Ryzen Owners Zen Garden

[TheLostSwede]

Most likely overtightened screws...
Thermal pads have some cushioning that thermal grease lacks...

 

[GoldenX]

^
Hell yeah, managed to get 3400MHz on tha sticks.

 

[biffzinker]

Most likely overtightened screws...
Thermal pads have some cushioning that thermal grease lacks...

The top of the heatsink is not in contact with the chokes though, still a gap between the two. Screws are only hand tight, didn't want to risk crushing the mosfets.

 

[TheLostSwede]

The top of the heatsink is not in contact with the chokes though, still a gap between the two. Screws are only hand tight, didn't want to risk crushing the mosfets.

Right, no idea then, never seen a broken choke before in my 25+ years of building PCs and working in the industry.
I've seen rusty chokes though...

 

[biffzinker]

^
Hell yeah, managed to get 3400MHz on tha sticks.

The switch to the B450 Tomahawk allowed for 3533 MHz from 3466 MHz.

 

[Mussels]

Three things happened today

1. I discovered HWinfo64 crashes the background service for corsairs iCue, which froze my fan speeds and caused thermal/performance issues
2. I discovered multi core enhancement in the BIOS (defaults to on) is a f*cking c*nt, turning it off dropped 20C and RAISED performance
3. I'm so annoyed at 1&2 i feel like merging my sytems together and running the 2700x + watercooled 1080 in the ITX case, leaving the big case to the older hardware

 

[TheLostSwede]

Three things happened today

1. I discovered HWinfo64 crashes the background service for corsairs iCue, which froze my fan speeds and caused thermal/performance issues
2. I discovered multi core enhancement in the BIOS (defaults to on) is a f*cking c*nt, turning it off dropped 20C and RAISED performance
3. I'm so annoyed at 1&2 i feel like merging my sytems together and running the 2700x + watercooled 1080 in the ITX case, leaving the big case to the older hardware

1. They warn about that and asks you if you want to disable the part that interferes with iCue.
2. Care to share a screenshot of that setting? As I haven't seen that option.

 

[Mussels]

Forgot reason for swapping the systems around: 2700x just feels like it belongs on B450 and not x370, and the ITX will fit on the desk vs the ATX on the floor.
This also lets me run several less fans and devices and have an overall less complicated build now that i solved the heat issues (90C linpack load dropped to 60C, since the fans now ramp up and the wattage dropped... so i dont really need a 4x 140mm fan setup to cool the CPU now)

1. The option to not read corsair doesnt work, it takes a full power cycle of the system + a reset of the service to get the device back up and running - it breaks it hard.
See the corsair thread here for the screenshots, rather than duplicating it here

2. I was thinking of the name for my previous intel system, its core performance boost on Gigabyte aourus, like two menus deep in the MIT settings

 

[TheLostSwede]

Forgot reason for swapping the systems around: 2700x just feels like it belongs on B450 and not x370, and the ITX will fit on the desk vs the ATX on the floor.
This also lets me run several less fans and devices and have an overall less complicated build now that i solved the heat issues (90C linpack load dropped to 60C, since the fans now ramp up and the wattage dropped... so i dont really need a 4x 140mm fan setup to cool the CPU now)

1. The option to not read corsair doesnt work, it takes a full power cycle of the system + a reset of the service to get the device back up and running - it breaks it hard.
See the corsair thread here for the screenshots, rather than duplicating it here

2. I was thinking of the name for my previous intel system, its core performance boost on Gigabyte aourus, like two menus deep in the MIT settings

1. Works fine for me, it simply ignores the iCue stuff.

2. We've been discussing the CPB in another thread and it prevents our CPUs from boosting past base clock, so if you can still boost with it disabled, means that there's something going on with certain boards. Have a look see...

Did you get a shiny new Ryzen 3000 CPU?

I have the original SX8200NP and a "last BIOS time" in Task Manager of around ~18 seconds. I previously thought it was a Ryzen 3000 + B450 problem. The fast boot options don't make a difference for me; once Windows starts booting everything is hella fast. You should probably just move it up to...

www.techpowerup.com

 

[biffzinker]

i demand better pics of the damage!

Heatsink, and CPU VRM Mosfets are all clean of the thermal compound. The thermal compound literally turned into a thermal pad.
It was firm with a little bit of stretch, and had the stick of a thermal pad. Half of it came off in a chunk stuck to the bottom of the heatsink.

What a mess though

 

[TheLostSwede]

Heatsink, and CPU VRM Mosfets are all clean of the thermal compound. The thermal compound literally turned into a thermal pad.
It was firm with a little bit of stretch, and had the stick of a thermal pad. Half of it came off in a chunk stuck to the bottom of the heatsink.

What a mess though

Care to share what thermal gunk it was, so the rest of us can avoid it?

 

[biffzinker]

Care to share what thermal gunk it was, so the rest of us can avoid it?

