Discussion Ryzen 3000 series benchmark thread ** Open **

[Hitman928]

Active cooling requires using energy to transfer heat away such as is used in thermoelectric cooling (Peltier) or heat pumps.

A fan doesn't use energy?

 

[IEC]

A fan doesn't use energy?

If you consider air the coolant and the fan the pump, sure why not. Doesn't matter to me. But some people might not consider it strictly active cooling.

 

[Fanatical Meat]

A fan doesn't use energy?

It means there's a tiny fan attached to the chipset heatsink. It's not actually active cooling (the computer dorks got that wrong 20 years ago) but that's what they mean.

I've always known this as a form of active cooling. What is your definition of active cooling that a fan to blow heat away from the heatsink doesn't fit?

Active cooling requires using energy to transfer heat away such as is used in thermoelectric cooling (Peltier) or heat pumps.

I was trying not to break this out but I cannot resist.

 

[IEC]

Content creation review video by EposVox and Wendell from L1T:

tl;dw
1) In DaVinci Resolve, the 3900X and 3700X utterly destroy h.264, Cineform and DNxHR renders
2) For content creation workflows, 3900X with less RAM is mostly beating EposVox's i9-7980XE overclocked to the limits with 128GB RAM... to the point he is trying to figure out how to squeeze his PCIe needs into X570 (and can't wait for Threadripper 3K)
3) Single thread in real-world workflows was mostly beating their 9900K setups... and MT workflows are heavily in favor of 3900X

P.S. I don't get why people say Blender or Cinebench R20 are artificial benchmarks. They are real applications used in real-world workflows and they are very representative of real-world performance in those workflows.

 

[Hitman928]

If you consider air the coolant and the fan the pump, sure why not. Doesn't matter to me. But some people might not consider it strictly active cooling.

Maybe you could call it a hybrid solution, but it definitely has an active component. I mean, if it's not active cooling, I don't see how you could consider water cooling active either, unless it's not considered active by those people?

 

[Hitman928]

P.S. I don't get why people say Blender or Cinebench R20 are artificial benchmarks. They are real applications used in real-world workflows and they are very representative of real-world performance in those workflows.

Because their CPU supplier of choice started to lose. There is something to say about the benchmark not being encompassing of every way you may use the renderer, but no benchmark will ever be perfect in that way.

 

[TheGiant]

Content creation review video by EposVox and Wendell from L1T:

tl;dw
1) In DaVinci Resolve, the 3900X and 3700X utterly destroy h.264, Cineform and DNxHR renders
2) For content creation workflows, 3900X with less RAM is mostly beating EposVox's i9-7980XE overclocked to the limits with 128GB RAM... to the point he is trying to figure out how to squeeze his PCIe needs into X570 (and can't wait for Threadripper 3K)
3) Single thread in real-world workflows was mostly beating their 9900K setups... and MT workflows are heavily in favor of 3900X

P.S. I don't get why people say Blender or Cinebench R20 are artificial benchmarks. They are real applications used in real-world workflows and they are very representative of real-world performance in those workflows.

just clicked though it, friend is asking me about upgrade of his 6900K especially for that kind of workload
I see a mixed bag of results with different video cards, combinations of CPU+GPU give a mix of results
I must say I cannot make any conclusion based on this video
Can you point me please?

 

[ElFenix]

Maybe you could call it a hybrid solution, but it definitely has an active component. I mean, if it's not active cooling, I don't see how you could consider water cooling active either, unless it's not considered active by those people?

there's no potential to move heat against the gradient, so, no. mechanical, yes. but it's like the old saying, "fans cool people, not rooms."

 

[Hitman928]

there's no potential to move heat against the gradient, so, no. mechanical, yes. but it's like the old saying, "fans cool people, not rooms."

I cannot find a definition of active cooling that requires to move heat against the gradient. By that definition, liquid cooling is not active either, do you consider it to be active cooling?

Fans can cool rooms, you just need the right setup.

 

[ElFenix]

I cannot find a definition of active cooling that requires to move heat against the gradient. By that definition, liquid cooling is not active either, do you consider it to be active cooling?

