austinpop

I'm the owner of build 17. Nothing yet, but given the update you posted, I am hopeful I'll hear from @paulhynes any day now.

What a fantastic rendition of Verklärte Nacht, one of my favorites! Thanks for the pointer. EDIT: here is the front cover:

I haven't used LMS in months, so it's possible it's a new problem.

Guys, I've played DSD through LMS to SL no problem. Do you have the DSDPlayer plugin installed on LMS? See my post:

A worthy Mahler 9.

The first post of this long thread will get you going.

Ok folks, this has gotten OT. Let’s get back to direct listening experiences please.

Thanks for the clarifications. 12m??!! Where I live in the Texas hill country, I'd be lucky if I had 1m of topsoil before I hit limestone.

Hi Miki, Could you explain in a little more detail what exactly you have added?

Beats me. I can only go by the Intel spec, which says that it consumes 8mW in "L1.2" state, which is one of the L1 sub-states in the PCIe standard. See https://semiengineering.com/using-pci-express-l1-sub-states-to-minimize-power-consumption-in-advanced-process-nodes/ . So I don't understand why Anand Tech would claim Optane does not implement any low power idle states. Maybe their benchmark doesn't test PCIe power states. In any case, this is taking us into the weeds. Remember - the only metric we care about in this discussion is: does it sound harsh? This started with the observation that people were finding NVMe SSDs to sound harsh. Folks looked at the specs for Optane and wondered - do these lower power specs translate to an SSD that doesn't sound as harsh? Upon trying it, the answer seems to be - yes. Now - does this correlate with the actual power consumption of these Optane SSDs in our servers? We don't know, unless we get in and do actual power measurements of Optane SSD vs traditional SSDs. Maybe someone really motivated could get in and do that, but as far as I'm concerned, the specs were just a selection method, and the true test was the listening test.

Sounds like it may be worth once again reiterating the rationale for Optane for streaming audio. Optane and its underlying 3-D X-point storage technology have 2 main characteristics that interest us audiophiles: Potentially better SQ than other (SLC/TLC/QLC) NVMe SSDs due to much lower power consumptions - per specs: Idle power: 8mW Active power: 2.5W Potentially better SQ than other (SLC/TLC/QLC) NVMe SSDs due to much lower latency for random I/O at small queue depths. I say potential as there is no proven correlation between SQ and the latency characteristics of the OS disk. Note that both of the above characteristics are available when the Optane disk is just used as a NVMe SSD in an M.2 PCIe slot. This use case should not limited to Intel CPUs. This is the same caveat as Larry's above - I have not tested this on AMD. There is another use case, that we audiophiles don't care about, which is to use the Optane drive as the caching layer of a 2-tier storage hierarchy. In this topology, the Optane drive acts as a cache for a slower SSD or HDD. This particular use case does in fact need certain Intel CPUs and certain OSes (Windows) to work. Confusing matters further, this is the use case most often associated with the use of Optane in the media, so that's why people get the mistaken notion this storage only applies to Intel (on Windows).

An excellent recording and performance of Schoenberg's Transfigured Night (Verklärte Nacht). The Haydn cello concertos are nice, although not sure how they fit under the "Transfigured Night" theme!

Since you can run with Turbo ON with the SMPS brick, this does seem to be a PSU limitation. Sadly, many PSUs make bold claims, but do not cover themselves with glory when called upon to deliver anywhere close to their rated current. If your PSU cannot work with Turbo ON, then it's debatable you would get any SQ benefit from frequencies higher than the base anyway. For now, your best bet may be to keep Turbo OFF until you can acquire a better PSU.

Roy is currently powering his 8700K server with 2 SR7 rails: a 19V rail powering an HDPlex DC-ATX converter, and a separate 12V rail powering the CPU (EPS). He finds this combo far exceeds the HDPlex 400W ATX LPS, even though this has separate rails for the ATX (3.3, 5, and 12V) and EPS (12V).

After some poking around, the SSC controls vary by motherboard: Asrock H370M-ITX/ac: not exposed at all Asrock Z390 Phantom Gaming-ITX/ac: exposed for CPU only as "BCLK Spread Spectrum" Asrock Z390 Phantom Gaming X ATX: exposed for CPU as "BCLK Spread Spectrum," and PCIe as "PCIE PLL SSC." Something to keep an eye on, depending on which mobo is used, but for me, since it's not exposed on my H370M-ITX/ac, I can now only hope it is disabled by default.