Jump to content
IGNORED

Audiolinux Server configurations, Software, Hardware, and Listening Impressions


lmitche

Recommended Posts

13 hours ago, lpost said:

Anyone else running the NUMA-off kernel notice sluggish response via SSH?

 

I don't really mind as it sounds great just curious as I've not seen it mentioned.

I didn't notice any differences in behaviour of SSH response. All working fine after the kernel update.

Link to comment
1 hour ago, LTG2010 said:

I would say a slight increase in clarity. Isolating CPU cores, has a bigger effect in conjunction with Pieros expert menu script. Basically scroll down and delete all the hashtags in front of  "echo". I have a 4 core celeron, I isolated 3 cores for audio. Set boot to extreme and real time priority to standard. Max performance and turbo off. Also as my server is Ethernet out, I changed rtirq.conf from USB to eno1 (my Ethernet bridge out to DAC).

That with ramroot, comfortably beats any other OS I've tried, the clarity, dynamics, bass etc is stunning.

 

Thanks. What ethernet bridge are you using?


"Don't Believe Everything You Think"

System

Link to comment
1 minute ago, mourip said:

Thanks. What ethernet bridge are you using?

My supermicro x10sba motherboard has 2 ports. They are clocked with sclk-ex. I was looking to upgrade the server bigger processor, jcat femto card etc. But following these results, there is no rush. :)

Link to comment
5 minutes ago, LTG2010 said:

My supermicro x10sba motherboard has 2 ports. They are clocked with sclk-ex. I was looking to upgrade the server bigger processor, jcat femto card etc. But following these results, there is no rush. :)

 

Thanks. I forgot that we were in the server thread! I was thinking ethernet to SPDIF converter for the endpoint.

 

I have bridged connections on my server, on board port and a PCIe Intel card. That made a very nice difference. I have a JCAT card ordered now based upon that success. Hoping that it will ship this week. I will power from one of my HDPlex outputs.


"Don't Believe Everything You Think"

System

Link to comment
38 minutes ago, mourip said:

I forgot that we were in the server thread!

LOL, me too, and I started the thread!

 

At the beginning we were still in denial that server side software tweaks matter with a NUC endpoint. Now we know there is no question that they matter, network isolation or not.

 

The server and endpoint combined clearly behave as a single system.

 

Therefore I am happy to see us discuss Audiolinux related server plus endpoint topics here! I'd remove the word "server" from the topic title if I could.

Pareto Audio aka nuckleheadaudio

Link to comment
55 minutes ago, davide256 said:

Has anyone played around with the NUC i8XXX series yet as AL server? Thinking about buying an i5/i7 in this series, when I looked at the CPU benchmark stats

for the NUC i7XXX series, wasn't impressed vs my existing AMD FX-8330 server

 

I remembered seeing two posts regarding the NUC8 and found them. Perhaps there’s others I don’t know of.

 

 

 

 

Link to comment

The new kernel is really a step forward In SQ. My PJYH SqueezeLite endpoint have never sounded that good. 

I added the asound.conf and set Realtime Priority = standard and isolated CPU cores to half and Extreme boot.

I agreed that the sound is more relaxed and also I hear more body to instruments.

Thanks

Theese NUC’s looks interestting

https://wccftech.com/intel-ghost-canyon-frost-canyon-nucs-9th-gen-core-i9-8-cores/

 

Link to comment
30 minutes ago, lmitche said:

Looking at the NUC roadmap raises fears that the new machines won't run at the low tdp (current) levels we enjoy on the 7th generation NUCs today.

1) I suspect the NUC Celeron family is on its way out, given the lack of supply since December in 7 series

2) comparing the 7 series vs 8 series, the critical difference is switch from 2 cores to 4 cores, more speed but also more watts consumption.

 

On the server side, not so bad, an i7 needed a heftier power supply to begin with as server, something like an SR4 or JS-2 seemed mandatory, not sure that using

an i8 changes anything there other than an aftermarket (Akasa?) fanless case now seems required for full throated CPU use?

 

On the endpoint side, challenging as what I've read so far seems to suggest 7CJYH/7PJYH are a step along the way to 7i7xxx in low power mode with LPS 1.2?

