NVMe SSD designed for audiophiles

[yamamoto2002]

I searched a bit, enterprise HDD firmware has “error recovery control” feature called TLER or CCTL to give up r/w retry early and let RAID controller to handle the error. consumer HDD do not have this feature and array is degraded when retry takes 10 seconds. I guess video recorder firmware is developed from this knowledge? on video recorder typical use case scenario, when recording video A, simultaneously, several hours ago recorded video B is played, and if read B retry takes 10 seconds, recording A is failed. video HDD should prioritize write A operation somehow

 

from https://en.wikipedia.org/wiki/Error_recovery_control

Quote

Modern hard drives feature an ability to recover from some read/write errors by internally remapping sectors and performing other forms of self-test and recovery. The process for this can sometimes take several seconds or (under heavy usage) minutes, during which time the drive is unresponsive.

 

Also found interview article with HGST engineers, they do not develop special firmware for audio industries at all, on circa 2013

[NewOldman]

OS on Pci-e adapter in slot 2,

Delock Products 89370 Delock PCI Express x4 Card > 1 x internal NVMe M.2 Key M 80 mm - Low Profile Form Factor

And ASUS Hyper-x card  for music in slot 6

 

and dont forget to settings Pci-e settings in Bios as Data under pci-e slot settings

 

[NewOldman]

13 minutes ago, yamamoto2002 said:

I searched a bit, enterprise HDD firmware has “error recovery control” feature called TLER or CCTL to give up r/w retry early and let RAID controller to handle the error. consumer HDD do not have this feature and array is degraded when retry takes 10 seconds. I guess video recorder firmware is developed from this knowledge?

 

Also found interview article with HGST engineers, they do not develop special firmware for audio industries at all, on circa 2013

that's why you should never use consumer HDD in hardware raid

[MarcelNL]

37 minutes ago, NewOldman said:

OS on Pci-e adapter in slot 2,

Delock Products 89370 Delock PCI Express x4 Card > 1 x internal NVMe M.2 Key M 80 mm - Low Profile Form Factor

And ASUS Hyper-x card  for music in slot 6

 

and dont forget to settings Pci-e settings in Bios as Data under pci-e slot settings

 

thanks for THAT pointer!

 

I have not even begun to go trough the many BIOS optios so it probably would have taken a day or two to figure that out!

 

 

[NewOldman]

6 minutes ago, MarcelNL said:

thanks for THAT pointer!

 

I have not even begun to go trough the many BIOS optios so it probably would have taken a day or two to figure that out!

 

 

it's for the Hyper-x card  not for os

[Rsmaximasr]

What I don’t get is why somebody is trying to improve the source drive when the data read has to pass thru many different pieces of hardware (l1, l2, l3 caches), and memory, what do you think you will get from the femto ssd?

Thats like an electrical utility putting a device at the dam/generator to make the power source pure when the facts are that power source will go thru many transformers/substations which degrades the quality of the power. 
Also, I’ve done many tests comparing enterprise hard drives, consumer hard drives, enterprise nvme/ssd, and consumer nvme/ssd as the source drive for Roon and you couldn’t tell which source drive was being used. 

[MarcelNL]

it's not as if we can change a lot inside the memory cache of a CPU but we can select CPU's for sonic merit and we do, same with RAM memory.

It's not until a system gets to a fairly high spec that the differences between hard drives etc become apparent, switching to NVME and switching off SATA controllers or even adding an external LPSU to an SSD can be heard fairly soon in the process.

 

 

[MarcelNL]

On 1/5/2024 at 8:56 PM, NewOldman said:

OS on Pci-e adapter in slot 2,

Delock Products 89370 Delock PCI Express x4 Card > 1 x internal NVMe M.2 Key M 80 mm - Low Profile Form Factor

And ASUS Hyper-x card  for music in slot 6

 

and dont forget to settings Pci-e settings in Bios as Data under pci-e slot settings

 

Now that I got everything working I'm wondering why you's put the OS and data NVME's in the slots for CPU2 ?

Do you run OS on CPU 2 and the player program on CPU1 (so the output PCIe card comes in slot 3 or 5 ?

 

 

[Ghmi]

On 1/8/2024 at 9:05 AM, MarcelNL said:

adding an external LPSU to an SSD can be heard fairly soon in the process.

Adding an external power supply to your SSD is going to :

 - break your motherboard ground plane design, tremendously boosting EMI

 - introduce a huge ground loop, acting like an antenna so that those EMI can be scattered in your whole room

Rule one of mixed electronic design : proper grounding and single path of return for each load. You're only making things worse.

