Receive Stutter

G3NPI
Posts: 39
Joined: Wed Apr 12, 2017 3:23 pm

Re: Receive stutter

Postby G3NPI » Sat Jun 10, 2017 2:10 pm

Just installed a new i7 motherboard (to replace i5) and noticed a low (in frequency and level) stuttering/popping noise using v3.4.1 and ANAN 10. As the radio is OK on the i5 board and reading this thread I concluded it was the network. The difference being I am using a local wired network.
Using Task Manager to look at the network useage, the i7 board with the stuttering noise shows 90% network useage and the i5 board 9% network useage.
So I am assuming the i7 has a network card which will not run above 100 Mb/s .I have ordered a 1Gb/s network card in the hope that will fix it.
Or maybe I have wasted my money.....

73

Geoff G3NPI
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Receive Stutter

Postby w-u-2-o » Sat Jun 10, 2017 4:28 pm

Geoff,

If you are running PowerSDR 3.4.1 then you are running Protocol 1, which only runs at speeds less than 100MBit/s. So the lack of a GigE interface should make no difference.

Also, it seems very unlikely that an new i7 motherboard does not have a GigE NIC on it. What make/model motherboard is it?

It certainly seems like you have a problem that is not hardware related.

When you are not running PowerSDR what is the network utilization? I recommend you use resmon to measure network utilization.

73,

Scott
G3NPI
Posts: 39
Joined: Wed Apr 12, 2017 3:23 pm

Re: Receive Stutter

Postby G3NPI » Sun Jun 11, 2017 3:39 pm

Well Scott, I have at last found why one PC shows a Network Utilitation of 90% and the other 9%. It is due to the 3 feet of Network cable linking the PC to the Router. Other than being a black cable, it is commercially made and looks like the other cables I have in my box. I suppose it was the ratio of the numbers which made me jump to the wrong conclusion about network card speeds!! Replacing the cable fixed the speed problem, I will now change back to the original i7 PC and see if the stutter has been fixed !!

73

Geoff G3NPI
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Receive Stutter

Postby w-u-2-o » Sun Jun 11, 2017 6:06 pm

Hard to find, but easy to fix! Good job!

73!

Scott
G3NPI
Posts: 39
Joined: Wed Apr 12, 2017 3:23 pm

Re: Receive Stutter

Postby G3NPI » Fri Jun 30, 2017 3:12 pm

Unfortunately I still have RX and TX stutter on the i7 machine. I can reduce it to an acceptable level by reducing the Primary Audio Sample Rate to 192000, but this of course reduces the visual band width. It must be down to a difference between the PCs, which is not obvious to me. In case someone can spot something, the PCs are as follows:

RIG: ANAN10 with HPSDRv3.4.1

PC 1 (with stutter)

Win 10 Pro
500Gb SSD
500Gb HDD
On Motherboard dual video (VGA+Digital)
i7-7700 CPU @ 3.60Ghz
RAM 16 Gb
64 bit system
Around 12% CPU usage running HPSDR


PC 2 (runs well with a 384000 Sample Rate)

Win 10 Pro
250Gb SSD
500Gb HDD
Separate dual video card (VGA+Digital)
i5 CPU 760@2.8Ghz
RAM 16 Gb
64 bit system
Around 50% CPU usage running HPSDR

Or perhaps it is somthing else ......

73

Geoff G3NPI Buckingham
User avatar
Tony EI7BMB
Posts: 651
Joined: Sun Apr 09, 2017 2:31 pm
Location: Dublin
Contact:

Re: Receive Stutter

Postby Tony EI7BMB » Fri Jun 30, 2017 4:30 pm

Perhaps try comparing the ethernet network settings on both PC's to see if there are any differences ?
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Receive Stutter

Postby w-u-2-o » Fri Jun 30, 2017 4:47 pm

Geoff,

More data is needed. Download and install LatencyMon.

Then set up the PC in question with PowerSDR running with the radio and everything else you normally have running along with it on the PC. Then run LatencyMon and start collecting data.

Your hardware looks quite good, much better on the new machine, with the exception of the video adapter. It would be quite interesting to take the video card out of the old machine and put it in the new machine, if you have an appropriate slot for it, and disable the motherboard video on the new machine.

If LatencyMon reports problems then do be sure to read, in detail, all of the information available at the LatencyMon web site about how to use LatencyMon to track down the source. It is often a driver problem somewhere on the machine. Video drivers are frequently the culprit, although it could be anti-virus software or any number of things.

