Anybody planning a Ryzen build?


Recommended Posts

It seems the issues with Ryzen are that the IOMMU groupings are not that nice.  What determines these groupings?... CPU, Motherboard, Motherboard Chipset, BIOS, combiation of any of those?

 

I'm wondering if this is something that can be "fixed" with a newer bios.  Like @Pauven mentioned about @Donach's post here: (https://forums.lime-technology.com/topic/55150-anybody-planning-a-ryzen-build/?page=13#comment-553895) ... if it's a function of the chipset would any motherboard with the same chipset as that one be good for unRAID?

 

Link to comment
1 hour ago, eric.frederich said:

It seems the issues with Ryzen are that the IOMMU groupings are not that nice.  What determines these groupings?... CPU, Motherboard, Motherboard Chipset, BIOS, combiation of any of those?

 

I'm wondering if this is something that can be "fixed" with a newer bios.  Like @Pauven mentioned about @Donach's post here: (https://forums.lime-technology.com/topic/55150-anybody-planning-a-ryzen-build/?page=13#comment-553895) ... if it's a function of the chipset would any motherboard with the same chipset as that one be good for unRAID?

 

 

It definitely can be patched, but for the moment acs override with the multi function option seems to get everything seperated.

Link to comment
13 hours ago, Bureaucromancer said:

And in further weirdness it seems in deleting my pinning I accidently set the vm to a single core and that it works with such.  It does not with the full 12 cores I usually have, experimenting now to see if I can track a hard limit down.

 

Another edit:  what I thought was a hard lock to 100% cpu may not actually be such.  I'm letting it sit witht he pegged cores at the moment on the full cpu set I want, having found that the time to get through the windows splash screen drastically increased every time I increased the core count.  The weird thing is that the progres spinner is moving now and then, so it looks like SOMETHING is happening behind all that CPU activity and its somehow related to number of cores in play.  I almost wonder if it's something odd about handling of the CCX units on Ryzen now?

 

I wouldn't think that pinning/unpinning the cpus would effect whether or not you could boot the vm.  Also, are you isolating any cpus for the host?  If so then you need use pinning, otherwise the vm will try and run all cpus on the host cpu.  For instance if you isolcpus=1-15, then you have specify via pinning which cores between 1-15 you'd like the vm to use, otherwise the system will try to run all vm cores on cpu 0 which is the only core "visable" to the system.  Imagine 12 cores trying to use one core.  This would be a problem and might be causing the slower boot times the more cores you add.  The only reason I have the cpu's unpinned is testing out the performance differences.  Also, are you using iommu=pt and kvm.ignore_msrs=1?  These helped me with win10 boot issues I was having (I think).  

 

 

Link to comment
3 hours ago, eric.frederich said:

It seems the issues with Ryzen are that the IOMMU groupings are not that nice.  What determines these groupings?... CPU, Motherboard, Motherboard Chipset, BIOS, combiation of any of those?

 

I'm wondering if this is something that can be "fixed" with a newer bios.  Like @Pauven mentioned about @Donach's post here: (https://forums.lime-technology.com/topic/55150-anybody-planning-a-ryzen-build/?page=13#comment-553895) ... if it's a function of the chipset would any motherboard with the same chipset as that one be good for unRAID?

 

 

Earlier this month, Tom's Hardware held an AMA (Ask Me Anything) with some folks from AMD.  The IOMMU grouping issue was discussed, but not much was thrown our way:

 

Presentato: As someone interested in doing virtualization and PCI pass-through of a GPU early reports of IOMMU groupings don't look promising for the consumer motherboards. Is that something AMD can address or are any improvements reliant on motherboard manufacturers?

 

DON WOLIGROSKI: This is something I've personally started to look at recently as a pet project. I'm playing with VM-Ware on my Ryzen system at home because, really, Ryzen's highly-threaded CPUs bring a lot of virtualization potential to the table in price segments where it hasn't been before. The sub-$300 segment has been limited to 4-thread processors on the Intel side, while Ryzen 5 ratchets that up to 12 threads. Boom.

 

Having said that, we're in launch mode right now, and virtualization isn't a top priority at the moment. We're laser focused on making the platform as fast as we can in the near future. I anticipate we'll look harder at virtualization as time goes on.