Which raises the question of whether the i8xxx series can work in low power mode with the LPS 1.2?

 

 

 

 

Regards,

Dave

 

Audio system

Link to comment
2 minutes ago, lpost said:

 

For me too, numa off has been the most significant improvement in AL for my system since it was released.

So interesting!  

Pareto Audio AMD 7700 Server --> Berkeley Alpha USB --> Jeff Rowland Aeris --> Jeff Rowland 625 S2 --> Focal Utopia 3 Diablos with 2 x Focal Electra SW 1000 BE subs

 

i7-6700K/Windows 10  --> EVGA Nu Audio Card --> Focal CMS50's 

Link to comment
15 minutes ago, davide256 said:

Which raises the question of whether the i8xxx series can work in low power mode with the LPS 1.2?

 

If you consider something like NUC8i7BEH, it is 28W with configurable TDP down of 20W with an i7-8559U.  That's still too much for an LPS-1.2.

Pareto Audio AMD 7700 Server --> Berkeley Alpha USB --> Jeff Rowland Aeris --> Jeff Rowland 625 S2 --> Focal Utopia 3 Diablos with 2 x Focal Electra SW 1000 BE subs

 

i7-6700K/Windows 10  --> EVGA Nu Audio Card --> Focal CMS50's 

Link to comment
21 minutes ago, rickca said:

If you consider something like NUC8i7BEH, it is 28W with configurable TDP down of 20W with an i7-8559U.  That's still too much for an LPS-1.2.

There are likely ways to get the TDP lower, but we won't know until we try.

 

It looks like there is fanless case or two from Akasa for Bean Canyon so that obstacle is removed.

 

 

Pareto Audio aka nuckleheadaudio

Link to comment

I am pretty impressed with the Intel supplied AC adapter (LITEON 19VDC 3.42A p/n PA-1650-90) smps for the NUC7i7DNHE1 in the setup below. The music is lively, lots of 'color' yet no fatigue, and the music/background contrast is great! I guess it has effective EMI filtering!

 

NUC7i7 2019-1.jpg

Link to comment
2 hours ago, lmitche said:

Its true, lots of the NUC sq is delivered with the stock smps. However an lps1.2 or sr4 takes things to a new level.

 I am expecting an SR7DR 9/19V this summer..If it does arrive I'll post impressions!

 

PS - I believe I do have an SQ preference (more palpable, air etc) with the LITEON above to the one I received with the Pentium NUC7PJYH1 which is an APD (Asian Power Devices Inc.) p/n DA65C19, and has the same 19VDC 3.42A output spec, when powering the NUC7i7 (though only briefly tested)

Link to comment

I am running AL with HQPe and Roon. I have a single PC system. 

I am wondering if or how to disable the ALSA Mixer or ALSA??

Piero says that HQPe already connects to ALSA with the best configuration. 

Is there something to modify in Roon or in AL only?

Is there any type of conflict that the code referencing default {type plug slave.pcm hw} and Roon calls “default” hw:X??

I am not using RAAT. 

 If I go into the expert menu, choose 4, systemwide configuration menu, and choose yes,  if the default is the “code” we want to put in,  what is there to change? 

Link to comment

I too run AL with HQPe and Roon on a single PC system.  I've made the ALSA edits but found no difference in SQ.  I don't believe it applies to HQPe systems.

 

The most significant improvement I've experienced since running AL has been the new kernel with NUMA disabled.  It's been a few days, and I'm still quite impressed with the changes/improvements.  Coherency as increased.

Link to comment

I hope you realize as much improvement that I have. 

 

For my setup, I find extreme1 boot, extreme real-time priority, ramroot, no-numa kernel, AMD version of Speedstep and hyper threading and turbo OFF, downcore from 8 to 4 cores (I don't upsample in HQPe and don't need the cores, they sit idle) and fix Ghz at lowest allowed 2.2Ghz.  No numa being the most significant.

 

Every few weeks I will alter one variable at a time and keep coming back to the above settings.  Because I run an AMD, some of the pstate software functions aren't fully supported.  I simply set them fixed in bios to achieve the same.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...