See also this thread: viewtopic.php?f=37&t=2473

73!

Scott

P.S. The "primary audio sample rate" is no such thing. That is the IF sample rate to/from the radio, each sample being a set of paired I and Q values. The audio terminology is a very unfortunately hold-over from the days when early SDRs delivered their IF data into the PC as actual analog audio data streams. To this day the PowerSDR user interface has still not been updated!
User avatar
WA0VY
Posts: 83
Joined: Sun Apr 09, 2017 3:41 pm

Re: Receive Stutter

Postby WA0VY » Fri Jun 30, 2017 4:48 pm

Geoff, could you download and run LatencyMon and share the results? That should identify potential driver and process issues causing latency and/or page faults which are usually at the heart of the problem. Also, see my recent posts in the LatencyMon Benchmarks thread. You will see some good and bad (more like horrid) results. Usually, the issue is related to component compatibility.

Also, please identify your motherboards and video cards.
73 Brent WA0VY
G3NPI
Posts: 39
Joined: Wed Apr 12, 2017 3:23 pm

Re: Receive Stutter

Postby G3NPI » Sat Jul 01, 2017 4:59 pm

Many thanks for the replies. I have started to work through the suggestions and accumulate data. However whilst using rescom I noticed it said the CPU was running at 1Gb instead of 3.6Gb probably because of a temperature problem. Tried to open the BIOS to have a look and the usual DEL button on boot up access does not work. So I am a bit stuck at the moment. I will have to look at the internet for suggestions. Will get back as soon as I have looked at the temp.

73

Geoff
User avatar
WA0VY
Posts: 83
Joined: Sun Apr 09, 2017 3:41 pm

Re: Receive Stutter

Postby WA0VY » Sat Jul 01, 2017 5:34 pm

Try using "F2" to get into BIOS. You have an obvious cooling problem almost certainly associated with your CPU cooling system (check connections to MB). The CPU will throttle down under those conditions. Try installing and running CPUID: http://www.cpuid.com/softwares/hwmonitor.html

It will give you a lot of diagnostic info. Let us know what you see.
73 Brent WA0VY
G3NPI
Posts: 39
Joined: Wed Apr 12, 2017 3:23 pm

Re: Receive Stutter

Postby G3NPI » Sun Jul 02, 2017 4:28 pm

Hello Brent,

I managed to get into the BIOS using the DEL key (must have tried 20 times before without success). Anyway the CPU temp looked OK at 41deg C in the BIOS.
Then tried CPUID monitor and the results are attached.

Screenshot 2017-07-02 17.05.37.png
Screenshot 2017-07-02 17.05.37.png (66.93 KiB) Viewed 20361 times

Screenshot 2017-07-02 17.18.18.png
Screenshot 2017-07-02 17.18.18.png (68.64 KiB) Viewed 20361 times


Any comments gratefully received.

Will continue with other suggestions and post results later.

73 for now

Geoff G3NPI
G3NPI
Posts: 39
Joined: Wed Apr 12, 2017 3:23 pm

Re: Receive Stutter

Postby G3NPI » Sun Jul 02, 2017 4:48 pm

Hello Scott,

I have now tried resmon as you suggest and post it for inspection.
I wonder what you think..

Screenshot 2017-07-02 17.44.08.png
Screenshot 2017-07-02 17.44.08.png (109.87 KiB) Viewed 20360 times


73
Geoff G3NPI
User avatar
WA0VY
Posts: 83
Joined: Sun Apr 09, 2017 3:41 pm

Re: Receive Stutter

Postby WA0VY » Sun Jul 02, 2017 7:08 pm

Geoff, speaking only about what I see in CPUID, what concerns me is what I'm not seeing. Unless I'm missing something, it looks like you have only one fan and its not identified as a CPU cooling fan. Is that correct? Do you have a CPU cooler? If so please identify the make and model? Is there more than one case fan? The CPU temps shown here are normal. The i7 7700 runs hotter than my Xeon but I am also using water cooling on that chip and the radiator is mounted outside the case. So, that would be expected. So let me know about your cooling.
73 Brent WA0VY
G3NPI
Posts: 39
Joined: Wed Apr 12, 2017 3:23 pm

Re: Receive Stutter

Postby G3NPI » Sun Jul 02, 2017 7:59 pm

Your right Brent. The fans are not up to spec. The main CPU has a Heat sink with a working fan mounted on it, but the only other fan is in the PSU. The case does not have its own fan. Also, I wonder if I have plugged the CPU fan into the case fan position, so it is not showing up? I shall have to correct that. Thank you.

