SOtM Launches sMB-Q370 Motherboard

12 hours ago, internethandle said:

@The Computer Audiophile Chris why let the "propeller heads" as they were rightfully referred to over at Audio Asylum for years re-tread the same old tired bits are bits dreck in the thread? Just banish them to Objective-fi where they belong, please.  

Because this is an "internet AUDIO DISCUSSION forum" where all points of view are welcome. If you don't like the thread then ignore it.

But usually the open discussion isn’t done in a thread with a specific title.  In this case the open discussion amounts to nothing more than pissing in the shoes of the OP.

Nobody should have to wade through pages upon pages of rehashed, ultimately pointless “debate” (read: entrenchment/tribalism) over objectivism in digital audio to get to any relevant information about the original topic, especially when that topic is a specific product. There’s a whole sub-forum to do that. Start a thread there, if you want.
 

Any assertions that the forum has historically operated any differently is ahistorical at best. The forum’s very former name is “Computer Audiophile” and was built in part on providing audiophile computer build recipes - what did you think was largely discussed here, foobar plugins? When a thread gets overly derailed with this kind of discussion, it’s usually reigned in. 

1 hour ago, internethandle said:

Nobody should have to wade through pages upon pages of rehashed, ultimately pointless “debate” (read: entrenchment/tribalism) over objectivism in digital audio to get to any relevant information about the original topic, especially when that topic is a specific product. There’s a whole sub-forum to do that. Start a thread there, if you want.
 

Any assertions that the forum has historically operated any differently is ahistorical at best. The forum’s very former name is “Computer Audiophile” and was built in part on providing audiophile computer build recipes - what did you think was largely discussed here, foobar plugins? When a thread gets overly derailed with this kind of discussion, it’s usually reigned in. 

Every discussion is that way on the internet, PERIOD. Very few people listen to others and just want to spout what they believe and move on.

59 minutes ago, kumakuma said:

 

Finally a post from @rando that I understand!

He, or she, is not called Rando for nothing!

17 hours ago, AudioDoctor said:

 

He, or she, is not called Rando for nothing!

Great job using gender identity as a negative and engaging in online discrimination.  

1 hour ago, rando said:

 

Great job using gender identity as a negative and engaging in online discrimination.  

What?

Pity theres no standards committee/agreement on what makes a CPU/MOBO combination work better for audio. Until that happens, each new product just looks like a shot in the dark.

3 hours ago, davide256 said:

Pity theres no standards committee/agreement on what makes a CPU/MOBO combination work better for audio. Until that happens, each new product just looks like a shot in the dark.

in mu experience thus far having many direct phases for the CPU is a bonus for sound quality, as well as having more cores do very little.

From a pure computing aspect the Intel NUC I started with was overpowered, now I'm abusing a 12 core CPU.

I guess that pro/heavy OC oriented boards will do better.

1 hour ago, MarcelNL said:

in mu experience thus far having many direct phases for the CPU is a bonus for sound quality, as well as having more cores do very little.

From a pure computing aspect the Intel NUC I started with was overpowered, now I'm abusing a 12 core CPU.

I guess that pro/heavy OC oriented boards will do better.

Curious, I'm now running an 11i3 NUC at the lowest CPU clock speed possible for maximum transparency with native rates. Cores don't seem to matter other than not running

the application on the same core(s) that are handling disk and eth I/O.  I do think the base CPU power matters, theres a clear difference in treble transient "crispness" between

something like an Atom processor unit vs more powerful CPU such as i3, i5, i7 etc. Haven't really seen CPU utilization impact SQ if PS is good, just causes undesirable pauses/stops

during playback. But I'm sure this gets flipped for those doing oversampling and other added post processing

that sounds comparable to my observations and those from others based on what I read and heard...although I cannot imagine that one needs 12 cores for running a linux distro and some I/O and I did not find as much of a difference when going to higher speed CPU;s as with more cores...I'm unable to explain. the end result is good enough not having to,,,,

No reviews ? No feedbacks ?

Back on topic:

The sMB-Q370 appears to accept two external clocks - 24Mhz and 25MHz.  Can anyone shed light on the following:

• Can external clock be added to both, or just one of these at at time?
• What motherboard functions do these two clock frequencies control?

Reason for my question is that if 24Mhz clock on the sMB-Q370 if for the onboard USB clocks, but you have a standalone USB card with a good clocks, then you could obviously look to try upgrading the 25MHz clock first.  In other words, there may be a preferable upgrade path in terms of money spend.

Equally if the clocks just affect the Intel Q370 chipset, then then may be no benefit for those out us who use PCI cards which are 'CPU direct' and avoid the chipset entirely.

Thanks

Spoke too soon!  I've answered my question on page 3:  https://docs.sotm-audio.com/lib/exe/fetch.php?media=manuals:smb-q370_productguide_v1.1c.pdf

Onboard clocks are for the Q370 chipset and the Ethernet controller.

In my current set up I bypass these by using CPU-direct PCI lanes - one 8x for a solarflare network card and the other 8x for JCAT USB XE.  Therefore I wouldn't get any benefits from the use of external clocks.

Will be interesting what other features standout in a sMB-Q370 review.

Great point you make here, I have been inquiring with Pink faun to modify my MB and you made me wonder why they propose two clocks...I probably only need one, since I'm using a Solarflare and have the ultra OCXO on the I2S bridge!

46 minutes ago, MarcelNL said:

Great point you make here, I have been inquiring with Pink faun to modify my MB and you made me wonder why they propose two clocks...I probably only need one, since I'm using a Solarflare and have the ultra OCXO on the I2S bridge!

Indeed - Put one card in your PCIEX16 slot and the other in PCIEX8 on your Aorus x570.  You'll be CPU direct, so clocks 'shouldn't' benefit you and you'll save a wad of cash.   sMB-Q370 is therefore interesting if you decided NOT to go for external cards, but with the cost of good quality external clocks being so high, I can't see a reason why you wouldn't do down this route like we have with Solarflare and JCAT/Pink Faun (somebody please shout out if this is incorrect).

Perhaps we know now why Taiko don't have any fancy clocks on their motherboards, it's just a standard SAGE C631e because the best thing you can do for accuracy of the CPU is a top quality linear supply and bios/OS optimisation (I presume there is no other form of clock external to the CPU).  Attach everything to the CPU directly, and power the CPU with the best power you can find and your timing / transient response will be the best.  Of course, clocks or not, design of the motherboard still matters (copper traces, VRMS etc...) to optimise the CPU and it's power delivery.

Same applies to Innuos (probably antipodes and others too) if you design around the CPU-direct philosophy, no reason for MB level clocks.

I think I learned something today!

That is exactly what I did...was just wondering if there is any part of the MB that would benefit from a better clock?

The chipset? Indeed the CPU? (always assumed they have their internal reference)?

10 minutes ago, MarcelNL said:

That is exactly what I did...was just wondering if there is any part of the MB that would benefit from a better clock?

The chipset? Indeed the CPU? (always assumed they have their internal reference)?

Same here - anyone?

(what I can say is that the ultra OCXO on the I2S transformed the output more than I expected)

what I read until now it seems as if the PCI bus has it's own clock...YMMV....I'll wait for the more informed response from Jord....