hoping for feedback or do you share just for fun of sharing?

Since there is no Zen, here is my Ryzen pain today:
Built a new ASRock build for myself, it wouldn't clock above 2.2Ghz. It turns out if you set a fixed voltage, this is a known issue on the latest BIOS of the ASRock Taichi. Nice one. A downgrading we go...

At least it'll save the battery! Or polar bears. Or something.

 

[R-T-B]

At least it'll save the battery! Or polar bears. Or something.

I prefer to save my hair.

Bah! The old bios does it too.

I know, it's only for benching as I'd never really run my Ryzen at 1.45v (even if it does clock to 4.2) but WTF ASRock? I want my fun.

 

[lexluthermiester]

Why is there no Zen in my Zen garden? I come here to share builds, not fight over their eccentricities. This is not pleasing.

Sorry man, I'll chill out. Sometimes it's hard to tell when someone is genuinely misunderstanding/misinformed or if they're just plain trolling.

Since there is no Zen, here is my Ryzen pain today:
Built a new ASRock build for myself, it wouldn't clock above 2.2Ghz. It turns out if you set a fixed voltage, this is a known issue on the latest BIOS of the ASRock Taichi. Nice one. A downgrading we go...

That is bizarre. ASRock boards are usually rock solid. And on a few ASRock boards, I've locked the voltage on several builds. That problem didn't happen. Are you adjusting the multiplier to match? Maybe a FW update is needed from ASRock?

But as far as CPUs and overclocking it's a vastly different reality... Even compared to a decade ago, when some really massive performance jumps were possible. Simple fact is: we got much closer to silicon limits and since so many CPUs will OC on their own (boost, auto tuning etc), you really need a good reason for manual tinkering. And if you're going hard on your CPU, you have to be careful.

That is true. And this is the evolution of OCing I mentioned. We are getting very close to the limits of Silicon. However, the game will be changing soon as the use of other metalloids such as Germanium, Arsenic, Antimony and Tellurium are being heavily explored. It's only a matter of time til a break-through is made in one of those areas. Not everyone wants an OC. But for those that do, teaching them how and why things work helps them understand the importance of proper care.

So apart from the general discussion about temperatures and coolers, I'm pretty amazed that he's doing heavy overclocked PCs for other people - with no guarantee whatsoever that they'll take care of dust, that they know how to read PC temperatures and so on...

But I teach them about proper maintenance and the importance of such. I literally drill it into them, if needed. The benefits of OCing can be remarkable, but it comes with diligent responsibility, as you clearly seem to know. In 2002 when the breathable foam came out, I was an early adopter because of the clear benefits of such use. That was when OCing for client's started to be more commonplace. When using such, a can of air and a vacuum was all that one needed to keep things cool and clean.

 

[R-T-B]

That is bizarre. ASRock boards are usually rock solid. And on a few ASRock boards, I've locked the voltage on several builds. That problem didn't happen. Are you adjusting the multiplier to match? Maybe a FW update is needed from ASRock?

Happens above 1.35v only, so not much of a concern for good ol' daily clocks, but very weird.

 

[_JP_]

My Dad's parts are ready

 

[grunt_408]

If and when I change my cooler it will be a custom water loop..

 

[Sempron Guy]

you guys are arguing over a 70c max load temp at 4ghz? I'm on a custom loop and breaks that 70c easily in WCG at 3.9ghz

 

[cdawall]

you guys are arguing over a 70c max load temp at 4ghz? I'm on a custom loop and breaks that 70c easily in WCG at 3.9ghz

What is wrong with your loop? The custom loop I built didn't even come close to that and it had a 1070 in it.

 

[Sempron Guy]

What is wrong with your loop? The custom loop I built didn't even come close to that and it had a 1070 in it.

30-32c ambient

 

[Johan45]

Merry Christmas, here's a fun song by some local talent

 

[cdawall]

30-32c ambient

That is about 8C higher than what I was dealing with makes more sense lol

 

[ozomax1997]

im looking at upgrading my PC some time next year, is the price difference between a 1600 and a 1600x justifiable on a budget?

 

[lexluthermiester]

im looking at upgrading my PC some time next year, is the price difference between a 1600 and a 1600x justifiable on a budget?

That depends on your budget and what your computing needs are. If you're not going to be doing much/serious gaming, go for a lower end or used video card and drop the extra money into the CPU. Also, prices are going to change next year as the new gen of Zen CPU's from AMD are going to hit the market, which will force the price of current gen ZEN lower. The best thing you can do right now is to save as much money you're willing to spend and then when you're ready price out parts at that time. If you're on a smaller budget, don't be afraid to look at used parts to save money.

 

[notb]

That is about 8C higher than what I was dealing with makes more sense lol

Which only supports the case discussed earlier: what will happen to a 70*C Ryzen setup when it's July, not December...

I think people really underestimate the fact that there are 8 cores under the bonnet. And a lot of efficiency and temp praise is based on poor utilization right now, yet usually they're accompanied by hopes that when software manufacturer catch up these CPUs will really shine...

 

[grunt_408]

I only built mine to go to the PUB G

 

[GoldenX]

Ryzen 1200 arrived, the biggest star are the Crucial Ballistix Sport modules, humble 2400MHz, stable at 3200 (with higher timmings).
Testing if 3900 @ 1.4v is stable, so far it is.