Post your Cinebench R23 Score

[Nordic]

It was run at stock with memory XMP speeds of 3600 Cas 16. By my observation it had a sustained 3.6ghz during multi and 4.8ghz during single. I am on air cooling.

 

[Tomgang]

It was run at stock with memory XMP speeds of 3600 Cas 16. By my observation it had a sustained 3.6ghz during multi and 4.8ghz during single. I am on air cooling.

Seems with in spec for at stock 5950X. I have an aircooled 5950X as well. Just with a Noctua NH-D15 aircooler.

Can i ask your single and max multi core temp since not so many runs these chips on air? Mine was 72C single core and 55C multi core fully loaded at stock. With my 4.65 GHz all core manuel oc at 1,375 volts it hits 84C in cinebench R23 with out any power limits.

Here is my stock score and PBO tuned score with the higest power limit i was alowed by motherboard to set.

Post your Cinebench R23 Score

Artic liquid freezer II 420 Ryzen 9 5900x - OC per CCX - CCX1@4675mhz, CCX2@4550mhz 1.25V 32GB DDR43600 cl16- B-die - OC@3733 cl14

www.techpowerup.com

Same setup but here i use asus dark hero party trick called Dynamic oc switch. Enable it to switch automatic betweem PBO where i can by pass the normal max 200 mhz+ and force even higher single core clocks and a manuelt all core OC with out power limits to hold it back. Still on air cooling. That alows me to get the higgest possible single core and multi core score at the same time. Hence why this has the higest single and multi core score.

Post your Cinebench R23 Score

Artic liquid freezer II 420 Ryzen 9 5900x - OC per CCX - CCX1@4675mhz, CCX2@4550mhz 1.25V 32GB DDR43600 cl16- B-die - OC@3733 cl14

www.techpowerup.com

 

[DemonicRyzen666]

Just over 12,000 on a 5600x at 4.7ghz all core manual overclock . single thread lags behind On corsair H150i

and here is PBO auto overclock about the same max all core 4.7ghz just single core is 4.85ghz

 

[tabascosauz]

View attachment 211526

Just over 12,000 on a 5600x at 4.7ghz all core manual overclock . single thread lags behind On corsair H150i

and here is PBO auto overclock about the same max all core 4.7ghz just single core is 4.85ghzView attachment 211530

Running a lot of background tasks? That's a lot of clock stretching. Sub-1500 is pretty low for what's supposed to be a 4.7GHz overclock. And 4.8GHz should be just under or at the 1600 mark.

 

[DemonicRyzen666]

Running a lot of background tasks? That's a lot of clock stretching. Sub-1500 is pretty low for what's supposed to be a 4.7GHz overclock. And 4.8GHz should be just under or at the 1600 mark.

well I'm only running infinity fabric at 1633mhz. I have not really tried to over clock my ram as it's only 3200mhz ram.
My temps aren't above 75C on all core and only 53C on single thread.

 

[Nordic]

Seems with in spec for at stock 5950X. I have an aircooled 5950X as well. Just with a Noctua NH-D15 aircooler.

Can i ask your single and max multi core temp since not so many runs these chips on air? Mine was 72C single core and 55C multi core fully loaded at stock. With my 4.65 GHz all core manuel oc at 1,375 volts it hits 84C in cinebench R23 with out any power limits.

The stock performance is good. I haven't even begun overclocking. Ryzen ClockTuner suggests I could run an all core overclock of 4575mhz to 4700mhz at 1.35. I haven't even messed with PBO settings yet. I am not sure what amount of tuning I want to do yet.

Ryzen master shows me holding 63C after 5 minutes on Single Threaded. It seems I am running even cooler on a multi threaded workload.

http://imgur.com/Ora703H

 

[tabascosauz]

well I'm only running infinity fabric at 1633mhz. I have not really tried to over clock my ram as it's only 3200mhz ram.
My temps aren't above 75C on all core and only 53C on single thread.

Shouldn't matter. You could run at 2133 JEDEC if you wanted and the difference on Ryzen in Cinevench should be like 10pts. Memory makes zero difference in Cinebench for Ryzen. It's only the ST score that looks out of place, which is why I asked if you're running stuff in the background. The ST test is so long that anything that runs during it will have an effect, even HWInfo.

