Resolume & native SDI output

[Meptik]

We figured out the freezing, which was a separate system all together, however, not sure what would be choking this systems bandwidth? We have PCIe bandwidth configured correctly, with GPU being in a 16x slot, and BM Duo in a 8x slot, despite it needing 4x. Thanks for all the help trouble shooting. Much appreciated.

[Joris]

This is the double edged sword of not shipping Resolume with dedicated hardware like other media servers do. The advantage is that you're not tied to overpriced and outdated hardware, and you can build your own machine with the latest components. The downside is that when something is not working on that hardware, and it's not a known issue, there's very little we can do to help you troubleshoot.

There is nothing in your spec and how you describe your setup that could indicate a problem with Resolume. Have you tried running output to the BM box using other software?

We figured out the freezing

Awesome! What was the issue? It could be helpful to other people that may come across your thread.

[Oaktown]

Windows 10 Pro
X99, Asus
Intel 6850k
32 GB Ram
3x SSDs: 2x M.2 - Samsung Evo 960 Pro, 1x Samsung 850 Pro
GTX 1080
BlackMagic Duo 2 Card
Resolume V5.1.3, Driver 10.8.5

A few questions for you.

• Which X99 do you have?
  How are your 2x M.2 SSDs connected?
  What PCIe slots are the GTX 1080 and BMD Duo 2 in?

[Meptik]

A few questions for you.
Which X99 do you have?
How are your 2x M.2 SSDs connected?
What PCIe slots are the GTX 1080 and BMD Duo 2 in?

All our machines are:
- Asus ROG X99 Mobo w/ Intel 6850k
- M.2 are:
1x via Mobo M.2 PCIe Slot
1x via Angelbird Wings PX1 PCIe x4 M.2 adapter
- PCIe Slots:
1x GTX 1080 in 16x slot
1x M.2 Adapter in PCIe 8x Slot
1x BM Duo 2 Card in PCIe 4x Slot

@Jorris - We completely understand the limitations of not using turnkey hardware. We're not complaining at all! Mainly just trying to best understand what's going on under the hood and how to get everything best streamlined. We're working to make sure we're not missing anything major! Very much appreciate your help!

[Oaktown]

Can you narrow it down to the PCIe slot #? The closest to the RAM would be #1.

[Meptik]

Yup! We've done different testing with different slot and card combinations and seem to get the same results:

1: 16x: GTX 1080
2: 1x: N/A
3: 4x: BM Card
4: 16x: GTX 1080
5: 1x: N/A
6: 8x: N/A

1: 16x: GTX 1080
2: 1x: N/A
3: 4x: M.2 Adapter
4: 16x: GTX 1080
5: 1x: N/A
6: 8x: BM Card

1: 16x: GTX 1080
2: 1x: N/A
3: 4x: N/A
4: 16x: BM Card
5: 1x: N/A
6: 8x: N/A

[Oaktown]

Not sure why the last configuration is slow but as a note, if you put your BM on #6, #1 will run at 8x instead of 16x. Also if you have USB 3.1 enabled, #3 will run at 1x instead of 4x.

Check out your Onboard Devices Configuration for both PCIEX4_1 Slot Bandwidth and PCIEX8_3 Slot Bandwidth are set correctly.

Try running your GTX1080 on #1 and your BM Duo 2 card on #3 in [X4 mode]?

[Meptik]

Thanks Oaktown, much appreciated. The servers just went out on shows, but when they get back we're going to continue testing. Will keep all posted about our finds. Thank you!

[Joris]

I still want to know what the problem with the freezes was on the other machines...

[Meptik]

I still want to know what the problem with the freezes was on the other machines...

haha! Sorry, forgot to respond to this....

It seemed to be due to the new Nvidia drivers. We ended up rolling back our drivers all the way to 372.90 (Don't quote me on the exact version, off the top of my head) and found stability. We built one X99 machine in December that is still running that 372.90(?) version and running solid.

Under the newer/newest versions, the machine would BSOD and restart 8-10 times a day randomly; we had just been running Resolume out to some monitors with content looping for days and it would randomly reboot. From all we could tell in the logs - sometimes it would crash before the system would collect the issue - it had something to do with memory instability in the GPU. The crashing would only occur with Resolume running, which is why we were having a hard time pin-pointing.