 

I guess this doesn't really answer the question about how much the BIOS is involved in dividing up the IOMMU groupings.  Anyway, at the very least, we're not the only ones asking this question, and AMD now knows it's an open issue.

 

Link:  http://www.tomshardware.com/reviews/amd-ryzen-ama,5018-2.html

 

- Bill

 

Link to comment
13 hours ago, GViz said:

Congrats on the long up time!

 

Thanks!  It so happens to be that I'm an hour away from reaching 21 days... blackjack!

 

With 21 days of up-time under my belt, I do have a few oddities to report.  I don't know which are Ryzen specific, which are 6.3.3 specific, and which are caused by plug-ins.  I figured I'd share them all here for review.

 

 

1)  The Case of the Hanging Console:  The main console (monitor/keyboard physically attached to the server) hangs shortly after boot.  While I can still Telnet into the server, and start up multiple Screen sessions, the main console remains hung 3 weeks later.  Now, this is purely a guess, but I'm starting to think that the hung console might be causing one or more of the issues described below.  I'm thinking that the hung console may actually be a hung process that was executing in the console, and when the process hangs perhaps it causes other processes to be blocked from running.

 

I know other Ryzen users have experienced the hanging main console, and I've only seen it reported on Ryzen.  If anyone has a different experience, please share.

 

I honestly didn't care too much about the hung console since I could still Telnet, but in light of the other anomalies I'm experiencing, I'm thinking the this is a bigger issue than I first surmised.

 

 

2)  The Case of the Reappearing Cache Drive:  In my previous build, I had a Samsung 840 1TB SSD installed as my cache drive.  In my upgrade to Ryzen, I installed a faster Samsung 960 1TB M.2 drive.  During this time, I left the old 840 installed and unassigned.  But on every server reboot, unRAID reassigns the 840 as the cache drive (and unassigns the 960).  I have to stop the array, change the assignment, and restart the array every time I reboot to get the 960 back into the cache drive role.  

 

Because I had so many crashes, I made a point of correcting this assignment and then rebooting the array before it crashed (thinking the change was somehow getting lost due to a crash), but this made no difference.  

 

Just a bit ago I physically removed the 840 from the server, but I haven't rebooted since then, so I don't know if this will allow the 960 to come up as cache drive.  I'm also a bit concerned that even if I fix it for now, if I ever re-install the 840 that this bad behavior will return.

 

 

3)  The Case of the Unfinishable Movie:  While it might just be a bad ISO, one movie being streamed from the server kept hanging near the end at various spots.  Now, it just so happens that this is the only movie I've streamed in the past 3 weeks, so I don't know if the issue is restricted to the single movie, or if it will affect more movies.  So, a single bad ISO doesn't seem to be such a big deal, right?  I had completely dismissed it until I saw my log file later with the following message at about the same time as the movie was hanging:

 

Apr 13 14:09:53 Tower shfs/user: share cache full
Apr 13 14:10:00 Tower shfs/user: share cache full
Apr 13 14:10:00 Tower shfs/user: share cache full
Apr 13 14:10:00 Tower shfs/user: share cache full
Apr 13 14:22:49 Tower shfs/user: share cache full
Apr 13 14:23:06 Tower shfs/user: share cache full
Apr 13 14:23:06 Tower shfs/user: share cache full
Apr 13 14:31:34 Tower shfs/user: share cache full
Apr 13 14:31:34 Tower shfs/user: share cache full
Apr 13 14:31:34 Tower shfs/user: share cache full
Apr 13 14:48:16 Tower shfs/user: share cache full
Apr 13 14:48:16 Tower shfs/user: share cache full
Apr 13 14:48:16 Tower shfs/user: share cache full
Apr 13 14:48:38 Tower shfs/user: share cache full
Apr 13 14:50:45 Tower shfs/user: share cache full

That message was repeated 157 times, it only occurred during the time we were watching the movie, and interestingly enough the movie being streamed was on the cache drive (it got moved onto the array later that night by the Mover).

 

The next day, a quick forum search prompted me to Stop/Start the array (but I did not reboot) to "fix" the problem, and I haven't seen the error since (but I also haven't streamed a movie since).  The cache drive was in no way full, with plenty of empty space, so I don't exactly understand the error.  I also didn't see any recent examples of this error on unRAID in the forum (perhaps I missed them?), but it might be that this is the first example of the problem returning from the dead, and might be caused by Ryzen.

