AMD Ryzen 3000 Builders Thread

[Fanatical Meat]

Thank you for the response but I think I identified what's causing it. Closing the NZXT CAM software makes it behave normally in long idle conditions. Tried it couple of times to verify. Could switch the CAM off altogether, but I need it for the cooler to work properly. Damn.

I almost picked up a NZXT cooler but was scarred because of the complaints about the software for it.
I read there is a open source program to use for NZXT coolers. I don’t have the link but I found it doing some google research regarding CAM software.

 

[Chicken76]

Thank you for the response but I think I identified what's causing it. Closing the NZXT CAM software makes it behave normally in long idle conditions. Tried it couple of times to verify. Could switch the CAM off altogether, but I need it for the cooler to work properly. Damn.

Which NZXT cooler do you have?

 

[lixlax]

Which NZXT cooler do you have?

X61

 

[Kenmitch]

I finally bought and installed a 3700x to replace my 1700. My overall impression is that it's a super nice chip. Everything is nice and snappy. I'm having the issue where I'm not hitting max boost even on a single core. I'm not surprised because it looks like the Asus version of BIOS 1.0.0.3AB for x370 was rushed. There are no less than 3 different spots where you can change the same PBO setting. Even with PBO enabled I've only seen 4.275 max. With PBO disabled the chip hits its 88w PPT limit and is maybe 50-100 mhz less on ST and MT than with PBO enabled. I'm optimistic AMD can fix this with a new BIOS.

You'll probably get a new uEFI pretty soon. Asus just dropped the 1.0.0.3abb for my C6H about a week ago.

On both of my 3xxx combos I've found that messing with PBO settings in the uEFI is a waste of my time. I just leave them on the defaults and use Ryzen Master to play around with PBO and all core overclocks.

Have you tried just using Ryzen Master to enable PBO?

On both of my combos enabling PBO in Ryzen Master only gets me a higher sustained all core clockspeed in the end. Both my 3600 and 3700x combos don't look to have max boost issues. 3600 hits 4.2 GHz and the 3700x hits 4.375 GHz.

 

[Markfw]

You know I have been fighting for days to gewt my 3rd 3900x rinning in an x370 board. Know what thre problem was ? AX860i PSU bad, one of my most recent PSI's. No more of the i series, I have a 1200i and now an 860i dead.

Currently testing 4100@1.2 vcore.

 

[DrMrLordX]

Thank you for the response but I think I identified what's causing it. Closing the NZXT CAM software makes it behave normally in long idle conditions. Tried it couple of times to verify. Could switch the CAM off altogether, but I need it for the cooler to work properly. Damn.

Ack. I guess you don't want that thing running full-bore all the time, do you? They can be a bit loud.

You know I have been fighting for days to gewt my 3rd 3900x rinning in an x370 board. Know what thre problem was ? AX860i PSU bad, one of my most recent PSI's. No more of the i series, I have a 1200i and now an 860i dead.

That's a lot of dead Corsairs. At least their RMA service is still good.

 

[Markfw]

Ack. I guess you don't want that thing running full-bore all the time, do you? They can be a bit loud.



That's a lot of dead Corsairs. At least their RMA service is still good.

To be honest, the 860i was a refub bought on 1/15/2019, so its not under warranty.

No more refurbs

BYW, it just rebooted at 1.175 vcore, I think 1.2 is it.

 

[VirtualLarry]

Yeah, refurb PSUs is not something I dabble in. And refurb HDDs barely. I happily buy open-box / refurb AMD mobos though, and networking gear.

 

[Markfw]

To be honest, the 860i was a refub bought on 1/15/2019, so its not under warranty.

No more refurbs

BYW, it just rebooted at 1.175 vcore, I think 1.2 is it.

BTW, of possible intrest. Both of these 2 CPU's are essentially working stock (it was runing 4050-4080) But one possible difference, the x470 mobo (a taichi) has an 8 pin and a 4 pin cpu power. The x370 only has the 8 min.

I have said it before, and now believe even more firmly, if it has both use both, they are there for a reason. In my case undervolting works better with the 2. I realize there are a lot of other variables, but 1 tenth of a volt, it 10% more vcore required. Anyway, I think its settled in a 1.2 vcore, and the one I am typing one is at 1.1 vcore.

 

[ClockHound]

BTW, of possible intrest. Both of these 2 CPU's are essentially working stock (it was runing 4050-4080) But one possible difference, the x470 mobo (a taichi) has an 8 pin and a 4 pin cpu power. The x370 only has the 8 min.

