DPC Latency, Windows 10 & Laptops (with Thunderbolt) in 2018

If you are new here check this forum first, your question may have been answered.
RELATED
PRODUCTS

Post

generaldiomedes wrote: Sat Feb 23, 2019 11:09 pm Somehow RME manages to get their windows USB 2.0 latency levels to near Mac thunderbolt levels with good driver development.
That "somehow" is because they're one of, if not the only interface manufacturers that develop their drivers in-house specifically for their interfaces.

Most other companies farm out the driver development to third parties or remake some generic ASIO/Core Audio frameworks for their drivers.

Post

BrokenTrance wrote: Fri Feb 22, 2019 11:12 pm Got not good result from Latencymon with new computer, that i bought second hand hahha. What could be wrong?
Perhaps nothing is wrong? It seems strange to me that you'd be talking about what Latencymon tells you, not that you found it impossible to play along with or anything else. Again, it seems that practical considerations don't come into your calculations, which is just the weirdest thing ever to me.
NOVAkILL : Legion GO, AMD Z1x, 16GB RAM, Win11 | Audient EVO 8 | Lumi Keys | Studio Pro 8
Korg Odyssey, bx-oberhausen, Proxima, PolyMax, GR8, JP6K, Union, Atomika,
Invader 2, Flow Motion, Olga, TRK 01, Thorn, Spire, VG Iron

Post

BrokenTrance wrote: Sat Feb 23, 2019 11:15 pm
Tappistry wrote: Sat Feb 23, 2019 12:40 am I haven't really gotten into much with windows laptops and realtime audio performance. I can speak to what I've dealth with on my desktops over the years, however.

Make sure your INF drivers and whatever else is appropriate for your motherboard's onboard USB controllers are up to date. That's one of the most easily overlooked items to update. After that, start trying to isolate problematic drivers. NVIDIA drivers have improved substantially IME over the last few years.

Lastly, I just built a new system and was still having some minor latency issues. I had to go into the BIOS and disable the processor c-states. By default, these power-saving features are turned on and they can cause random spikes in DPC latency when the system starts trying to turn off processor cores and move threads around.
I actually think it was the INF drivers that was especially problematic according to LatencyMon.
That's interesting. You made sure to use the one's specifically from your motherboard manufacturer's website?

Post

BONES wrote: Sun Feb 24, 2019 3:14 am
BrokenTrance wrote: Fri Feb 22, 2019 11:12 pm Got not good result from Latencymon with new computer, that i bought second hand hahha. What could be wrong?
Perhaps nothing is wrong? It seems strange to me that you'd be talking about what Latencymon tells you, not that you found it impossible to play along with or anything else. Again, it seems that practical considerations don't come into your calculations, which is just the weirdest thing ever to me.
I don't use music program very often to be honest. But if i choose to do i think it should be better now with handling audio then before.

Post

Tappistry wrote: Sun Feb 24, 2019 3:15 am
BrokenTrance wrote: Sat Feb 23, 2019 11:15 pm
Tappistry wrote: Sat Feb 23, 2019 12:40 am I haven't really gotten into much with windows laptops and realtime audio performance. I can speak to what I've dealth with on my desktops over the years, however.

Make sure your INF drivers and whatever else is appropriate for your motherboard's onboard USB controllers are up to date. That's one of the most easily overlooked items to update. After that, start trying to isolate problematic drivers. NVIDIA drivers have improved substantially IME over the last few years.

Lastly, I just built a new system and was still having some minor latency issues. I had to go into the BIOS and disable the processor c-states. By default, these power-saving features are turned on and they can cause random spikes in DPC latency when the system starts trying to turn off processor cores and move threads around.
I actually think it was the INF drivers that was especially problematic according to LatencyMon.
That's interesting. You made sure to use the one's specifically from your motherboard manufacturer's website?
I don't really know i used the driver that DriverBooster had.

Post

BrokenTrance wrote: Sun Feb 24, 2019 11:00 am
Tappistry wrote: Sun Feb 24, 2019 3:15 am
BrokenTrance wrote: Sat Feb 23, 2019 11:15 pm
Tappistry wrote: Sat Feb 23, 2019 12:40 am I haven't really gotten into much with windows laptops and realtime audio performance. I can speak to what I've dealth with on my desktops over the years, however.

Make sure your INF drivers and whatever else is appropriate for your motherboard's onboard USB controllers are up to date. That's one of the most easily overlooked items to update. After that, start trying to isolate problematic drivers. NVIDIA drivers have improved substantially IME over the last few years.

Lastly, I just built a new system and was still having some minor latency issues. I had to go into the BIOS and disable the processor c-states. By default, these power-saving features are turned on and they can cause random spikes in DPC latency when the system starts trying to turn off processor cores and move threads around.
I actually think it was the INF drivers that was especially problematic according to LatencyMon.
That's interesting. You made sure to use the one's specifically from your motherboard manufacturer's website?
I don't really know i used the driver that DriverBooster had.
I don’t know what Driver Booster is, but unless a driver update app is made by your computer’s or motherboard’s manufacturer, it’s typically a good idea to stay away.

Post

Yeah you are right. Looked through the drivers i have now with Driverbooster. Seems to be that it has installed Intel driver with some motherboard peripheral. So i think it is ok though.

Post

BONES wrote: Fri Feb 22, 2019 12:55 am TB exists to sell more computers to people who don't think about what they really need. Seriously, what difference does it make if TB is 80 times faster when USB 2.0 is already 8 times faster than you need?
That's a bit harsh, TB exists to service the users who need it and storage and people who need to hook up high bandwidth devices like external GPU's do actually need it.