 

I plan to do some more testing with this when I get some time.  I'll try and finish the movie (but of course it is now in the array and not on the cache drive), and I'll try to stream other movies from both the cache and the array.  Until I figure out more, it might be best just to chalk this up as a one-time anomaly.

 

 

4)  The Case of the Reappearing Cache Drive, Episode 2:  Just when you thought the story was over... HAH!  You're so wrong, there's more to tell.  When I physically removed the old 840 cache drive today, I also installed two new 3TB HD's so I could pre-clear them.  But back in the GUI, unRAID still lists the 840 under "Unassigned Devices".  unRAID also fails to show the two new HD's I just installed.  Even more odd, I started pre-clears on both drives, and in the GUI unRAID indicates "Preclear in progress... Pre-Read.  19% @ 119MB/s (1:20:28)" for the 840 SSD.  

 

Now, to be fair, the 840 was drive ID "sdd", which is also the same drive ID for one of the two new HD's I installed (same slot), so this seems to be simply that unRAID is correctly reporting that sdd is being pre-cleared, but falsely reporting that sdd is the 840, and not one of the new drives.

 

And it is worth mentioning that I had dlandon's "Unassigned Devices" installed until just an hour ago.  I had noticed that I was seeing two listing of Unassigned Devices on the unRAID Main page, the first with a blue tic-tac-toe icon (which lists the old 840), and the second with the padlock icon (which correctly listed the two new drives).  Not liking what I was seeing, I uninstalled the plug-in, and was surprised to see that the tic-tac-toe version of Unassigned Devices was still showing - I'm now thinking that this is built into unRAID stock (when did that happen?...), and by having the separate plugin installed I was seeing two lists.

 

I don't know if having the plugin installed contributed to the problem, but it does seem that the plugin Unassigned Devices worked better (it at least showed the correct drives) but it did not show the running pre-clears.  Removing it made no improvement, but I also haven't restarted the array or rebooted.

 

 

5)  The Case of the Unspinnable Drives:  In the GUI, if I press the small gray up-pointing triangle to the left of the drive, the drive does not spin up.  If I access the drive through the shares, it will spin-up just fine.  I'm also thinking that the all disks "Spin Up" button at the bottom of the Main page does work, but I might need to test again to be sure.

 

Here's the error message that I get in the log when I try to spin-up an individual drive:

Quote

Apr 13 13:16:03 Tower emhttp: error: cmdSpinup: missing csrf_token

 

 

6)  The Case of the Unexplained Error Messages:  These error messages occurred for the very first time today, and if I'm reading the log correctly it was at the same time I was removing the 840 SSD and installing a couple new HD's. 

 

Quote

Apr 24 12:01:08 Tower kernel: CPU: 6 PID: 10828 Comm: kworker/u32:0 Not tainted 4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 6 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 6 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 6 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 6 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 6 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 6 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 4 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 4 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 4 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 4 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:
Apr 24 12:01:08 Tower kernel: CPU: 4 PID: 10828 Comm: kworker/u32:0 Tainted: G        W       4.9.19-unRAID #1
Apr 24 12:01:08 Tower kernel: Call Trace:

 

 

Well, I guess that's it for now.  While it is nice to have achieved 3 weeks of up-time, these little issues are concerning, and taken as a whole makes me wonder if unRAID is actually running correctly on Ryzen.  Perhaps one or more of these are Ryzen caused issues that are unrelated to the stability issues we addressed by disabling "Global C-states Control".

 

-Paul

Link to comment
15 minutes ago, Pauven said:

2)  The Case of the Reappearing Cache Drive:  In my previous build, I had a Samsung 840 1TB SSD installed as my cache drive.  In my upgrade to Ryzen, I installed a faster Samsung 960 1TB M.2 drive.  During this time, I left the old 840 installed and unassigned.  But on every server reboot, unRAID reassigns the 840 as the cache drive (and unassigns the 960).  I have to stop the array, change the assignment, and restart the array every time I reboot to get the 960 back into the cache drive role.  

 

Because I had so many crashes, I made a point of correcting this assignment and then rebooting the array before it crashed (thinking the change was somehow getting lost due to a crash), but this made no difference.  

