9 thoughts on “Testing Manifold Systems, Ltd. Experimental Parallel Processing Engine (Radian) – along with some other products

  1. Nice post Art. Three tiny points.
    (1) Manifold Software Ltd, not Manifold System Ltd.
    (2) Please name me if you want to quote me. Real name or user name both fine.
    (3) If you want to edit my words, please ask first. Though you did a nice job–I couldn’t have done better.

  2. Thanks Tim. I was going to ask your permission on the forum, but given the time change I put you in anonymously as I wanted to get this posted. You beat me to it, and I will take your post here as a permission to use your name. That was my intention all along.

    I will, however, maintain the editing of the words, as it is my paraphrase of what you said so that it is better understood by the readers of the blog. In no way was the editing done to put words in your mouth or change the way you said something, but rather to clarify things for the typical reader of this blog. I think your comments were spot on and insightful.

    • Cool Art, I thought you were only being extra polite, and you improved my wording no end. Any chance to collaborate is good.

      It’s really great we can talk about this “out loud” to some degree. Thanks Dimitri!

  3. Is there any chance you could test a function where neighbours are involved?

    I’m targeting limits of parallelization of the difference we see in Mfd8 between the GUI function NormalizeTopology(geomSet) and the SQL function Normalize(geom), the later producing gaps and not solving all overlaps just now without parallelization while the sophisticated process behind the GUI function does not.

  4. Great to see. You might not be able to answer, but I see your setup doesn’t include GPU capability. If present, would this further influence the Radian timings?

  5. Dan, actually, my setup does have a Fermi card – so it is making use of the GPU. However, for the task that I outline above, I am not certain if it is making use of the GPU. We have written our own GPU software that I will write about next week. I’m not certain that the GPU is actually doing anything in this task.

  6. Art, thanks for posting about this although I find it more discouraging then encouraging as it seems the product is no where near ready for release. While saving time with speed increases are always welcome, to date all the parallel processing work has just delayed bug fixes and new features that could have saved loads of time. That’s what I’d encourage you to pass along to Manifold in a very tactful way :-) In the meantime, thanks again for posting and all that you do with GIS.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s