Audio users? Well, that's a little more questionable.

I agree the bandwidth argument is bull, it was always a waste of time.

So what's it good for?

Bus power delivery and signal RTL.

Ever had an interface where plugging in a pair of condensers at the same time causes one to fail? That's because you're not getting stable power delivery over the USB to the interface. Fixable by plugging in a wall wart to work around it, but people don't always wish too.

That aside, RTL on TB is far tighter than USB tends to be. Yes, RME can get the UFX within spitting distance, but given they spend the vast amount of outgoing cash on R&D, it's unsurprising as that's the business model they adopted and whilst my approval is well known, they really are the exception than the rule in general. Further down the market with a device like the Quantum, they get better speeds with less development going on, simply because they are working with a far more efficient connection protocol out of the box.

I don't disagree, all implementations of USB should be enough, but then if they can do it quicker and cheaper on TB with its direct PCIe connection then I can't blame people for getting excited.

Although... if you want a rock solid, stable PCIe performing unit for a good price. Well, there are still some PCIe cards out there.

Post

Thought I'd just add that I'm currently having issues with this set up:

Asus Zenbook Pro Duo (i9 version) running Cubase 10.5 <-Thunderbolt 3-> UA Apollo X4/X8P

Currently going through different bits to try and address following tonnes of clicks/pops caused by our old friend DPC, amongst other things. Really bad LatencyMon score.

Thought I'd add this as it's a relatively new, expensive laptop and if it will work in the end, great having the mixer window on the second screen. But so far, no dice. I'll keep reporting back, but if this doesn't work, it's going back and being replaced with a new Macbook Pro 16 I think.

Post

necro bump... :D

Since this never got much interest, and I had to choose back then, it became this:
alec.tron wrote: Wed Apr 25, 2018 11:41 pm Dell XPS-15 9560
https://www.notebookcheck.net/Dell-XPS- ... 668.0.html
And what a shit decision (for audio) that was...
As well known now - the XPS line of Dell laptops has been plagued with multiple hard/impossible to fix DPC latency issues, which rear their head regularly, even on non-taxing realtime audio situations, leading to drop outs, glitches, etc pp.

And even after 2 years of Dell attempting to fix this clusterf*, we still have new Bios/Driver releases that fundamentally break everything for audio... like
https://www.reddit.com/r/Dell/comments/ ... ds_of_smi/
https://www.dell.com/community/XPS/Dell ... -p/7501047

Just some anecdotal turd for the XPS line, in case anyone is considering these....
c.

Post

AUTO-ADMIN: Non-MP3, WAV, OGG, SoundCloud, YouTube, Vimeo, Twitter and Facebook links in this post have been protected automatically. Once the member reaches 5 posts the links will function as normal.
Thanks Alec for pointing me out to this forum and thread.


my XPS-9560 experiences are as follows:
* disable Dell services to stop the SECONDS of latency (even the mouse stops!)
** info: https://github.com/pestrela/smi_counter ... -2-concept (https://github.com/pestrela/smi_counter/blob/master/README.md#ring--2-concept)
* even then, I still have yellow performance on latencyMon. These are milliseconds, however.
** For this I've just raised the buffer
* back in the days I used to have very often BSOD. With all the latest updates this was almost erradicated, no comparison to 2018.

In short: I've kept the Dell stuff fix the BSODs. This added seconds of latency. Measuring and stopping the right service fixed that last problem. Happiness now!

Post

dj_estrela wrote: Sun Mar 08, 2020 11:00 pm Thanks Alec for pointing me out to this forum and thread.


my XPS-9560 experiences are as follows:
* disable Dell services to stop the SECONDS of latency (even the mouse stops!)
** info: https://github.com/pestrela/smi_counter ... -2-concept
* even then, I still have yellow performance on latencyMon. These are milliseconds, however.
** For this I've just raised the buffer
* back in the days I used to have very often BSOD. With all the latest updates this was almost erradicated, no comparison to 2018.

In short: I've kept the Dell stuff fix the BSODs. This added seconds of latency. Measuring and stopping the right service fixed that last problem. Happiness now!
Hey, I can't quite follow what you've done. I have a Dell XPS 9575 and I get sudden slow downs, drop outs and seconds long cpu spikes. LatencyMon reports that my system is not good for real time audio. Usually it's fine but hits issues sometimes.

I don't see any of those Dell services in my task manager under services. Is that not where to look? Which ones specifically did you disable? I can't really understand your last paragraph. Can you summarize your changes? Thanks!

Post

To avoid latency problems, working tweaks especially for NVIDIA/Windows users at
viewtopic.php?p=7624448#p7624448

Post

Echoes in the Attic wrote: Mon Mar 09, 2020 3:49 pmI don't see any of those Dell services in my task manager under services. Is that not where to look? Which ones specifically did you disable?
Easiest way:
- Windows10 Start Button click & search (simply by starting to type the following:)
msconfig
- Then switch to Services tab in msconfig window
- enable 'Hide All Microsoft Services'
- look for Dell services running. You can disable/re-enable them here easily, and disabling all has had the biggest effect as far as I can tell

Also, for the XPS 9650 switching off the wireless network altogether is mandatory, while you're dealing with realtime audio and want to limit DPC latency spikes... as is addressing the rumoured/suggested culprits - Nvidia driver settings and Killer Network software are 2 that I addressed and that seemed to have helped too. Alas, this is far from scientific...

c.

Post Reply

Return to “Getting Started (AKA What is the best...?)”