 

Just a bit ago I physically removed the 840 from the server, but I haven't rebooted since then, so I don't know if this will allow the 960 to come up as cache drive.  I'm also a bit concerned that even if I fix it for now, if I ever re-install the 840 that this bad behavior will return.

 

 

This is a common issue, one or the other solution from the FAQ should fix it:

 

 

Edited by johnnie.black
Link to comment
5 minutes ago, johnnie.black said:

This is a common issue, one or the other solution from the FAQ should fix it:

 

Wait... there's a FAQ?! [searches Lime-Tech documentation... nope.  Searches forum... oh, there it is!].  I guess the FAQ isn't official Lime-Tech documentation, but it would have been nice if they at least linked to it.  The only FAQ they seem to be linking to is for 4.7.

 

Anyway, thanks for the solution johnnie.black.  I am using Chrome.  The next time I go through a reboot, I'll try the operation with Firefox or IE.  Any reports on Edge?

 

-Paul

Link to comment
4 hours ago, Bureaucromancer said:

 

It definitely can be patched, but for the moment acs override with the multi function option seems to get everything seperated.

 

For that one mobo inparticular or for all of them?  That one was one of the more expensive ones.

What downsides are there to using acs override?

 

2 hours ago, ufopinball said:

 

Earlier this month, Tom's Hardware held an AMA (Ask Me Anything) with some folks from AMD.  The IOMMU grouping issue was discussed, but not much was thrown our way:

 

 [...]

 

I guess this doesn't really answer the question about how much the BIOS is involved in dividing up the IOMMU groupings.  Anyway, at the very least, we're not the only ones asking this question, and AMD now knows it's an open issue.

 

Link:  http://www.tomshardware.com/reviews/amd-ryzen-ama,5018-2.html

 

- Bill

 

 

Thanks for the info.

Link to comment

So I read an interesting article this morning that has me wondering if an old compiler version might be to blame for unRAID on Ryzen issues.

 

http://www.phoronix.com/scan.php?page=news_item&px=Kernel-GCC-April-Bumping

 

It seems reasonable to me (without knowing for a fact) that Lime-Tech has to compile their own kernels to include the "special sauce" that makes unRAID what it is.

 

Does anyone know what compiler version Lime-Tech uses to compile the unRAID kernel?  It stands to reason that a newer compiler that is "Ryzen aware" would be important to prevent any performance or compatibility issues with the Linux kernel on AMD's newest architecture.

 

Even the recommended compiler versions are several years old, so I would anticipate that LT is on something compatible, but who knows...

 

Perhaps Lime-Tech would care to comment?

 

-Paul

Link to comment
On 4/24/2017 at 2:08 PM, Bureaucromancer said:

Yeah, I've tried with and without pinning, ignore msrs is on and iommu=pt.  For the moment I'm going to do some experiments on bare metal and see what I get, there's definitely something up with the cores and I'd like, among other things, rule out a bad cpu.

 

So I've did some playing around with bench marking and my 8 core fx/unraid setup on win10.  With NPT disabled I am seeing what you described as progressively worse multi-threaded performance when adding additional cpus.  It's downright horrible in some tests with all 8 cores.  Single threaded performance, however, seems to be pretty consistent with ~12-15% performance drop across the board.  4 cpus looks like the sweet spot for overall system performance.   It seems to be a mixed bag though.  My cinebench tests did not show this behavior and performance seemed to scale up pretty well as I added additional cores.  I'm also running a Linux Mint 18.1 vm and while I saw some noticeable degradation in win10 (login time, internet browsing, some desktop apps, etc), disabling NPT seemed to improve the overall desktop experience and I was not able reproduce the same weird multi-threaded issues.  

 

So it looks like I'll be dealing with same issues with Ryzen...  for a second I started to wander if I should hold off and wait a while longer to see if some of these things get ironed out before taking the dive. But then I realized that I've been with amd since the k5. What am going to do, go out and buy intel? 

Edited by Some Dude
Link to comment

Hi all

 

Just wanted to post my story so far.

