Polling Rates Almost all Wrong

[Keith@HPTuners]

Can you post your log and channel config for the 6 channel setup? Like I said before, just because you want 10ms doesn't mean the PCM will give it to you.

Also, for your particular pcm, 2048K Copperhead, you have room for about 8 parameters in the buckets. Everything else will be requested one by one which is a much slower process.

Please understand that what you ask for via interval and what you get isn't software controlled, it's controller restricted. In you case, because the buckets are small, and you are trying to log a lot of parameters, and to further add, you have many parameters that share the same interval, the software will struggle with predicting what you want.

[Keith@HPTuners]

btw, it appears that parameters are being added to the buckets in a top down order from the channels list.

This is something that should be easy for you to test.

I'm looking into a solution.

[Keith@HPTuners]

Tomorrows beta should give the lower interval parameters priority to the faster methods.

[CCS86]

Can you post your log and channel config for the 6 channel setup? Like I said before, just because you want 10ms doesn't mean the PCM will give it to you.

Also, for your particular pcm, 2048K Copperhead, you have room for about 8 parameters in the buckets. Everything else will be requested one by one which is a much slower process.

Please understand that what you ask for via interval and what you get isn't software controlled, it's controller restricted. In you case, because the buckets are small, and you are trying to log a lot of parameters, and to further add, you have many parameters that share the same interval, the software will struggle with predicting what you want.

Hmmm, that doesn't make much sense. 22 of my parameters are logging at 31 samples / sec. (most are requests at 5 or 2 Hz)

btw, it appears that parameters are being added to the buckets in a top down order from the channels list.

This is something that should be easy for you to test.

I'm looking into a solution.

I did try reordering my channels list with higher priority items at the top... no change in the result.

Tomorrows beta should give the lower interval parameters priority to the faster methods.

Keep us posted!

[dermotw]

Hmm, I just looked at mine (v3.6 and with many channels, most set quite slow). Same kind of thing. Basically, the sample rate is all over the place. What disturbs me is that apparently setting many channels to slow rates actually does not help. Whats the point of being able to set e.g. every 5s for coolant temp if it still retrieves it every 200mS???

I don't think my old 1999 LS1 ECU is even doing things like broadcasts, like later/other ECUs.

Below my XL with the channel data vs. 60 of real data...

channel_ints.xlsx

-D

[dermotw]

Ok, update, I made a few more logs to test, logging only 4 channels. No matter what interval I set, it chooses its own log rate and its nothing like what I set....

[CCS86]

Check out the latest beta.

They definitely improved the polling rates!

[dermotw]

ok I will...

[dermotw]

Ha, yes, it "behaves as expected" now in the Beta. Slow = slow and fast = fast (more or less...). -I wonder how the hell it was not before... those setting have been around a looong time... oh well...

-D

[CCS86]

I wonder how the hell it was not before... those setting have been around a looong time... oh well...

-D

I was pretty surprised to find that behavior! Sometimes my unwillingness to take no for an answer helps out.