I have run Latency Mon and at first it seemed OK, then I realised that HPSDR Sample Rate was set at 19200. So I changed it to 384000 and that changed the Latency Mon picture !! see 384000 attachment. It gives lots of suggestions and, to be honest I am not too sure how to tackle the suggestions it makes. If you are able to spell it out I would be much obliged.
Specifically- A DPC routine belonging to a driver?? How does this relate to HPSDR, is it fixable?
- Looking at the network info it seems OK, or am I missing something?
- I am not using a WLAN adapter, as far as I know.
- Is disabling CPU throttling a good idea?
- What is the process for getting a BIOS update, I have only just purchased the board. Could it be old stock?

Screenshot 2017-07-02 @384000.png
Screenshot 2017-07-02 @384000.png (51.38 KiB) Viewed 20354 times

Screenshot 2017-07-02 @19200.png
Screenshot 2017-07-02 @19200.png (51.48 KiB) Viewed 20354 times


73

Geoff G3NPI
User avatar
WA0VY
Posts: 83
Joined: Sun Apr 09, 2017 3:41 pm

Re: Receive Stutter

Postby WA0VY » Sun Jul 02, 2017 10:53 pm

Geoff, a few comments:

1. I still do not know what MB you are using. Can you provide that information please? However, it is unlikely that a bios update will solve the issue.

2. Running LatencyMon (LM) for a short period of time does not give useful data. When I had latency issues, LM would always give me a good few minutes and then explode. So, try running HPSDR at 192k for 20-30 minutes and see what happens in LM.

3. You have a fan and CPU cooling issue for sure. First, I would like to know what cooler you are using. Perhaps a picture would tell me what I need to know. Second, I could make a recommendation on a cooler but need to know about your case as some cooling solutions need fairly large cases. Third, you should have case fans installed as well to move air through the computer case, front to back. I would advise at least two (I have 5 but I'm not running a normal computer) and most cases have a provision for that on the rear panel. Fourth, based on what CPUID is telling me, your CPU fan is not plugged into the CPU fan header. Check your MB and manual for the location of that header which is normally very close to the CPU itself. That will make a difference since if the MB BIOS is controlling your fan speed, it will be sending the control signals to the CPU header, not to just any old fan header.

4. Good news, bad news on your 384k LM results. The bad news is high interrupt-to-process latency coming from somewhere. I had that problem and never found it but it disappeared in my new build. And, I never knew why. If you Google it, you will find that it is very nebulous and not particularly associated with any driver or piece of hardware. More bad news: storport.sys latency is killing you and has been identified with exactly the problem you are having (i.e. drops and pops). Google this phrase: "storport.sys dpc latency." You will see a lot of primarily audio blogs where people discuss this issue and some solutions they have found. I suspect it is a very individualized issue and depends on the build and OS. Some have reported solving the issue by reloading Win 10. Also, there's is no way to know from the info sent if there are other drivers also causing high dpc latency. You can check that by going to the drivers tab in LM after it has been running a bit and sort for highest latency. Your problem could be just storport.sys or you could have a gaggle of other drivers waiting right behind it, all with high latency. So, now the good news: no page faults! Also, I didn't see anything jumping out at me regarding the network itself but Scott will have to report back to you on the resmon results. I have not used that software before and have never had any network issues to work through myself.

5. Do NOT disable CPU throttling until your cooling situation is resolved. You could fry your CPU.

6. With time and money, everything is fixable. It took me a lot of both.
73 Brent WA0VY
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Receive Stutter

Postby w-u-2-o » Mon Jul 03, 2017 1:27 am

What Brent said, plus:

- I never suggested running ResMon. I specifically suggested running LatencyMon, and you did, and with predictable results.

- Is it fair to say that you have no audio glitches at 192K sample rate?

- I see no serious problem with your CPU temperatures. They are typical for an i7-7700 in an average cooling environment, and, even if you think they are too high, they are certainly not anywhere near high enough to cause CPU throttling to occur.

Bottom line: you need to fix the driver problems that are killing your ability to get good performance. This is a highly individualized problem. Google is your friend, here. Try to find people that are having the same problem as you and emulate their solutions.

