AMD Ryzen 5000 Builders Thread

[B-Riz]

	Cores Threads	Base Freq	Turbo Freq	L3 Cache	TDP	MSRP
Ryzen 9 5950X	16c/32t	3400	4900	64 MB	105 W	$799
Ryzen 9 5900X	12c/24t	3700	4800	64 MB	105 W	$549
Ryzen 7 5800X	8c/16t	3800	4700	32 MB	105 W	$449
Ryzen 5 5600X	6c/12t	3700	4600	32 MB	65 W	$299

5600X comes with cooler, the rest do not.

above from https://www.anandtech.com/show/1614...en-3-on-nov-5th-19-ipc-claims-best-gaming-cpu

 

[richierich1212]

I'm in CA so I won't be of any help. Maybe somebody else will volunteer to give you a hand.

Larry could help if he didn’t have an obsession with cheap/older computer parts 😂

 

[repoman0]

Got another 5900x in exchange at MC just now so we will see if I am incompetent or the old CPU!

 

[Justinus]

Got another 5900x in exchange at MC just now so we will see if I am incompetent or the old CPU!

Did you install the original cpu in the correct orientation?

 

[Midwayman]

Did you install the original cpu in the correct orientation?

Pins up, right?

 

[repoman0]

New CPU up and running with RAM at XMP! It booted easily right away. I guess CPU duds exist, that's a first for me. Doing my usual Prime95 Blend stability test to make sure. 12 cores at 4.6 GHz is using a similar amount of power to my 3700X at 4.1-4.2. Not bad at all, I'm impressed.

 

[Justinus]

New CPU up and running with RAM at XMP! It booted easily right away. I guess CPU duds exist, that's a first for me. Doing my usual Prime95 Blend stability test to make sure. 12 cores at 4.6 GHz is using a similar amount of power to my 3700X at 4.1-4.2. Not bad at all, I'm impressed.

It was that nasty glofo IO die, TSMC wouldn't have made a dud like that

 

[repoman0]

It was that nasty glofo IO die, TSMC wouldn't have made a dud like that

The other option is the TUF X570 Pro was a dud and killed the CPU, considering it did boot once or twice before going unresponsive. I bought it for seemingly improved features vs the Prime X570 Pro at a lower price, but meh, I’ll keep my existing known working board and send this one back.

 

[MrTeal]

Well, I finally got shipping notification on my 5900X. It was a wait after I sold my 5800X, but in the end it only ended up costing me an extra CAD$60 over the 5800X, so I think long term it will have been worth it.

Excited to see what temps look like here compared to what I was seeing with my 5800X.

 

[amrnuke]

New CPU up and running with RAM at XMP! It booted easily right away. I guess CPU duds exist, that's a first for me. Doing my usual Prime95 Blend stability test to make sure. 12 cores at 4.6 GHz is using a similar amount of power to my 3700X at 4.1-4.2. Not bad at all, I'm impressed.

Glad it's working well -- I'm very happy my initial suspicion was quite wrong, and also that they had a replacement so quick!

Let us know how it goes, let's see how much rubber you can make that baby burn!

 

[Kenmitch]

Well, I finally got shipping notification on my 5900X. It was a wait after I sold my 5800X, but in the end it only ended up costing me an extra CAD$60 over the 5800X, so I think long term it will have been worth it.

Excited to see what temps look like here compared to what I was seeing with my 5800X.

Custom water? You'll have more cores to load, but at least they'll be split up on two die. I'm sure you'll chime back in with a update once your up and running.

 

[MrTeal]

Yeah, I'm pretty confident the issues I was having reaching Tjmax under ycruncher were due to the density of the 8 core chiplet and individual CPU variance as water temp wasn't an issue at all.

 

[repoman0]

Glad it's working well -- I'm very happy my initial suspicion was quite wrong, and also that they had a replacement so quick!

Let us know how it goes, let's see how much rubber you can make that baby burn!

Thanks! Yeah, they had a stack of like 25 it looked like. Took it back with basically no questions asked. On the other hand, AMD has yet to respond to my warranty request.

I’m weak these days when it comes to tweaking .. most I will do is reapply my ram clocks and see if I can improve sustained boost clocks and reduce power and heat with curve optimizer (after a few weeks of usage at stock to get comfortable that it’s fully stable). This CPU should last a few generations and get me through the initial growing pains of DDR5.

 

[LikeLinus]

I'm about to set out on my 5950x build. I purchased the CPU from another member who was downgrading to a 5900. Not a (+/- $20) mark-up and I was very happy with the deal. I was also able to trade my 1070ti + cash with a member for an 3080 GPU. The build gets interesting from there.