 

58 minutes ago, MarcelNL said:

Now that I got everything working I'm wondering why you's put the OS and data NVME's in the slots for CPU2 ?

Do you run OS on CPU 2 and the player program on CPU1 (so the output PCIe card comes in slot 3 or 5 ?

Whatever you do, your OS will be running on both CPUs. Your player program can only run on top of the OS.

[MarcelNL]

1 minute ago, Ghmi said:

Adding an external power supply to your SSD is going to :

 - break your motherboard ground plane design, tremendously boosting EMI

 - introduce a huge ground loop, acting like an antenna so that those EMI can be scattered in your whole room

Rule one of mixed electronic design : proper grounding and single path of return for each load. You're only making things worse.

 

Whatever you do, your OS will be running on both CPUs. Your player program can only run on top of the OS.

whatever you do please try something and LISTEN to the results rather than theoretisizing and stating it cannot work, this is getting old fast.

[NewOldman]

1 hour ago, MarcelNL said:

Now that I got everything working I'm wondering why you's put the OS and data NVME's in the slots for CPU2 ?

Do you run OS on CPU 2 and the player program on CPU1 (so the output PCIe card comes in slot 3 or 5 ?

 

 

if you look at the block diagram for the MB you may understand 

 

page 119 in manual

 

[MarcelNL]

10 minutes ago, NewOldman said:

if you look at the block diagram for the MB you may understand 

 

page 119 in manual

 

I have it open in front of me...

slots 1, 2, 6 and 7 are linked to CPU 2

slots 3,4 and 5 to CPU 1 I understand you'd want to stay away from slot 3 and 4 which seems to be shared through a sortof controller.

 

I'd expect CPU to be assigned to handle most of the OS, and be linked to the OS NVME and the player program assigned to CPU 2 and the data files linked to that. 

[Ghmi]

33 minutes ago, MarcelNL said:

whatever you do please try something and LISTEN to the results rather than theoretisizing and stating it cannot work, this is getting old fast.

 

Slot 2 and 6 are the only X8 PCI ports on this mobo which go straigth to the CPU. The 3 and 4 are switched, 1, 5 and 7 are x16 ports. Also, using a dual CPU board really isn't a very good idea if you're concerned about noise. Twice the CPUs = more than twice the electronic noise.

Also, that's not "theory". Unlike you, I'm a practicing engineer.

 

13 minutes ago, NewOldman said:

if you look at the block diagram for the MB you may understand 

page 119 in manual

 

He most definitely won't.

[Ghmi]

1 minute ago, MarcelNL said:

I have it open in front of me...

slots 1, 2, 6 and 7 are linked to CPU 2

slots 3,4 and 5 to CPU 1 I understand you'd want to stay away from slot 3 and 4 which seems to be shared through a sortof controller.

 

I'd expect CPU to be assigned to handle most of the OS, and be linked to the OS NVME and the player program assigned to CPU 2 and the data files linked to that. 

 

You're thinking along the lines of "putting some stuff on one CPU, and some other stuff on the other".

That's not how any of this works. The two CPUs are linked and synchronized anyway.

 

What matters isn't CPU1/CPU2, but the number of lanes of each PCI slot and whether they're switched or not.

[NewOldman]

5 minutes ago, MarcelNL said:

I have it open in front of me...

slots 1, 2, 6 and 7 are linked to CPU 2

slots 3,4 and 5 to CPU 1 I understand you'd want to stay away from slot 3 and 4 which seems to be shared through a sortof controller.

 

I'd expect CPU to be assigned to handle most of the OS, and be linked to the OS NVME and the player program assigned to CPU 2 and the data files linked to that. 

OS/player program  and Music files on CPU2 

network and USB card on CPU 1

[MarcelNL]

[NewOldman]

9 minutes ago, MarcelNL said:

OS on Pci-e adapter in slot 3 

and listen if you like it more

 

[MarcelNL]

Thanks ! I'll do that once I have the passive cooling in place, the two fans make listening for that sort of differences impossible.

[NewOldman]

[NewOldman]

I have used this combination
Dynatron B12
https://www.dynatron.co/product-page/b12
and Noctua 90x90 fan on top of it with rubber stands
with external power supply

 

"but let continue this discussion in right thread instead"

[MarcelNL]

HAHA, I was just looking at that same picture of the Extreme ....

 

I have the passive cooling ready, just need to install it, drill some holes, cut some thread and all that.. 4 heatpipes per CPU and two 300W heat sinks for passive air cooling should do the trick.