73,

Scott
G3NPI
Posts: 39
Joined: Wed Apr 12, 2017 3:23 pm

Re: Receive Stutter

Postby G3NPI » Mon Jul 03, 2017 9:17 am

Many thanks for the detailed response.

1. The MB is a GIGABYTE GA-H110M-S2H

2. Tried running LM for nearly an hour at 192000 and as predicted it shows errors.

Screenshot 2017-07-03 RX@192000.png
Screenshot 2017-07-03 RX@192000.png (63.03 KiB) Viewed 20325 times


3. Regards cooling.

Picture of fan:-

IMG_7918.JPG
IMG_7918.JPG (1.54 MiB) Viewed 20325 times


From what you say my case is not really suitable as it has only one unused fan position at the rear. I will fit a fan there and plug it into a socket on the MB labelled SYS_FAN. Then see how it goes in terms of a longer term plan.

The CPU fan is plugged into a socket called CPU_FAN, so something odd going on there.

4. OK. I will Google this and be prepared for a long struggle. I looked at the LM driver tab and attach a copy. The numbers do not mean much to me, but the network DPC count seems large compared with the Storage Port Driver. Is that a clue?
What does DPC stand for?

Screenshot 2017-07-03 LM Driver.png
Screenshot 2017-07-03 LM Driver.png (94.49 KiB) Viewed 20325 times


5. I will not disable CPU throttling. It did not sound a good idea when I first read it.

6. Unfortunately everthing has gone up 20% here, apart from pensions.

73

Geoff G3NPI
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Receive Stutter

Postby w-u-2-o » Mon Jul 03, 2017 11:14 am

Geoff,

Number one recommendation: go through all of your drivers with a fine tooth comb and make sure they are all fully up to date. Mostly that means going to the motherboard manufacturer's web site, downloading and installing all the latest drivers for your chipsets, etc.

Again, try swapping the video card from the old computer into the new one, if you can.

And I still don't see any dramatic cooling problems.

73,

Scott
User avatar
WA0VY
Posts: 83
Joined: Sun Apr 09, 2017 3:41 pm

Re: Receive Stutter

Postby WA0VY » Mon Jul 03, 2017 2:38 pm

Geoff, First, this is the definition of DPC from Wikipedia:

"A Deferred Procedure Call (DPC) is a Microsoft Windows operating system mechanism which allows high-priority tasks (e.g. an interrupt handler) to defer required but lower-priority tasks for later execution. This permits device drivers and other low-level event consumers to perform the high-priority part of their processing quickly, and schedule non-critical additional processing for execution at a lower priority.

DPCs are implemented by DPC objects which are created and initialized by the kernel when a device driver or some other kernel mode program issues requests for DPC. The DPC request is then added to the end of a DPC queue. Each processor has a separate DPC queue. DPCs have three priority levels: low, medium and high. By default, all DPCs are set to medium priority. When Windows drops to an IRQL of Dispatch/DPC level, it checks the DPC queue for any pending DPCs and executes them until the queue is empty or some other interrupt with a higher IRQL occurs.

For example, when the clock interrupt is generated, the clock interrupt handler generally increments the counter of the current thread to calculate the total execution time of that thread, and decrements its quantum time remaining by 1. When the counter drops to zero, the thread scheduler has to be invoked to choose the next thread to be executed on that processor and dispatcher to perform a context switch. Since the clock interrupt occurs at a much higher IRQL, it will be desirable to perform this thread dispatching which is a less critical task at a later time when the processor's IRQL drops. So the clock interrupt handler requests a DPC object and adds it to the end of the DPC queue which will process the dispatching when the processor's IRQL drops to DPC/Dispatch level.

When working with streaming audio or video that uses interrupts, DPCs are used to process the audio in each buffer as they stream in. If another DPC (from a poorly written driver) takes too long and another interrupt generates a new buffer of data, before the first one can be processed, a drop-out results.[1]"

(Emphasis added). The other drivers showing latency may not be causing your issue. Other than storport.sys, only ACPI.sys has a highest execution time which could be of interest and that did not occur very often. It is likely that if you resolve the storport related issue, you will be in good shape. You may want to read this article which explains what storport.sys does:

https://docs.microsoft.com/en-us/window ... ort-driver

Based on my understanding of storport (and dumbed down to what a lawyer can understand), your issue may not be the actual driver but rather how it is handling data it controls. I noted that you have both an SSD and a hard drive in your system. I'm assuming that your OS is on the SSD and the hard drive is for storage. If that is correct, and just for experimental purposes, lets see what happens if you disconnect the hard drive and run LM again with the sample rate at 192k.

