Re: bug feedback - post newly found progdvb bugs here

#916
Prog wrote:Please send me your plugins directory with this plugin.
[Deleted by Prog]
Windows: 7 Service pack 1 , 32bit _-_ DVB: Sky Star 2.8a PCI _-_ CPU: 2.6 GHZ Dual core X2 AMD _-_ RAM: installed 4Gig , 2.75 Gig usable _-_ Graphic: ATI Radeon x1200 series total available memory : 1663

Re: bug feedback - post newly found progdvb bugs here

#917
Prog wrote:
reeven wrote:i did not know that for PIP and mozaic to work, i need to :

in timeshift mode is set to : "All channels from transponder store in buffer" OR "Not Close Device and buffer of old channel when open new"


Now pip and mosaic is working in v6.73,v6.74, and v6.80.
I am not have idea. Please select "one channel in TS" mode and send me log of bad mosaic and PIP operation.
what do you mean select one channel in TS, what we should do to help 8) :wink:
Windows: 7 Service pack 1 , 32bit _-_ DVB: Sky Star 2.8a PCI _-_ CPU: 2.6 GHZ Dual core X2 AMD _-_ RAM: installed 4Gig , 2.75 Gig usable _-_ Graphic: ATI Radeon x1200 series total available memory : 1663

Re: bug feedback - post newly found progdvb bugs here

#919
Hi
version 6.80, I can't set Diseqc switch setting and make crash .
option of Add Switch doesn't appear so you have to remove the selected one by default first .
so if the option appear directly when do you want to manage it first time would be simply easy .

Re: bug feedback - post newly found progdvb bugs here

#922
I'm trying to search this TP : 11623 V 27500 27500 3/4 on Hotbird 13^E to bring the "RTL 102.5 TV" to my channel list.scan this Freq with "manual search", channel find and i can see it in the list of founded channel on this Freq:(see Pic)
http://up1.iranblog.com/images/n37hj7ivkvbkdpgvg04.jpg
but when i click OK on this window, i cant find the "RTL 102.5 TV" on the my channel list.
also its not in the undelete window. tried with different setting but wasn't useful.need help.tnx
Windows: 7 Service pack 1 , 32bit _-_ DVB: Sky Star 2.8a PCI _-_ CPU: 2.6 GHZ Dual core X2 AMD _-_ RAM: installed 4Gig , 2.75 Gig usable _-_ Graphic: ATI Radeon x1200 series total available memory : 1663

Re: bug feedback - post newly found progdvb bugs here

#923
I've discovered a very strange issue in ProgDVB. It's there since a very long time (I think it was there in 6.63.x and still present in 6.80.3).

I found that ProgDVB is using quite a lot of CPU resources. It's been using continuously > 20% CPU on my Intel Q6600 CPU. As it is a Quad-Core this means almost full load on one of the cores just by ProgDVB. If this load is caused by single thread, then it might sometimes even hit the limit of what one core can process.

Accidentially I've discovered that CPU load seems to be heavily dependent on Timeshift settings. Although in very weird way: When Timeshift is disabled (my preferred settings) then CPU load goes up (I would have expected it a bit lower if Timeshift is not used). If set to default settings of "Only current channel in buffer" then CPU load trops to 5-10% on the same machine.
Strangely also the settings "All channels from transponder store in buffer." and "Not close device an dbuffer of old channel when open new." cause increased CPU load.

Also changing the buffer location or even buffer size causes this increased CPU load.

As I am often not using timeshift I am using ProgDVB as s aimple display tool and like to keep memory footprint low. Even with Timeshift disabled ProgDVB has a quite large memory footprint of 250MB on my system. With timeshift it grows to ~500MB of course. I am a software developer too but somehow a footprint of 250MB seems to be quite high for my understanding, even taking into account various I/O buffers etc.

So I would have two questions:
- Is the increased CPU load on any derivation from the default Timeshift settings intentional? Why?
- Any possibilities to reduce memory footprint?

Re: bug feedback - post newly found progdvb bugs here

#925
1. No. I am usually using Microsoft Demux. I ought to remember that there were some fixes in Prog demuxer recently. I just tried switching to Prog Demuxer. The issue remains. As soon as I select Timeshift mode: "Disabled" then CPU load goes up to 20-30% (full load on a single CPU in a quad-core environment). So I even think that it might be some idle-loop or similar which causes 100% load on a single core (adds 25% load on a Quad-Core machine).

2. I might have been wrong regarding buffer size changes. At least in current version I can reduce the in-memory buffer to 1MB (0 seems to work too) which seems to disable the timeshift functionality too. At least the timeshift-bar is inactive then. Memory footprint is then pretty stable at ~250MB; which is still a lot.

So currently disabling timeshift by setting the in-memory buffer to 0 seems to work while disabling it in the "Main" section still seems to be broken.

Using ProgDVB 6.80.03 x64 on Windows 7 x64

Re: bug feedback - post newly found progdvb bugs here

#927
Hi prog

plaese please ...
make older prog dvb channels compatible with newer version's
i have prog 6.51.7 and my channel list doesn work in 6.8x versions .
the problem is that most of the channels in my channel list doesn't chenge (Switch) when i select the prefered channel .and i have to reselect the correct switch option in channel properties to make it show the channel.
thx

Re: bug feedback - post newly found progdvb bugs here

#929
reeven wrote:on my quad q6600@3.2ghz, i have 6-10% comsumption in progdvb.
From time to time it raise to 11-12%.
If you have 20%, you could raide your FSB at 333x9 mean 3.2ghz.Then you will have more power.
Q6600 at 2.4ghz is slow.
I think you didn't get the point here. It's not about the fact that my machine is too slow for ProgDVB to run (which would be ridiculous anyway for SDTV). It's about the fact that when I switch the timeshift mode - or even disable it - there seems to be a looping thread or similar which puts full load on a single core (+25%). No matter how high my CPU clock will be, one of the CPUs will be fully loaded by this thread. So even if one has a 10+ GHz machine it will be 25% loaded.

So this is a bug to be fixed.