 

Specs

CPU: Ryzen 1700

MOBO: Gigabyte GA-AX370-Gaming-K7 (with bios upgraded from F2 to F3)

RAM: Crucial 16GB RAM(Non-ECC) - CT16G4DFD8213 (only single ram stick atm. Will add more in the near future)

SSD(Cache): Kingston SSDNow UV400 240GB

PSU: Corsair RM550X(550W)

HDD(3 of them, 1 of them as parity): WD Red 3TB (WD30EFRX)

Graphics Card: GT710 or 20 or 30 (forgot)

 

First of all, I'm very new to the unRAID as a whole (in fact these kind of technology as a whole).

What I wanted out of unRAID was something that can host multiple docker, VMs and most importantly work as a NAS + PlexServer.

 

It has been about 5 days since I got unRAID up and running and during this time, my goal was to simply get the server running without freezing/crashing.

 

In the beginning, unRAID kept on crashing/freezing/hanging for me every 12~24 hour period. 

Like general consensus on this forum, I ran memtest for a day or so to see if it was my memory issue but thankfully my memory was fine.

 

Than luckily, I encountered this thread where @Pauven mentioned that turning off Global C-state was the key to his stability issue.

 

I tried to look for this option as well in the bios but I couldn't find it originally. This issue was solved by upgrading the bios to F3.

 

Now my server's uptime has reached just over 1 day. 

 

I know it isn't probably a time to feel safe just yet but I am very relieved to see this result.

 

Once I move over all the files scattered all over my various external HDDs, I'll install PLEX and start on the second part of my journey.

 

Thank you everyone who posted here. I especially want to thank @Pauven, whom I wouldn't have been able to solve(hopefully for good) this stability issue without.

 

0ad15d249e.jpg

 

Cheers,

Q

   
Edited by Roang
Made motherboard detail clearer
Link to comment

So has anyone who has bought the Ryzen chip had any good luck in terms of playing games at around 60 fps? At the moment games just jump all over the place from a steady fps down to the 10-20 fps. Portal and Portal 2 work like a charm, but any game like dying light, witcher 3, borderlands, or anything with some extra umph seems to go all over the place in terms of performance. Running on 3d Mark had no issues. I'm at a bit of a loss on what to do.

I've changed the minimum cpu frequency to match the max to try and reduce any latency that may be causing any issues, changed some performance metrics, etc. I'm not sure on where else to go from here. 

I'm loving everything else about unraid, just this last hump i'm trying to get over and this system will be great. Any and all help will be appreciative!

Edited by GViz
Link to comment
10 hours ago, GViz said:

So has anyone who has bought the Ryzen chip had any good luck in terms of playing games at around 60 fps? At the moment games just jump all over the place from a steady fps down to the 10-20 fps. Portal and Portal 2 work like a charm, but any game like dying light, witcher 3, borderlands, or anything with some extra umph seems to go all over the place in terms of performance. Running on 3d Mark had no issues. I'm at a bit of a loss on what to do.

I've changed the minimum cpu frequency to match the max to try and reduce any latency that may be causing any issues, changed some performance metrics, etc. I'm not sure on where else to go from here. 

I'm loving everything else about unraid, just this last hump i'm trying to get over and this system will be great. Any and all help will be appreciative!

I do not have firsthand experience, but have read about it in a few places.

 

For some reason, Ryzen/AMD appears to tank framerates unless you disable NPT (Nested Page Tables).

Disabling it, though, degrades CPU performance; the extent of which seems to vary on who you ask.

 

We've only got a few places it's referenced, but the story is still fairly similar all around.

It's been posted on this thread, at reddit (twice), and on the VFIO mailing list with no response.

 

Hopefully a less harmful workaround than disabling NPT is found; the second reddit post looks interesting.

Edited by dkabot
Link to comment
5 hours ago, dkabot said:

I do not have firsthand experience, but have read about it in a few places.

 

For some reason, Ryzen/AMD appears to tank framerates unless you disable NPT (Nested Page Tables).

Disabling it, though, degrades CPU performance; the extent of which seems to vary on who you ask.

 

We've only got a few places it's referenced, but the story is still fairly similar all around.

It's been posted on this thread, at reddit (twice), and on the VFIO mailing list with no response.

 

Hopefully a less harmful workaround than disabling NPT is found; the second reddit post looks interesting.

 

How did i miss this within this thread. Ugh. Thank you for this information. I'm going to try and give it a whirl tonight,  see what my results are.  I'll report back when i have anything meaningful.  