Fans can cool rooms, you just need the right setup.

it got difficult to find since every computer person considered fans to be active cooling, so there's billions of hits on google that will tell you that. i'm going by what my physics textbook back in high school said (there was a unit on air conditioners). the fan is just blowing air around, not actually moving heat.

and, as i answered in the previous post, no (though it's been over 20 years since i took high school physics so maybe i'm wrong on that one - i distinctly remember the fan thing because a/c vs fans was the example).

 

[IEC]

I cannot find a definition of active cooling that requires to move heat against the gradient. By that definition, liquid cooling is not active either, do you consider it to be active cooling?

Fans can cool rooms, you just need the right setup.

OT: Place a fan in a sealed container and run it for an hour, measuring the temperature inside the container at start and afterwards. It will actually be higher, because the fan converts electrical energy into mechanical energy + heat. Now, it will feel cooler to us meatbags because of evaporative cooling effects, but it does not remove heat.

Back on topic, elmor @ OCN has been testing memory on X370/X470/X570 with Ryzen 3000 series chips. The results are quite promising:

Summary

• X370/X470 BIOS is still immature when using Ryzen 3000 (depending on the board)
• Highest achieved memory frequency was 4333 MT/s on X570, 4200 MT/s on X470 and 3666 MT/s on X370. The results will most likely get closer with BIOS updates.
• Fabric clock target 1700-1900 MHz
• Memory latency penalty ~10ns when moving from synchronous to asynchronous memory/fabric clock, which can be reduced by further increasing memory frequency
• Highest possible fabric clock with memory synchronized (MCLK = 2*FCLK) yields the highest read bandwidth and lowest latency with 1xCCD
• Write bandwidth only scales with fabric clock with 1xCCD
• Copy bandwidth benefits from both memory and fabric clock, and is not hurt much by increased latency
• 2xCCD read, write and copy bandwidth keeps scaling with both memory and fabric clock without being hurt by asynchronous mode. In this case, it comes down to choosing between lower latency or higher bandwidth

Source:
https://www.overclock.net/forum/13-amd-general/1728878-ryzen-3000-memory-fabric-x370-x470-x570.html

As a bonus, it appears PCIe 4.0 support is not blocked by AMD, and is instead up to the motherboard maker to support. He was able to get Gen4 speeds on an X470 board. I will try to replicate using a Radeon VII if my BIOS will allow it.

 

[Abwx]

the fan is just blowing air around, not actually moving heat.

In the case of a CPU cooler it is blowing an air that is at a temp lower than the cooling surface, so this "fresh" air is moving heat from the cooler and transfering it to the rest of the air in the room, and utimately to the walls of the room...

 

[maddie]

I cannot find a definition of active cooling that requires to move heat against the gradient. By that definition, liquid cooling is not active either, do you consider it to be active cooling?

Fans can cool rooms, you just need the right setup.

What he's trying to tell you is that a fan (or any non-active cooling) can't cool below ambient. It's limited to the surrounding temperature.

Active cooling can, such as refrigeration, peltier, etc.

 

[Hitman928]

OT: Place a fan in a sealed container and run it for an hour, measuring the temperature inside the container at start and afterwards. It will actually be higher, because the fan converts electrical energy into mechanical energy + heat. Now, it will feel cooler to us meatbags because of evaporative cooling effects, but it does not remove heat.

Back on topic, elmor @ OCN has been testing memory on X370/X470/X570 with Ryzen 3000 series chips. The results are quite promising:



Source:
https://www.overclock.net/forum/13-amd-general/1728878-ryzen-3000-memory-fabric-x370-x470-x570.html

As a bonus, it appears PCIe 4.0 support is not blocked by AMD, and is instead up to the motherboard maker to support. He was able to get Gen4 speeds on an X470 board. I will try to replicate using a Radeon VII if my BIOS will allow it.

Put a dome around your house and tell me if your AC unit on your house causes the temperature to go up or down.

 

[Hitman928]

What he's trying to tell you is that a fan (or any non-active cooling) can't cool below ambient. It's limited to the surrounding temperature.

Active cooling can, such as refrigeration, peltier, etc.

I understand what he was saying, but where is that definition of active cooling that requires sub ambient temps? The general definition of active cooling that I've always seen is uses energy to transfer heat. Watering or a fan heatsink do just that. Housing contractors also consider an electric fan as active cooling when talking about attics so it's not just computer people.

 

[Markfw]

Can we stop with the active cooling stuff ? not related to benchmarks.

