2.5x speedup in analog processing

2.5x speedup in analog processing

Hi All, just quick note to mention that in our most recent beta release the performance of analog processing has improved x2.5. Based on your computer, we're using SSE2, AVX2, FMA3, or FMA4 SIMD instructions; as a result, analog filtering is no longer the bottleneck in analog performance. (now there is a new bottleneck of course) I think we can probably improve it by at least that much again before hitting what I like to call 'maximum theoretic performance'.

It's taken a while but it looks like this week we can push the first sneak-peak version of the in-progress UI swap-out. Since we want to ship as frequently as possible, we will ship 'transitional' versions that will look a bit weird because they will be a mix of UI styles -- but it'll give people a chance to see where things are heading and provide early feedback.

In addition to the UI refresh we're working a new application state management system as well as an incredibly cool new "nodes and pipes" data processing back end. Our aim is to ship what we internally call 'first class real time view' and a realtime oscilloscope-like view by the end of this quarter.

Should have the next update in a week. Until then!

Joe

Reading next

LED protocol analyzer for WS2811, WS2812B, WS2813, TM1809, TM1804, UCS1903, and LPD1886.