Link to comment

Count me in the group that are testing.  Brought my R5 1600x online yesterday evening.  Will begin playing with passthrough today. Testing firestrike and games for today with Windows, then off to play with KVM.

Specs

R5 1600x 

MSI X370 SLI PLUS bios 3.2

NVMe M.2 WD black 256GB cache

1TB data drive

1080 GTX hoping to passthrough. 

Link to comment

Something odd I've encountered but not seen in here based on some searches: Windows 8.1/10 refused to start the installer on the default VM settings.

It'd get to the Windows logo flag, I'd get some HDD access for a little while, followed by absolutely nothing at all to indicate it was doing anything.

 

I followed some instructions I saw elsewhere about doing an upgrade to 10 where they said to set CPU to "Emulated (QEMU64)" with one core, and that got it booting.

The passed through GPU seems to be irrelevant to this occurring, as the failure and success cases were the same with VNC or a passed through GPU.

 

It seems once it's been installed and downloaded a round of updates (probably unnecessary?) I was able to switch it back and have it not fail to boot.

May be common knowledge, I'm not certain, but tossing it in this thread for completeness' sake since it doesn't appear to be mentioned.

Link to comment
On 5/3/2017 at 9:11 AM, phbigred said:

Count me in the group that are testing.  Brought my R5 1600x online yesterday evening.  Will begin playing with passthrough today. Testing firestrike and games for today with Windows, then off to play with KVM.

Specs

R5 1600x 

MSI X370 SLI PLUS bios 3.2

NVMe M.2 WD black 256GB cache

1TB data drive

1080 GTX hoping to passthrough. 

If you're concerned with how to passthrough the gpu, I would look up spaceinvader one's youtube tutorial on it. I linked it earlier in this threat. No longer do you need two gpu's to do passthroughs. You'll still need another computer to edit a hex file.

 

 

14 hours ago, dkabot said:

Something odd I've encountered but not seen in here based on some searches: Windows 8.1/10 refused to start the installer on the default VM settings.

It'd get to the Windows logo flag, I'd get some HDD access for a little while, followed by absolutely nothing at all to indicate it was doing anything.

 

I followed some instructions I saw elsewhere about doing an upgrade to 10 where they said to set CPU to "Emulated (QEMU64)" with one core, and that got it booting.

The passed through GPU seems to be irrelevant to this occurring, as the failure and success cases were the same with VNC or a passed through GPU.

 

It seems once it's been installed and downloaded a round of updates (probably unnecessary?) I was able to switch it back and have it not fail to boot.

May be common knowledge, I'm not certain, but tossing it in this thread for completeness' sake since it doesn't appear to be mentioned.

I was able to get mine up and running without any issues. I am using cpu cores 8-15, it should be the same since you're on the 1700 and i'm using a 1700x. 30GB virtual hard drive, 10 gb of ram (there is a bare minimum of at least 2), stored within the domains folder that unraid creates for VM's. Are you still having issues setting up a vm?

I'm still having issues getting to turn off npt tables. I found an older discussion lime tech (https://goo.gl/4aRAoL), but it doesn't seem i'm making any good changes. Even looking up else where. I'm most likely doing it incorrectly as even if I have it set the npt to 0 or 1, game's still have some issues. Were you able to have any success phbigred?

Link to comment
11 hours ago, GViz said:

I was able to get mine up and running without any issues. I am using cpu cores 8-15, it should be the same since you're on the 1700 and i'm using a 1700x. 30GB virtual hard drive, 10 gb of ram (there is a bare minimum of at least 2), stored within the domains folder that unraid creates for VM's. Are you still having issues setting up a vm?

No; while I did encounter issues with setting up Windows, they were alleviated fairly easily with the instructions I noted.

 

My current issue is that as a guest, Kubuntu's installer will fairly consistently hang the system, Ubuntu's will occasionally do so and even Windows has managed to do it once.

Trying to poke at other Linux host setups to see if they change the situation, and so far the only one I got far enough to work (kernel 4.10) hung the guest when I tried to load Kubuntu, with the host then noting that the process for KVM was hung every ~2 minutes.

 

Because of that, I haven't even gotten into trying disabling NPT or any other workarounds for the GPU hit; just trying to sort out if there's anything to deal with hard locks.