TG-7

TG-7 Thermaltake high performance and reliable Thermal grease with diamond powder

www.thermaltake.com

There's a bit between the syringe, and cap that turned into elastic rubber but sticky. Got it from Best Buy because it was an emergency situation.

 

[GoldenX]

The switch to the B450 Tomahawk allowed for 3533 MHz from 3466 MHz.

3466 boots but isn't stable, so it might be the board.

 

[biffzinker]

3466 boots but isn't stable, so it might be the board.

The B350M Gaming Pro capped at 3400 MHz with the 2600X. Swapping to the 3600 got it up to 3466 on the same board. Haven't messed around with the Tomahawk that much yet. 3600 MHz might be doable since it leaves the primary timings at there default 16-18-18-18-36.

 

[Mussels]

1. Works fine for me, it simply ignores the iCue stuff.

2. We've been discussing the CPB in another thread and it prevents our CPUs from boosting past base clock, so if you can still boost with it disabled, means that there's something going on with certain boards. Have a look see...

Did you get a shiny new Ryzen 3000 CPU?

I have the original SX8200NP and a "last BIOS time" in Task Manager of around ~18 seconds. I previously thought it was a Ryzen 3000 + B450 problem. The fast boot options don't make a difference for me; once Windows starts booting everything is hella fast. You should probably just move it up to...

www.techpowerup.com

iCue *works* but items disappear or control stops until system is rebooted and service restarted - if you dont have a commander pro (or other affected device) everything appears normal.


System moved over to my ITX build with zero issues, other than lack of optical port giving me trouble because my sound bar has a damaged 3.5mm jack... so USB soundcard it is. Luckily still had one.

Plastic side panel is a disapointment after having tempered glass, but it *fits* in the damn room, which is more important.
Watercooled 1080 and 2700x sits at 70c in burntests at 600RPM fan speeds

BIOS between the two brands is so annoying, settings with the same name behave totally different.

CPB on this board controls if it even turbos at all, and it seems like changing almost ANYTHING breaks turbo - this wasnt apparent with the ryzen 1400 i had previously.

 

[biffzinker]

 

[tabascosauz]

@Mussels it appears from the picture that your 2700X is keeping its boost clocks with CPB off. This is very interesting as every Ryzen 3000 complainant I've come across (myself included) has reported that CPB off locks all Ryzen 3000 SKUs to base clock only.

Maybe I'll send Gigabyte an email. It doesn't sound like a Matisse "feature", sounds more like they missed something in their updated BIOSes.

 

[TheLostSwede]

Gigabyte released a "final" stable F5 UEFI for my board and I'm back to 3800MHz memory.

 

[GoldenX]

Got a 16GB Micron E-die kit, managed 3400MHz with it, quite happy, for a 2400MHz kit.

 

[Mussels]

@Mussels it appears from the picture that your 2700X is keeping its boost clocks with CPB off. This is very interesting as every Ryzen 3000 complainant I've come across (myself included) has reported that CPB off locks all Ryzen 3000 SKUs to base clock only.

Maybe I'll send Gigabyte an email. It doesn't sound like a Matisse "feature", sounds more like they missed something in their updated BIOSes.

once i moved the 2700x to the asus b450 it behaves totally different - CPB on the giga board just boosted temps and voltages and nothing else, yet on the asus board (its listed twice as seperate settings, for some reason) it disables turbo.

On the giga board i could also adjust bclk and keep turbo, cant on the asus. annoying.

i should really try a ram OC now, better board and all.

 

[lexluthermiester]

Got a 16GB Micron E-die kit, managed 3400MHz with it, quite happy, for a 2400MHz kit.

That is a great OC for ram! Very nice. Did you need to bump the ram voltage to get there?

once i moved the 2700x to the asus b450 it behaves totally different - CPB on the giga board just boosted temps and voltages and nothing else, yet on the asus board (its listed twice as seperate settings, for some reason) it disables turbo.

On the giga board i could also adjust bclk and keep turbo, cant on the asus. annoying.

i should really try a ram OC now, better board and all.

I think that board makers are doing what they did with the first gen Ryzen, learning. While it's a similar architecture, there are enough differences that they're still having "growing pains". I predict that in a few months most of the issues people are having will be worked out..

 

[TheLostSwede]

I think that board makers are doing what they did with the first gen Ryzen, learning. While it's a similar architecture, there are enough differences that they're still having "growing pains". I predict that in a few months most of the issues people are having will be worked out..

Most of the issues I was having seems to have been more or less worked out. However, I do still believe people are expecting more frequent and higher boosts, as well as "proper" manual overclocking, but we might never see those.

 

[GoldenX]

That is a great OC for ram! Very nice. Did you need to bump the ram voltage to get there?

1.39v, nothing radical.

 

[Mussels]

tried a variety of OC options, even with temps hitting 90C i got like a 3% boost

Back to stock settings with an undervolt and calling it a day, AMD really did a top notch job with their XFR settings

 

[lexluthermiester]

1.39v, nothing radical.

Perhaps. Personally wouldn't be comfortable with that much voltage long term, but if it works stable, rock on!