MarketFeeder Pro 8.0 - BETA testing

Hi
The weird thing is the dates and times, somehow the date and time on my computer was reset. So, now I'm unsure if the computer caused Pro8 to crash or vice-versa. I will try to recreate this and see if I can of any help.

When I came to my computer today after being away for the weekend my computer was displaying its time as the year 1899!  My computer restarts each day at 7am and mfp vers 7 auto loads sometime later.  I've restarted the computer now and it has kept the correct time so I don't think my motherboard bios battery is dead, so maybe our problem is related.  (I had managed to start the latest build of vers8 as well via teamviewer so that might have been involved in the problem like you)

My settings file was also trashed in vers 8 & vers 7 when I restarted the computer.  I've sent the log to support in case it helps Oxa work out what happened

Hi Armani

so maybe our problem is related.  (I had managed to start the latest build of vers8 as well via teamviewer so that might have been involved in the problem like you)

My settings file was also trashed in vers 8 when I restarted the computer.  I've sent the log to support in case it helps Oxa work out what happened

Yeah its the same thing alright. Not a techie so I dont want to keep recreating the crash and xxxxing my puter.

larp

Can i test this version too,please?

Thanks
samuka66

armarni and LarpHager, do you have the setting "Synchronise local date and time with BetFair's" on? It's in "Monitoring Options".

armarni and LarpHager, do you have the setting "Synchronise local date and time with BetFair's" on? It's in "Monitoring Options".

Yes, I do.  

There must be something in common with those who are getting problems, but obviously it's not very widespread or else more people would be reporting the loss of markets settling and this new crash which resets the computers date and trashes all the settings.  Larp mentioned that he was running a 'deep' trigger to test vers 8, which implies some size and complexity to it.  My test trigger is over 600k in size and runs about 70 different triggers that can all put bets into the markets at certain times, so maybe the issue is to do with trigger file size.

Also my operating system is windows professional 64 bit, so maybe it's OS related.  The earlier thread in the troubleshooting section about the windows explorer recent update files which caused MFP vers 7 to behave erratically may also have been OS related, as it effected some people, but not others.

Today I'll try and do those screengrabs of markets not settling and I will also run a trigger that is only 300k in size with half the normal test bets removed, to see if that makes a difference, and if there is a culprit buried in my large test trigger somewhere, to start tracking it down.

Not sure if its related, but in post 101 of this thread you mention that with the latest build of version 8 we will need to email to request a one off password to enter into the program.  When I installed the latest build of vers 8 it didn't ask for this information and just went to the usual log in screen.

Today I'll try and do those screengrabs of markets not settling and I will also run a trigger that it only 300k in size with half the normal test bets removed, to see if that makes a difference, and if there is a culprit buried in my large test trigger somewhere, to start tracking it down.

Thank you armarni, look forward to it!

Not sure if its related, but in post 101 of this thread you mention that with the latest build of version 8 we will need to email to request a one off password to enter into the program.  When I installed the latest build of vers 8 it didn't ask for this information and just went to the usual log in screen.

That's only because I told you the serial key in a private message when you were applying for the beta-test, didn't I? You already entered this key, so the program doesn't need it any more.

Hi,

Yes Oxa, I always have it sychronised.

I have halved the trigger file size I was using, and I can see the difference already, much smoother.My trigger file was 490k. I use Windows Home Premium 64bit, if that makes a difference. I will test today on the dogs/horses

Larp

I have halved the trigger file size I was using, and I can see the difference already, much smoother.My trigger file was 490k.

I may have found a fault in the time synchronising setting that could reset your system clock to 1899 if there was a problem with getting the current BF date and time, and I'm fixing it right now.

As for the rest, well, to be honest I don't think it is realistic to run four copies of MF Pro, with all greyhound races in each, with a trigger file of that size. I cannot guarantee it won't crash with a load like this. Try opening four Internet Explorers, with 100+ BetFair pages in each of them, and see if this is going to be stable for a lengthy period of time. I did this experiment myself, and I didn't even get to open the third copy of the browser when it all went to a complete freeze.

As for the rest, well, to be honest I don't think it is realistic to run four copies of MF Pro, with all greyhound races in each, with a trigger file of that size. I cannot guarantee it won't crash with a load like this.

I will try and take a screen grab of the processor load when the markets are running this afternoon to see if it is maxing out.  My large test trigger is only monitoring horse race win and place markets from a few minutes before the off.  What is curious is that vers7 and vers6 before never had a trouble with running this large trigger, which is testament to what an amazing piece of software market feeder is!

Hi,

I agree Oxa, it was an overload. Like Armani, I was only running for 2 minutes on each dog race and 5 minutes on the horseraces, but with 3 or 4 profiles each with a different version of the trigger file, it was too much work; i have unctions of free cpu and it was boiling at 75% or more when they were all active. I don't expect any problems if I curtail the trigger files.

The good thing is I now have a pretty good idea of what an acceptable workload is, which is I suppose what testing is all about.

larp

What is curious is that vers7 and vers6 before never had a trouble with running this large trigger, which is testament to what an amazing piece of software market feeder is!

Like I said above, the crashing may have something to do with synchronising the time with BF, so let me make a fix, and then we'll see if it is an improvement.

I am also having problems with monitoring markets, with 6 I could monitor as many as I like at once, with version 7 icould only monitor 6 markets , and with 8 I can only monitor 4 and it is still not working reliably, I am sure it is something to do with the size of the application, version 6 would run at 12000 on the cpu version 7 at 28000 but version 8 is running at 50000.

hi oxa
there seems to be a problem with the retrieve bet list and account statement update in the general options.
when i am testing 2 versions i like to use test mode on the new version and live on the main version.
i have unticked both the retrieve bet list and account statement boxes in version 8, but the live bets still show, so no test bets are placed.
this works correct in all other versions.

mcbee

Hi Oxa

Halving the size of the trigger down to just over 300k doesn't appear to have made a difference, as some markets are still not settling.  I've included a screen grab of my resource monitor that seems to show that the computer has plenty of headroom left in terms of memory and processing power with mfp vers 7 & 8 actively monitoring markets, so I'm not sure that the issue is trigger size related.  



I don't know what is significant, but here are my observations:

1. The horses are listed in a different order in vers 7 and vers 8.  This is also true of other markets that settled correctly in both vers 7 and vers 8 so that's probably of no significance.  
2. Perhaps more importantly, the horses which have won in the vers 8 markets that aren't correctly settling have no sp price listed next to them.  
3. Also no horses in the vers 8 races have a last traded price recorded above the chart data, where as this info is available in vers 7 even after the market has settled.

I will now halve the test trigger again down to 150k and see if theres any difference.

A quick question.  Windows task manager shows a large difference in memory allocation between vers 7 & vers 8 (in the attached screengrab vers 8 is using 56,748k versus vers 7 using 14,652.)  Is this a function of the new api being more resource hungry or is it that  the new code hasn't been optimized yet?

that is wht I have noticed about the cpu usage it is massive in version 8, I have had version 7 running next to version 8, I autodutched 6 markets on each, obviously different markets , every bet got matched in version 7 but over half were left unmatched in version 8, it doesn't seem to be performing at the speed needed
Colin