 

[Tomgang]

The stock performance is good. I haven't even begun overclocking. Ryzen ClockTuner suggests I could run an all core overclock of 4575mhz to 4700mhz at 1.35. I haven't even messed with PBO settings yet. I am not sure what amount of tuning I want to do yet.

Ryzen master shows me holding 63C after 5 minutes on Single Threaded. It seems I am running even cooler on a multi threaded workload.

http://imgur.com/Ora703H

Nope. What you are running there is multi core stress test. CPU full load hence why you are only at 3.6 GHz. You are actually running 6C-7C higher than i am in multi core test. I hit 55C max there.

To run single core test. Use the test below under the test you run in the screenshot. That is single core test. It will hit higher boost clock but also get even hotter.

 

[Nordic]

Nope. What you are running there is multi core stress test. CPU full load hence why you are only at 3.6 GHz. You are actually running 6C-7C higher than i am in multi core test. I hit 55C max there.

To run single core test. Use the test below under the test you run in the screenshot. That is single core test. It will hit higher boost clock but also get even hotter.

I only shared the screenshot for the multicore run. The single threaded had a max temp of 63c. Single threaded holds a sustained ~4.8ghz during cinebench.

 

[Tomgang]

I only shared the screenshot for the multicore run. The single threaded had a max temp of 63c.

ah i see. will run my own test again to compare. then. i will be back.

 

[Nordic]

ah i see. will run my own test again to compare. then. i will be back.

I turned on PBO and set it to the motherboard limits, leaving cpu voltage on auto. Single threaded heat is now about 65c while multi threaded is about 83c.

Single threaded:

http://imgur.com/l38O2ar

Multi-threaded:

http://imgur.com/mIcVkIE

 

[freeagent]

I should plug my 5600X back in. those are all core clocks.. and it doesn't really like 4 sticks that much.

 

[Tomgang]

I turned on PBO and set it to the motherboard limits, leaving cpu voltage on auto. Single threaded heat is now about 65c while multi threaded is about 83c.

Single threaded:

http://imgur.com/l38O2ar

Multi-threaded:

http://imgur.com/mIcVkIE

back with my own test.

Stock single is a round 60C

Multi stock is a round 53C

With PBO single core is pretty much the same at 60C

PBO multi is a round 76C

So i do have better temp over all.

 

[Nordic]

back with my own test.

So i do have better temp over all.

As you should. You have the better air cooler. I think you have 8 heat pipes to my 6. Your fans go up to 3000 rpm while mine go to 2000 rpm.

 

[Tomgang]

As you should. You have the better air cooler. I think you have 8 heat pipes to my 6. Your fans go up to 3000 rpm while mine go to 2000 rpm.

NH-D15 is actually only a 6 heat pipe cooler. But it does however have two towers and 2 fans. Also it can´t be denied either that my high RPM fans does a thing or two on temp as well.

But it´s nice to get comfirmed that the ekstra money i spend also gets me lower temp.

 

[Nordic]

NH-D15 is actually only a 6 heat pipe cooler. But it does however have two towers and 2 fans. Also it can´t be denied either that my high RPM fans does a thing or two on temp as well.

But it´s nice to get comfirmed that the ekstra money i spend also gets me lower temp.

I should have gotten a noctua heatsink to begin with. They are expensive but the fans are that much better than the competition. I got my Scythe because it had a good performance per dollar, but then upgraded to Noctua fans months to get the performance I wanted. The total cost of the Scythe and Noctua fans was the same as just getting a noctua heatsink in the first place.

This is with PBO on set to motherboard limits. Same memory settings. Still on air. Multithreaded ran at about 4.1ghz. Not sure why I am getting lower single threaded scores, although not by much. A 12% bump in multi threaded is decent.

 

[tabascosauz]

I should have gotten a noctua heatsink to begin with. They are expensive but the fans are that much better than the competition. I got my Scythe because it had a good performance per dollar, but then upgraded to Noctua fans months to get the performance I wanted. The total cost of the Scythe and Noctua fans was the same as just getting a noctua heatsink in the first place.

This is with PBO on set to motherboard limits. Same memory settings. Still on air. Multithreaded ran at about 4.1ghz. Not sure why I am getting lower single threaded scores, although not by much. A 12% bump in multi threaded is decent.

On a lot of Ryzens if you just adjust (increase) power limits alone via PBO, you might see a bit of a ST dip, if you don't boost ST by other methods like CO. Not sure why it is, but that's how it's been since my 3700X.