 

[Asterox]

Content creation review video by EposVox and Wendell from L1T:

tl;dw
1) In DaVinci Resolve, the 3900X and 3700X utterly destroy h.264, Cineform and DNxHR renders
2) For content creation workflows, 3900X with less RAM is mostly beating EposVox's i9-7980XE overclocked to the limits with 128GB RAM... to the point he is trying to figure out how to squeeze his PCIe needs into X570 (and can't wait for Threadripper 3K)
3) Single thread in real-world workflows was mostly beating their 9900K setups... and MT workflows are heavily in favor of 3900X

P.S. I don't get why people say Blender or Cinebench R20 are artificial benchmarks. They are real applications used in real-world workflows and they are very representative of real-world performance in those workflows.

The most important detail is who say or claims, not what they say or claims.

 

[maddie]

I understand what he was saying, but where is that definition of active cooling that requires sub ambient temps? The general definition of active cooling that I've always seen is uses energy to transfer heat. Watering or a fan heatsink do just that. Housing contractors also consider an electric fan as active cooling when talking about attics so it's not just computer people.

Semantics then.

By the way, I never consider facts or truth of the real world to be subjected to democratic principles. As I tell my friends, believe whatever you want but the real world trumps all theory.

 

[misuspita]

For me, passive is something that doesn't use any kind of (electrical/mechanical) energy to cool. Radiators = passive. Peltier, ventilators, AC, LN2 = active. /OT

Are there any reviews with 3800X and 3600X? 3600X I seen only that youtube guy from Tech deals, but none 3800X

 

[DrMrLordX]

Are there any reviews with 3800X and 3600X? 3600X I seen only that youtube guy from Tech deals, but none 3800X

3800x reviews are coming later. Day 1 reviewers got a package with 3900x, 3700x, and I think 3600.

 

[BUTCH1]

Wow, folks at Intel have to be nervous at this point, for arguments sake, let's say most people buy in the mid-range price point and AMD is just killing it here. For $250 you can buy a 6 core/12 thread CPU that wins most benchmarks over Intel, (which cost's $100-125 more). Intel also did itself zero favors by requiring a new MOBO for every refresh as well, it will be interesting in upcoming weeks/months to see if Intel slashes prices to try and stay competitive.

 

[Markfw]

Wow, folks at Intel have to be nervous at this point, for arguments sake, let's say most people buy in the mid-range price point and AMD is just killing it here. For $250 you can buy a 6 core/12 thread CPU that wins most benchmarks over Intel, (which cost's $100-125 more). Intel also did itself zero favors by requiring a new MOBO for every refresh as well, it will be interesting in upcoming weeks/months to see if Intel slashes prices to try and stay competitive.

Cheap is one thing. But when you also have a dead socket (next gen will be a new socket I am sure), it makes it worse.

They are in deep doo-doo PERIOD.

 

[Zucker2k]

The 3800x are being sold right now. Why were they missing from officially sanctioned reviews? I hazard a guess. Same performance as a 3700x and $70 higher, perhaps? Would make the 9900k look better - price-wise.

 

[VirtualLarry]

3800x reviews are coming later. Day 1 reviewers got a package with 3900x, 3700x, and I think 3600.

I think HardwareUnboxed said in their 3600 review, that they had to pay for it with their Patreon money, that AMD might sample them a 3600 later on. AMD's review kit apparently included a 3700X and a 3900X, and a couple of specific X570 mobos. (Loaned?)

 

[Fanatical Meat]

Content creation review video by EposVox and Wendell from L1T:

tl;dw
1) In DaVinci Resolve, the 3900X and 3700X utterly destroy h.264, Cineform and DNxHR renders
2) For content creation workflows, 3900X with less RAM is mostly beating EposVox's i9-7980XE overclocked to the limits with 128GB RAM... to the point he is trying to figure out how to squeeze his PCIe needs into X570 (and can't wait for Threadripper 3K)
3) Single thread in real-world workflows was mostly beating their 9900K setups... and MT workflows are heavily in favor of 3900X

P.S. I don't get why people say Blender or Cinebench R20 are artificial benchmarks. They are real applications used in real-world workflows and they are very representative of real-world performance in those workflows.

Everyone, above is the way to post a video. Simple, easy to follow summary with the imbedded video.
I salute you @IEC

Sorry for off topic compliment.