I have said it before, and now believe even more firmly, if it has both use both, they are there for a reason. In my case undervolting works better with the 2. I realize there are a lot of other variables, but 1 tenth of a volt, it 10% more vcore required. Anyway, I think its settled in a 1.2 vcore, and the one I am typing one is at 1.1 vcore.

That's it - you can stop pitching. Just bought the Taichi X470. ;-)

Will wait another month week or two for the 3900x/3950x supply situation to stabilize before buying a discounted 3800X.

 

[Despoiler]

You'll probably get a new uEFI pretty soon. Asus just dropped the 1.0.0.3abb for my C6H about a week ago.

On both of my 3xxx combos I've found that messing with PBO settings in the uEFI is a waste of my time. I just leave them on the defaults and use Ryzen Master to play around with PBO and all core overclocks.

Have you tried just using Ryzen Master to enable PBO?

On both of my combos enabling PBO in Ryzen Master only gets me a higher sustained all core clockspeed in the end. Both my 3600 and 3700x combos don't look to have max boost issues. 3600 hits 4.2 GHz and the 3700x hits 4.375 GHz.

I have not tried using Ryzen Master. I'll see if that works. Ryzen Master does report the board limits have been raised, but the chip isn't really taking advantage of it like I'd expect. Thanks!

 

[Despoiler]

I have not tried using Ryzen Master. I'll see if that works. Ryzen Master does report the board limits have been raised, but the chip isn't really taking advantage of it like I'd expect. Thanks!

@Kenmitch no dice with Ryzen Master. If I enable PBO the program keeps the chip's 88w PPT value while extending the other two values per the mobo capability. I've uninstalled, reinstalled, etc. The feature definitely does work from the BIOS. I get 150-200mhz more all core boost with it enabled.

 

[lopri]

It's my first day of using MSI X470 Carbon, and the board is already doing 1933 fclk, synced with uclk. It is rather incredulous to me at the moment how easy it is. On Gigabyte X570 I have never managed to boot at a higher than 1900 fclk.



The board did not come equipped with a proper BIOS for Matisse so I had to acquire a 2200G to flash it. It was not a bad experience over all and I am also impressed by how capable the 2200G appears to be. I am looking forward to exploring MSI BIOS from here on.

 

[lopri]

Please update when you know. I’d like a 32GB system with a possible upgrade path to 64GB
Also from my amateurish poking around it appears ASROCK boards are best for all 4 dims being populated. I think it has to do with the T topology thing.

To make a long story short, I can run 64 GB on Gigabyte X570, albeit at reduced max frequency and higher voltage. Obviously that was expected, but the voltage required at high frequency is rather uncomfortable and unsatisfactory. For instance, a pair of sticks do 1900MHz/CL16 at 1.38V and another pair do it at 1.40V. But to run all 4 of them I had to give 1.44V for 1800MHz/CL16. And it is hard to boot into 1800 uclk/mclk, even though it does 1733 fairly predictably at around 1.40V. So in a way 1800 (fclk/mclk/uclk) with 4 sticks behaves like 1900 with 2 sticks on this board, and requires higher voltages.

Strangely I did not need more voltage for the memory controller. Higher vSOC did not seem to do anything so I pretty much stayed within 1.02~1.05V for all the stable configurations.

Here is a Geekbench run with 64GB.
https://browser.geekbench.com/v4/cpu/14489293

 

[Fanatical Meat]

To make a long story short, I can run 64 GB on Gigabyte X570, albeit at reduced max frequency and higher voltage. Obviously that was expected, but the voltage required at high frequency is rather uncomfortable and unsatisfactory. For instance, a pair of sticks do 1900MHz/CL16 at 1.38V and another pair do it at 1.40V. But to run all 4 of them I had to give 1.44V for 1800MHz/CL16. And it is hard to boot into 1800 uclk/mclk, even though it does 1733 fairly predictably at 1.40V. So in a way 1800 (fclk/mclk/uclk) with 4 sticks is like 1900 with 2 sticks on this board, with higher voltages.

Strangely I did not need more voltage for the memory controller. Higher vSOC did not seem to do anything so I pretty much stayed within 1.02~1.05V for all the stable configurations.

Thanks for the update. Please give another when you test again.
Going to be doing my build by end of month.

Was 3200 memory speed possible with 64GB?

Apologies if above sounds stupid, I get confused is 1900mhz ram = ddr3800 or ddr1900 or ddr2400.

 

[lopri]

Yes. DDR4-3200/CL14 with all 4 sticks is possible at or below 1.35V.

 

[lopri]

Already matching the best latency I've seen on the Gigabyte board. Holy cow. I have yet to touch the timings yet!

 

[Fanatical Meat]

Already matching the best latency I've seen on the Gigabyte board. Holy cow. I have yet to touch the timings yet!