As for cooling, it looks like the fan may have come with the i7 chip and is plugged into the MB correctly. The stock cooler is OK but if you experienced an instance of CPU throttling, then your CPU temp had to be near 100C. At this time, my advice is to get some case fans installed and keep an eye on the CPU temp. If the stock cooler plus case fans can't keep you running at a max of 65-70C, I'd look at a better CPU cooler. As for the single chassis fan header, if you have two fan spaces on your case, you can use a y-cable to run two fans from the same header. If you do so, be sure that only one fan has four leads connected and the other only three. Otherwise the BIOS may get confused as it measures the fan speeds. Many y-cables will have that already taken care of.
73 Brent WA0VY
G3NPI
Posts: 39
Joined: Wed Apr 12, 2017 3:23 pm

Re: Receive Stutter

Postby G3NPI » Mon Jul 03, 2017 2:54 pm

Great news. The stutter has gone (fingers crossed).

I updated 3 MB drivers. Each tested with HPSDR at 384000 + N1MM+VAC running

1. Audio. 1hr test shows DPC execution time down to 395us. Process latency still high at 8200us. A good start.
2. Management Engine+Serial Over LAN+Local Management Service. Process latency still high at 1980us but improving.
3. Ethernet Controller Driver. After this update the Latency Mon gave the picture in the attachment. Not perfect I know, but the rig works properly
and that was my aim.

A very big thank you to both of you for sticking with it and providing clear guidance.

73

Geoff G3NPI

Screenshot 2017-07-03 After MB driver updates.png
Screenshot 2017-07-03 After MB driver updates.png (51.53 KiB) Viewed 20294 times
User avatar
WA0VY
Posts: 83
Joined: Sun Apr 09, 2017 3:41 pm

Re: Receive Stutter

Postby WA0VY » Mon Jul 03, 2017 3:01 pm

Geoff, this looks very good. There is always going to be some latency, but as long as you're in the green, you should be virtually stutter free. Great job!!
73 Brent WA0VY
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Receive Stutter

Postby w-u-2-o » Mon Jul 03, 2017 4:51 pm

Great job, Geoff :)

Keep updating/grooming those drivers and maybe you'll knock it down some more. To stay on the cautious side, see if you can obtain the current driver installer as well, that way if a new driver makes something worse you can always go back.

73!

Scott
W4WMT
Posts: 325
Joined: Sun Apr 09, 2017 10:12 pm

Re: Receive Stutter

Postby W4WMT » Tue Jul 04, 2017 12:40 pm

I was hoping that one of you LatencyMon jocks could explain to me why it's reporting tons of hard page faults, even though I have the vmm system disabled!!??!!

73, Bryan W4WMT
User avatar
WA0VY
Posts: 83
Joined: Sun Apr 09, 2017 3:41 pm

Re: Receive Stutter

Postby WA0VY » Tue Jul 04, 2017 2:15 pm

By "vmm" system do you mean the page file? If you are trying to force swapping from the hard drive to RAM, I can tell you it doesn't work and will not solve an excess page fault issue. Been there, done that. I was only able to solve the page faulting issue by going with an M.2 SSD. With an SSD on the M.2 bus (not sure "bus' is the correct word when talking about an M.2 connection), data moves direct to/from the CPU. You no longer use the SATA bus at all for operating, just for storage if you have SATA devices. Scott also has an M.2 SSD and is not seeing any page faulting. I'm not sure Geoff's is a SATA or M.2 version; they make both. M.2 connections offer a huge speed advantage as well. In my case 4-5 times faster than my SATA SSD, which was itself way faster than even my SCSI HDD.
73 Brent WA0VY
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Receive Stutter

Postby w-u-2-o » Tue Jul 04, 2017 5:14 pm

The reason is because there is paging that goes on that has nothing to do with the page file. All kinds of things page in and out of memory, only certain things use the page file area.

This is why it is never a good idea to disable the page file, it really doesn't buy you anything and can actually hurt performance because memory starts getting hogged up with stuff that would ordinarily be, and should be, paged out.

73,

Scott

Return to “PowerSDR mRX”