Edited by dkabot
Link to comment

New Ryzen BIOS updates will soon be in the works, as AMD is working on an AGESA 1.0.0.6 microcode update with improved memory support:  

 

On a separate note, I'm happy to announce that my Ryzen server now has over 32 days of up-time, more than a full month.  Disabling Global C-state Control really made a big difference for my motherboard.

 

The one real problem I have with disabling "Global C-state Control" is that my power usage has increased.  Instead of idling closer to 90w, the server is idling at almost 120w.  More than anything else, this is shortening how long the UPS can keep the server running.  Before upgrading to Ryzen, my Celeron G1610 idled with the same drives at about 71w, and I could get 90 minutes easy from my UPS.  Now my runtime is about 60 minutes.  Even getting the idle down to 90w would give me 20 more minutes of runtime on the UPS.  I was prepared for a higher idle with the upgrade to Ryzen, but never anticipated it would be 45w higher, that's ridiculous.

 

Hopefully whatever is broken in unRAID gets fixed so that we no longer have to disable the C-states.

 

I have been able to watch a couple movies without issues, both directly from the cache drive and from the array.  The one movie I couldn't finish earlier must have just been a bad ISO, but since that the movie is "Monster Trucks" and isn't worth finishing, I haven't gone back and tried it again.  Perhaps Ryzen has some artificial intelligence and attempted to save me by corrupting the ISO.  B|

 

I've successfully pre-cleared a couple new drives (using Joe L.'s classic command line utility, not the plugin) and added them to my array without any issues, other than slowing down my monthly parity check by a couple minutes.

 

I've also added the Docker for CrashPlan without any issues. I haven't had a chance to do any more VM testing.

 

4 hours ago, dkabot said:

My current issue is that as a guest, Kubuntu's installer will fairly consistently hang the system, Ubuntu's will occasionally do so and even Windows has managed to do it once.

 

I found that the Lubuntu installer consistently hung the system too when installing a new VM.

 

Haven't seen anything from Lime-Tech regarding Ryzen in a long time.  Curious if they have had any progress.

 

-Paul

  • Upvote 1
Link to comment
1 hour ago, Pauven said:

New Ryzen BIOS updates will soon be in the works, as AMD is working on an AGESA 1.0.0.6 microcode update with improved memory support:  

On a separate note, I'm happy to announce that my Ryzen server now has over 32 days of up-time, more than a full month.  Disabling Global C-state Control really made a big difference for my motherboard.

 

The one real problem I have with disabling "Global C-state Control" is that my power usage has increased.  Instead of idling closer to 90w, the server is idling at almost 120w.  More than anything else, this is shortening how long the UPS can keep the server running.  Before upgrading to Ryzen, my Celeron G1610 idled with the same drives at about 71w, and I could get 90 minutes easy from my UPS.  Now my runtime is about 60 minutes.  Even getting the idle down to 90w would give me 20 more minutes of runtime on the UPS.  I was prepared for a higher idle with the upgrade to Ryzen, but never anticipated it would be 45w higher, that's ridiculous.

 

 

Glad to hear there are new updates.  I'm not having any issues with my RAM, but expansion on memory compatibility is still a plus for the platform.

 

The second article did mention: 

 

We resolved the “overclock sleep bug” where an incorrect CPU frequency could be reported after resuming from S3 sleep.

 

Dunno if that will help solve the C-State issue at all?

 

BTW, do you keep your drives spinning all the time?  My system idles at 54 watts (drives spun down, Win10 VM active but idle), which is substantially better than the 90 watts you're seeing with C-states enabled.  Considering our system configurations are reasonably similar, I'd expect the power usage to be roughly the same?  I'm measuring off the UPS, and the unRAID tower is the only active item plugged into it.  My C-States setting is the default value, enabled.

 

- Bill

 

Link to comment
1 minute ago, ufopinball said:

 

Glad to hear there are new updates.  I'm not having any issues with my RAM, but expansion on memory compatibility is still a plus for the platform.

 

The second article did mention: 

 

We resolved the “overclock sleep bug” where an incorrect CPU frequency could be reported after resuming from S3 sleep.

 

Dunno if that will help solve the C-State issue at all?