I purchased a Phanteks Eclipse P600S w/ the mini ITX and PCIe 16x cable. Basically I can run two systems in one case.

Build 1A
CPU - AMD Ryzen 5950X
Cooler - Noctua NH-D15
Motherboard - GIGABYTE X570 AORUS Ultra (got this because 3 M.2 slots w/ 16x GPU and 4x for 10GB Card)
RAM - G.SKILL Trident Z Neo (For AMD Ryzen, 1 rank) DDR4 3600 4x16GB = 64GB
GPU - ZOTAC Gaming GeForce RTX 3080 AMP Holo
NIC - Mellanox 10gb Fiber
NVME - I have (3) 1TB Samsung 980
SSD- 4 drives (probably the 1TB Hynix SSD or 3 Hynix and the Crucial 2TB SSD) 2 of them are lost if I use the third M.2, no biggy.

Build 2B
CPU - Intel I7-9700K
Cooler - Corsair H60 or I have two air alternatives (Scythe Big Shuriken 3 or Noctua NH-L12S) so, TBD!
Motherboard - GIGABYTE Z390 I AORUS PRO WiFi
RAM - G.SKILL DDR4 3600 2x16GB 32GB ( acutally have 4 sticks of this, but the mini-ITX only allows for 2 sticks, shame!)
GPU - ASUS GeForce GTX 1060TI 6GB
Storage - (2) 500TB Samsung 960 or possibly the two 1TB WD Blue. Not sure I need speed here.

Extra parts may change depending on space and such.

I have (4) SK Hynix Gold 1 TB SSD, (2) 2TB Crucial Drive, (2) WD Blue 1TB Drive, (2) Crucial MX500 2TB, (2) 500GB WD Blue SSD, plus a couple of 1TB Samsung drives (860? Been awhile since I've looked). I do a ton of video editing and trying to figure out my system. Obviously the 5950x is the main editing machine, but I plan on using the 9700k as an additional render machine or as gaming machine for one of my son when he is here. Gaming would only be when both of my sons are in the room wanting to play games. Otherwise, I may use it to do other things while my main machine is rendering After Effects (can take hours!)

This project will start tomorrow when I have all of the parts. Remaining parts should be delivered today!

EDIT: Totally forgot that I purchased a 55" 4K TCL TV to use as a monitor for what we can see from a real TV. I'm planning on putting it up for review in the same space. Long story, but I have a design idea and it's basically a preview monitor.

 

[KentState]

I really haven't spent time looking at clocks, temps and overclocking. At stock, this is what I'm seeing when running Prime95 for an extended period of time using their heavy workload mix. If I'm seeing mid-40's and 4.4ish all core, how much more can I really get out of it by raising the power limit past 142W PPT? If I can see 5Ghz all-core, then I will try overclocking but if only a 100Mhz more, then I'm fine.

 

[repoman0]

I really haven't spent time looking at clocks, temps and overclocking. At stock, this is what I'm seeing when running Prime95 for an extended period of time using their heavy workload mix. If I'm seeing mid-40's and 4.4ish all core, how much more can I really get out of it by raising the power limit past 142W PPT? If I can see 5Ghz all-core, then I will try overclocking but if only a 100Mhz more, then I'm fine.

View attachment 37364

wtf? How are your temperatures that low? Crazy fancy custom WC loop? Mine hit 75-85 in Prime95 blend with an Arctic 280mm AIO

 

[KentState]

wtf? How are your temperatures that low? Crazy fancy custom WC loop? Mine hit 75-85 in Prime95 blend with an Arctic 280mm AIO

That's what I'm trying to understand. When I run the Prime95 Blend it seems like it's going full out, but the temps are really good. I've also verified the temps in Ryzen Master and they match.

Here is what RM is showing.

 

[Justinus]

That's what I'm trying to understand. When I run the Prime95 Blend it seems like it's going full out, but the temps are really good. I've also verified the temps in Ryzen Master and they match.

Here is what RM is showing.

View attachment 37366

Even with my hampered (missing o-ring) water block, prime95 is not nearly as hot as realbench, cinebench r23 on loop, or blender benchmark. By like 10-20 degrees celsius, depending on which of the 3 you're comparing it to.

This is comparing them with a power unlimited static overclock. With your restrictive PPT of 142W you'll likely see these other benchmarks run at a lower clockspeed at similar temperatures instead.

 

[repoman0]

That's what I'm trying to understand. When I run the Prime95 Blend it seems like it's going full out, but the temps are really good. I've also verified the temps in Ryzen Master and they match.

