My M2 Ultra Mac Studio is on Ethernet…well it is on both Ethernet and WiFi but it doesn’t use WiFi for anything but AirDrop.
I use 1x=poly-sinc-gauss-long Nx=poly-sinc-gauss-hires-lp and I tried it with a 24/48 file…I had 48K DSD on but I had hissing and popping issues along with the drop outs.
I’ll try again sometime…but I don’t have time right now.
Strange indeed. Next I would check if the network flow control works, because Ultra can certainly put some pressure on the NAA.
My test case above is with Spring 2, connected to my Xeon workstation (running Ubuntu 22.04 and networkaudiod service). Going through HPE 1920S switch with flow control activated from settings.
In my mind, it’s become clear that the performance deltas for HQP between M1 → M1 xxx → M2 → M2 xxx → M2 Ultra are not significant. Certainly not nearly significant enough to justify thousands of dollars on a Mac Studio. If you are happy with DSD256 and latest EC modulators you can get away with a baseline M1. If you want DSD512 then build a (mini) PC with fast Intel for not much more $$ than the M1.
Sure, I meant relative to a M1 Mac mini, not top of line M2. For example, I am looking at building a B660 deskmini with i7-13700 and Noctua for about $750 all in that should do DSD512 with various EC modulators.
I have a stock Dell desktop: i7-13700, 16gb ram running HQP Embedded 5. DSD 512x48, gauss long/ hires-lp, ASDM7ECv3. Works well even with the stock fan. Will add a Noctua fan shortly. (Quiet and better cooling than the stock fan).
Around $800 usd.
For me the difference between my Mac Mini M1 and MacBook Pro M1Max is pretty big. Thanks to double number of performance cores and over 4x higher memory bandwidth of the Max.
I guess I see it this way - if I can’t go from most desired filters/modulators @ DSD256 to same filter/modulator @ DSD512 then it’s not significant, particularly given the exponential cost vs performance relationship in the M1/M2 lineup with respect to HQP.
What exponential cost versus performance relationship? Just because the difference in single core performance between the base M1 and the M2 Ultra is not huge does not mean the difference between the base M1 and M2 Ultra in overall performance is not gigantic. The cost difference between the two SoCs is not out of line. In fact, it’s a bargain.
If your only use case is single core, I can see you not thinking there is a big difference. But, that is not reality for the most of the computing world…
Correct, this was just in the context of HQP as I originally noted … " the performance deltas for HQP ". Should have been more clear on that.
It doesn’t make sense to me to spend 3x price (or more) of say an M1 Mac Mini for what I see as incremental gains in choice of filters, and hope of squeezing out DSD512 at 7EC-light, etc. Better off moving to a PC at that point imo. That’s all.
Thx - I’ll try but I’m having difficulty with local shops.
The prices include the GPU so a couple of places I talked with online that are about $2400–$2800 are not far off your estimate.