 

BTW, do you keep your drives spinning all the time?  My system idles at 54 watts (drives spun down, Win10 VM active but idle), which is substantially better than the 90 watts you're seeing with C-states enabled.  Considering our system configurations are reasonably similar, I'd expect the power usage to be roughly the same?  I'm measuring off the UPS, and the unRAID tower is the only active item plugged into it.  My C-States setting is the default value, enabled.

 

- Bill

 

 

The overclock sleep bug is unrelated to the C-state issue.  Basically, if you overclock your system, that bug causes the reported frequency to be really high after resuming from sleep.  It was fixed in AGESA 1.0.0.4, so recent BIOS' already have this fix.

 

I definitely spin the drives down.  But even then, with 18 HD's (approx 5-10w spun down), my 2760A controller card (a massive 27w at idle), 3x 120mm case fans (about 5w at idle speeds), the six SAS backplanes in my case (another 10w) and the GT 710 GPU (guessing 5-10w at idle), combined with a Ryzen idling at 38w, and my server easily gets into the 90w-100w range at best.  

 

That was one of my surprises when I set out to build a sub 20w system around the Celeron G1610, and ended up with 70w idles, the majority of which came from my 24-port SAS adapter card and my very nice X-Case RM-424 Pro server case, two things I never even anticipated.  Add those 40w+ to your 54w build, and you're in the mid 90's too.

 

Going from the Celeron G1610 to the Ryzen 7 1800X should have added about 20-30w to my server, factoring in the higher CPU idle, 4x the memory and the external GPU. With C-states enabled I'm actually fairly close to that.

 

-Paul

Link to comment
2 hours ago, Pauven said:

 

The overclock sleep bug is unrelated to the C-state issue.  Basically, if you overclock your system, that bug causes the reported frequency to be really high after resuming from sleep.  It was fixed in AGESA 1.0.0.4, so recent BIOS' already have this fix.

 

I definitely spin the drives down.  But even then, with 18 HD's (approx 5-10w spun down), my 2760A controller card (a massive 27w at idle), 3x 120mm case fans (about 5w at idle speeds), the six SAS backplanes in my case (another 10w) and the GT 710 GPU (guessing 5-10w at idle), combined with a Ryzen idling at 38w, and my server easily gets into the 90w-100w range at best.  

 

That was one of my surprises when I set out to build a sub 20w system around the Celeron G1610, and ended up with 70w idles, the majority of which came from my 24-port SAS adapter card and my very nice X-Case RM-424 Pro server case, two things I never even anticipated.  Add those 40w+ to your 54w build, and you're in the mid 90's too.

 

Going from the Celeron G1610 to the Ryzen 7 1800X should have added about 20-30w to my server, factoring in the higher CPU idle, 4x the memory and the external GPU. With C-states enabled I'm actually fairly close to that.

 

-Paul

 

Okay, so if the overclock sleep bug is previously fixed, is the new AGESA 1.0.0.6 all about memory compatibility?  Is there something specific you're looking for in this release?

 

Regarding power consumption ... I was actually looking at your HighPoint 2760A.  If I upgraded to a 20 or 24-drive case, I'm going to need more ports.  Right now I have 8 off the motherboard, and 8 off the Dell Perc H310 controller.  I guess at best I'd only need 8 more ports, but it's nice to know what's available.

 

Interestingly enough, I did some power testing on the various controllers I have in-house.  Here's what I recorded:

 

Dell HV52W PERC H310 (8 ports) -- 7.2 watts
SuperMicro AOC-SAS2LP-MV8 (8 ports) -- 8.8 watts
Adaptec RAID 1430A (4 ports) -- 5.2 watts

 

At the time my motherboard had only five SATA ports ... so to support 15 drive slots (three 5x3 cages), I would either need two 8-ports, or three 4-ports (plus MB ports).

 

The two Dell H310s worked out best at 14.4 watts, the two SuperMicro SAS2LPs come in at 17.6 watts, and the three Adaptec 1430As total 15.6 watts.

 

Considering your power report I'll probably spend the upgrading money moving to 8TB drives, instead of moving to a whole new case and add/upgrade drive controllers, just to support more 4TB drives.

 

Also, since I just dropped a lot of $$$ on the Ryzen build to begin with, I figure the next update will have to wait for a bit.  My server is full-ish, but there's still a fair amount of space available.  I'm not adding content *that* fast, and there are a few things I can totally get rid of and reclaim that space.

 

- Bill

 

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.