Here is what RM is showing.

View attachment 37366

Hmm maybe I’m thinking small fft for my temperature figures. I haven’t messed with it enough yet either.

FWIW I’ve read that the best way to get better boost behavior is lowering your voltage curve with the curve optimizer, which reduces power usage and temps and tricks the boost algorithm into higher clocks that way.

 

[Justinus]

Hmm maybe I’m thinking small fft for my temperature figures. I haven’t messed with it enough yet either.

FWIW I’ve read that the best way to get better boost behavior is lowering your voltage curve with the curve optimizer, which reduces power usage and temps and tricks the boost algorithm into higher clocks that way.

What kind of clockspeeds/voltage are you seeing in prime95 to get 75-80C?

 

[repoman0]

What kind of clockspeeds/voltage are you seeing in prime95 to get 75-80C?

I will need to run it again and check, just going by a quick run after I got the second chip to make sure it didn’t BSOD and kill itself like the first. But IIRC prime blend starting off running at 4.6 with temps in the 50s, then must have switched to a smaller fft mode because power and temps shot up. I believe clock speed was 3.7-3.9 at that point. Also possible my fan curve was screwed up because this aio is new to me. I’ve since tweaked them a bit but have really only played CP2077 (which peaked according to hwinfo with 3080 at 71 and 5900 at 75, all cores showing that they had hit 4950 MHz max at some point)

 

[Justinus]

I will need to run it again and check, just going by a quick run after I got the second chip to make sure it didn’t BSOD and kill itself like the first. But IIRC prime blend starting off running at 4.6 with temps in the 50s, then must have switched to a smaller fft mode because power and temps shot up. I believe clock speed was 3.7-3.9 at that point. Also possible my fan curve was screwed up because this aio is new to me. I’ve since tweaked them a bit but have really only played CP2077 (which peaked according to hwinfo with 3080 at 71 and 5900 at 75, all cores showing that they had hit 4950 MHz max at some point)

If you have a bit of time to spare, running hwinfo and then setting up cinebench r23 on multi 30 minute (stability) test would be useful. Reset hwinfo about 10 seconds after you start the test then screengrab it right before the end and post it up here. Looking for core clocks, core/ccd temps, and voltages.

I have to move the temps up to the top so they are visible onscreen with the voltages and clockspeeds.

 

[KentState]

Here is my r23 30 minute run. Capture the last cycle of the run so you can see the actual values while running as well as the max.

 

[repoman0]

If you have a bit of time to spare, running hwinfo and then setting up cinebench r23 on multi 30 minute (stability) test would be useful. Reset hwinfo about 10 seconds after you start the test then screengrab it right before the end and post it up here. Looking for core clocks, core/ccd temps, and voltages.

I have to move the temps up to the top so they are visible onscreen with the voltages and clockspeeds.

Sitting pretty solid at 4.175 to 4.2 GHZ, 1.138-1.15V with the following values under CPU:



edit: scored 20778 in multicore .. need to look up and compare if that's any good!

Tried curve optimizer quickly with default PBO limits (142W, 95A, 140A) and -10 voltage offset and got a quick BSOD for uncorrectable WHEA error, so guess it's not as easy as it seems.

 

[Justinus]

Here is my r23 30 minute run. Capture the last cycle of the run so you can see the actual values while running as well as the max.

[SNIP]

Seems pretty close to in line with my temps when I leave it on auto. Similar frequencies on my 5950x.

Sitting pretty solid at 4.175 to 4.2 GHZ, 1.138-1.15V with the following values under CPU:

edit: scored 20778 in multicore .. need to look up and compare if that's any good!

Tried curve optimizer quickly with default PBO limits (142W, 95A, 140A) and -10 voltage offset and got a quick BSOD for uncorrectable WHEA error, so guess it's not as easy as it seems.

For the lulz I set my 5950x to 4.2 ghz all core and 1.144v (reported under load). On my (hampered) custom loop I'm seeing around 55C averages with peaks closer to 60C.

I think that falls in line with what you'd expect from an AIO... I think. It's honestly hard to tell.

As for score, locked at 4.2 I see 28,225, which would be slightly higher prorated to 12 cores at 21,169. Probably a function of me locking my clocks and your boost fluctuating a bit.

 

[Noid]

I just installed a 5900x bought today.
I have post.
But, on reboot after F10 saving settings in BIOS ...
... goes right back into BIOS screen
( loop )
Never leaves BIOS ... wtf

Tried the last 3 BIOS releases.
( surprised I could flash back )

Temps are good.
All devices present.

any ideas ?