Multithreading
Echoview's underlying architecture has been re-engineered to provide multithreaded data reading and calculation, resulting in a faster and more responsive experience.
On this page
# Next-generation computing
# Data reading
# Data calculation
Major version increments are the time when we have traditionally made big changes to Echoview. Echoview 5.0 follows this pattern with what, under the hood, is one of our biggest improvements yet - multithreading, or the ability to perform multiple tasks at the same time.
Multi-core processing power
Large datasets and increasingly-complex processing algorithms demand greater computing power. We’ve made a major overhaul of Echoview’s entire data-processing engine to support the next generation of high-performance multi-core processors.
This means that Echoview can calculate your data much faster via its multithreaded operators, and you can use and interact with Echoview and the rest of your computer at the same time.
Out with the old, in with the new
Echoview versions 4.90 and older used only one processor core. So while Echoview was reading data or making a calculation, you couldn’t interact with it. For those with multi-core processors (now common with all new computers), Echoview versions 4.90 and older couldn't take full advantage of the available computing power.
We have addressed these restrictions with our new architecture. Echoview 5.0 is now faster, smoother and better than ever, and will continue to improve over coming releases.
The first thing you might notice when you open version 5.0 is that Echoview continues to be responsive and interactive when you add a data file.
Data reading (scanning a datafile and adding its pings) now happens in the background – not with any lower priority, it’s simply that Echoview can do these two things at once.
Multi-tasking
When you add a large datafile, a new progress bar will appear in the lower-right corner.
You can continue to work with Echoview while this occurs. For example, you can start creating dataflows and viewing your data quicker.
When each datafile is fully read, its pings are seamlessly appended to the variables. So when you are loading large amounts of data - whether it’s in a few large files or spread over dozens of smaller files - you won’t have to wait for Echoview the way you used to; you can begin work immediately.
The most computationally-intensive task that Echoview performs is calculating data for virtual variables.
So in addition to multithreaded data reading, Echoview 5.0 can now perform multithreaded virtual-data calculation. This means that echograms generated with multithreaded operators can take less time to load on faster computers, and you can continue to work with Echoview and your computer while calculations are ongoing.
Multithreaded operators
Virtual-variable operators that can take advantage of this new framework are marked on the New Variable dialog with the icon of a running hare.
In Echoview 5.0, about a third of the operators are multithreaded. We will continue to convert the remainder over future releases.
Dataflows containing single-threaded operators
Mixing multithreaded and single-threaded (old-style) operators in a dataflow is fine. However, please note that the single-threaded operators cannot run any faster than before. They may therefore limit the benefits of the multithreaded operators until they are converted.
In general terms, the more multithreaded operators you use, the faster you can get to your data and the more responsive your computer will be while calculations are occurring.
