Why does my application slow down?

[roberto.venetz]

I found how to set 0.09 for X-Axis step. Then I tried your program but it's worse than the one I made.

- I want to choose to save or not at the end of the execution;

- nothing is displayed on the graphes.. the Y-Axis is set as auto-scale but is doesn't work. thus, I don't see anything;

- I'm really surprised but your program is slower than mine, I have the time to read all the values before they change (it was too quick with my application).

I'll start with mine again and initialize shift registers as it's done in your program. Hope it will bring amelioration.

[altenbach]

The running speed of the program is limited by the BT communication speed. To compare the speed of the different code version, make a benchmarking version that uses locally syntesized data.

How do you measure speed?

Remember that if you allocate a fixed array size and use autoscaling on the x-axis, you might not see much until about one pixel wide worth of data has arrived.

Have you tried reading 6 bytes at a time via BT?

[roberto.venetz]

Hey!

The Bluetooth communication is not the problem. I tried my application without storing data and it works perfectly (correct rate and it doesn't slow down). We now can be sure than it's all about memory space allocation.

[altenbach]

---

roberto.venetz wrote:

The Bluetooth communication is not the problem. I tried my application without storing data and it works perfectly (correct rate and it doesn't slow down). We now can be sure than it's all about memory space allocation.

---

Of course it is, but you are twisting my words around. You were saying that your allocation scheme is superior to mine. Is it still? I was suggestion to test the allocation codes independently from BT for a fair comparison. Did you? What are the timing differences as a function of output size?

[roberto.venetz]

I'm able to determine a time difference to make a comparison. By performing some tests I guess both our scheme are the same, or almost. I may have spoken too quickly.

[roberto.venetz]

I'm NOT able to determine a time difference to make a comparison. SORRY

[roberto.venetz]

I continued to perform tests and I was able to determine what was really slowing down my application. The real problem is the time the PDA needs to refresh the screen. I tried my application without the waveforms, displaying only the numeric value and the work rate is respected.