View attachment 10631

Excellent, I’m particularly interested in the B450 board. Thing I can’t figure out is the X570 boards appear to have better memory support.

 

[DisEnchantment]

I have been having issues with my Corsair H110i GT and H100i Pro, my X570 board is not detecting the CPU fan speed.

1. The H110i would start with the fan always not running at all, although the pump is running since the CPU is cool.
2. The H100i Pro would not even work and the Cooler head starts blinking rapidly.
3. The Seasonic PSU use to trip when I connect the AIO and the SSD.
4. I power the pump and the AIO fans from single sata power cable. Is this OK? the AIO has a splitter cable for the AIO radiator fans.
5. Is the Motherboard supposed to detect the AIO fan speed (not radiator fan speed)?

I switched the coolers thinking that it is an AIO problem, but both the coolers have the same problem on the X570 but not on my B350 board.
I thought this was a PSU problem and I switched the PSU from Seasonic 750W to CoolerMaster 700W PSU but same issue.
The Board cannot detect the CPU fan. Any ideas.
I contacted Seasonic for RMA and also Corsair for RMA but I am not sure these are having the issues..
I wonder if there is an issue with the SSD or the X570 or the PSU or the AIO.

1. The H100i Pro and the H110i v2 has two splitter for the Radiator fans. These fans and the Pump are normally powered from the same SATA cable. This is how I power them on my B350 system
2. Connecting the Pump CPU fan header to any of the CPU header or sys fan header on the Motherboard does not help the issue.
3. Connecting the Pump fans directly to the motherboard of course works, is this how is it supposed to be done? In this case what is the use of the cpu fan header of the AIO.
4. Is it due to the the fact that the sata is powering two fans from the splitte on the AIO connector that the PSU is tripping?

Anyone can help me,

CPU is 3400G placeholder for 3900X (which has not arrived yet).

 

[Topweasel]

This is my configuration.

H100i.

One split header to 2 PWM fans, plugged into the Pump.
Motherboard header plugged into CPU Fan1 /WP (for Water Pump).
Sata power plugged into a sata connector (same one my spindle drive).

Read the instructions but I am pretty sure the Mobo connector and Sata connectors need to connected. The Mobo one I think drives the pump and the Sata one the upwards of 4 fans.

 

[DrMrLordX]

But to run all 4 of them I had to give 1.44V for 1800MHz/CL16.

I'm feeding 1.46v to my DDR4-4400 sticks to get them to run DDR4-3733 14-16-14-28. That ain't too bad.

 

[IEC]

I have had no issues running 1.0375V SOC for 3733+ memory speeds. Might even be able to go lower.

 

[Chicken76]

1. I power the pump and the AIO fans from single sata power cable. Is this OK? the AIO has a splitter cable for the AIO radiator fans.

Don't do that. Pumps usually draw more current than fans, and on top of that you're also powering multiple fans from the same connector. Split them up. Your motherboard might have a dedicated pump connector, check the specs/manual.

 

[DisEnchantment]

Don't do that. Pumps usually draw more current than fans, and on top of that you're also powering multiple fans from the same connector. Split them up. Your motherboard might have a dedicated pump connector, check the specs/manual.

Too late . That was the problem. I dont know why the AIO came with the splitter. Fried the RGB on the H110i and the Pump is no longer accessible via USB.

Apparently on my older rig, the 550 W PSU cannot trigger this issue.
With the new rig the 750W PSU was tripping sometimes and it seems to be due to plugging the Radiator Fans to the AIO splitter cable which is powered by from a single SATA connector.
I used ML140 fans and SP140 fans, did not know those can draw that much current.
I fried the RGB on the H110i. But apparently the pump works and I just need to connect the pump to the SATA connector and the Radiator fans to other headers, the pump can no longer power the fan via the splitters.

With the H100i Pro, thankfully everything now works, I connect the PWM fans to pump headers on the X570 and the SATA to the Cooler pump. And it is better this way, I can monitor all the fans and cooler speeds.
Also the SATA SSDs are powered from other Cables not the same one.
Now the temps are low 30s degrees Celsius
Now monitoring the system if issue appears. FIngers crossed.

 

[Kenmitch]

3950x on LN2

5.2GHz Ryzen 9 3950X Benchmarks Blitzes the competition - OC3D

5.2GHz Ryzen 9 3950X Benchmarks Blitzes the competition AMD’s Ryzen 3rd Generation processors are coming, offering up to twelve cores at launch and up to 16 cores from September onwards, delivering compelling performance levels over a wide range of price points. In September, AMD plans to...

www.overclock3d.net