We seem to be pretty close on both ST scores and temps.

Mugen 5 is an admirable cooler, just not in the same class as D15. Hard to expect a single tower to match a dual tower ever. I mean, you also get to enjoy the nice clearance all around - D15 compatibility is hilarious by comparison trying to play nice with even average height RAM and PCIe slots.

U12A looks like a good competitor though, might come out ahead of the Mugen, but more expensive than D15 usually

 

[Nordic]

On a lot of Ryzens if you just adjust (increase) power limits alone via PBO, you might see a bit of a ST dip, if you don't boost ST by other methods like CO. Not sure why it is, but that's how it's been since my 3700X.

We seem to be pretty close on both ST scores and temps.

Mugen 5 is an admirable cooler, just not in the same class as D15. Hard to expect a single tower to match a dual tower ever. I mean, you also get to enjoy the nice clearance all around - D15 compatibility is hilarious by comparison trying to play nice with even average height RAM and PCIe slots.

U12A looks like a good competitor though, might come out ahead of the Mugen

Just upgrading to dual Noctua's on my Mugen gave me a 7c drop under full load when I was on a 3900x. It makes a difference.

 

[Tomgang]

I should have gotten a noctua heatsink to begin with. They are expensive but the fans are that much better than the competition. I got my Scythe because it had a good performance per dollar, but then upgraded to Noctua fans months to get the performance I wanted. The total cost of the Scythe and Noctua fans was the same as just getting a noctua heatsink in the first place.

This is with PBO on set to motherboard limits. Same memory settings. Still on air. Multithreaded ran at about 4.1ghz. Not sure why I am getting lower single threaded scores, although not by much. A 12% bump in multi threaded is decent.

I was never in any dout what cooler to get. But i dit everything i cut do to get the best possible aircooling setup. Besides NH-D15 and using Noctua industrial IPPC 3000 RPM fans, i also use Thermal Grizzly Kryonaut extreme cooling paste.

I think i know why your single core gets lower with PBO on. You said it raised by 2C. Zen 3 boost actually funktion much like GPU boost. Besides power limits, boost clocks is also desided by temp. So the higher temp your cpu runs at, the lower boost clock you get and that has a negative effect on your single core clock and single core score. Your multi score score is lower than mine because your board seems only to allow for lower power limits. Compare yours to mine. Specially the two last limits cpu in R23 and those two are higher on my board than yours.

So a combination of lower temp and higher PBO power limits gave med a multi score of 29512.

 

[tabascosauz]

Just upgrading to dual Noctua's on my Mugen gave me a 7c drop under full load when I was on a 3900x. It makes a difference.

I didn't say it didn't?

That's more a function of the fact that Scythe's fans are almost always terrible because they always lack rpm.

 

[Nordic]

I am intending to get back into water cooling this winter. I have one of those 80mm thick 240mm radiators. I think I am going to get a 480mm something though. Watercooling can only do so much though.

 

[freeagent]

There are plenty of good coolers out there, don’t stress over a brand

 

[Tomgang]

There are plenty of good coolers out there, don’t stress over a brand

No we want a all out brand flame war

Just like the Intel vs. AMD and AMD vs. Nvidia fanboys have. Just with coolers

 

[Nordic]

There are plenty of good coolers out there, don’t stress over a brand

There are plenty of great coolers. I was planning on buying a Noctua but decided to instead get a more budget option. I believe this was a mistake because the cost of the Scythe cooler and the noctua fans I later purchased for it ended up costing as much as the expensive noctua but with less performance than I would have had. I should have just bought the Noctua or other equivalently performing cooler.

 

[tabascosauz]

I am intending to get back into water cooling this winter. I have one of those 80mm thick 240mm radiators. I think I am going to get a 480mm something though. Watercooling can only do so much though.

If you pull out all the stops (Kryonaut/H2/SYY/KPX paste, good CO curve to bring down ST temps, well balanced power limits, good Ryzen-optimized block), wouldn't be surprised if it both runs appreciably cooler and performs better. Even a single thick 280mm (or was it 240mm?) gets impressively close to handling a 5950X and 3090. So if you have less GPU, or two rads, I'd expect v good results.

When these Zen 3 cores push above 15W per-core in ST they do some crazy temperature shit, like transient 10-15C spikes. Games like BFV and MW19 are pretty challenging for even D15 for that reason, not MT or pure wattage. Be curious to see what watercooling can do for that.