Announcement

Collapse
No announcement yet.

LibreOffice Lands A Ton Of GPU OpenCL Functions

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #11
    Originally posted by oleid View Post
    Actually, that is what I think, too. Proof of concept for AMD. Nevertheless, this code might come handy for some people.
    Fair enough . From a research perspective and in the name of science i can't complain.

    Originally posted by oleid View Post
    Yet, at the moment it won't work very well for awk, as it would force an X11 dependency for awk. This will probably change, as soon as we get render nodes.
    Are you serious? Long live to awk!

    Comment


    • #12
      Originally posted by clavko View Post
      Ofcourse it's not an amateur project - it has a couple of millions LOC, ffs.
      Also, I'm not saying you should lower your expectations, I'm saying you
      shouldn't have any to begin with

      You see, this software is free of charge. No one said it would be
      fit for any particular purpose. You didn't pay a cent for it. Also,
      you're not forced to use it, there are many others, both commercial
      and free.

      Your personal feelings towards level of usefulness of some of the free
      contributions to a free project is really just wishful thinking. Therefore,
      either roll up your sleeves and code it yourself or pay someone to do it.

      NHF

      I should not have any expectations to begin with!? oh my...

      Comment


      • #13
        Originally posted by Alex Sarmiento View Post
        Is someone planning to do bimolecular simulations using spreadsheets?
        You would be surprised. I have been asked to implement some really silly stuff in Excel because some professor in pharmacology or something is scared of "programming languages". Try telling your boss that you refuse to write an Excel macro because his or her friend refuses to learn Matlab or Python.
        Last edited by TheBlackCat; 10-30-2013, 07:23 AM.

        Comment


        • #14
          Does LibreOffice support any time-domain to frecuency-domain transformation functions, like FFT? Having these functions to transform datasets would be very handy, and this use case would benefit greatly from anOpenlCL implementation. If I'm not mistaken, Excel does support FFT with a the Analysis ToolPack add-in.
          Last edited by newwen; 10-30-2013, 08:25 AM.

          Comment


          • #15
            Cute, but pointless, since a word processor/etc. doesn't do nearly enough actual work to benefit perceptibly from this kind of thing. In fact, it'll run *just fine* on a 286.

            What libreoffice really NEEDS.... RIGHT NOW, in an EMERGENCY, is to build support for mobile platforms, particularly ANDROID. There are nightlies for Android, but they do absolutely nothing and haven't changed AT ALL since they started being built many months ago.

            Comment


            • #16
              Since Calligra began to stabilize I'm rating all the LO stuff "meh"... Even if I still need it to send something to our beloved government...

              Comment


              • #17
                I was about to be surprised about the amount of whiners around here, until I forgot that I'm at phoronix where the majority of people who post have something to complain about.

                What does surprise me is how self-centered some of you are. Sure, the average person has no use for hardware-accelerated spreadsheets, but you can't possibly speak for everyone. I have found LO is incredibly slow in comparison to Excel when it comes to spreadsheets longer than 5000 rows and more than 20 columns wide. Just using 1 more CPU core to take some of the load would make a huge difference. The sad part is sometimes the program crashes if the calculation is big enough, but works fine in Excel. I might hate MS and the bloatware that Office is, but I do have to admit that Office is a pretty solid platform and runs nicely.

                I think this was a necessary feature to add, but it's too bad it's really only for financial purposes.

                Comment


                • #18
                  I don't get it...

                  Why are they all so big on gpgpu?

                  They should first make use of the grammar correction some cheap display engines provide:
                  http://cgit.freedesktop.org/~libv/xf...23006ce81#n981

                  No amount of 3D engine trickery can beat that!

                  Comment


                  • #19
                    Originally posted by libv View Post
                    Why are they all so big on gpgpu?

                    They should first make use of the grammar correction some cheap display engines provide:
                    http://cgit.freedesktop.org/~libv/xf...23006ce81#n981

                    No amount of 3D engine trickery can beat that!
                    But there is grammar correction, it's been around for a while, but the linux version doesn't ship with it by default.

                    Comment


                    • #20
                      Originally posted by oleid View Post
                      Actually, that is what I think, too. Proof of concept for AMD. Nevertheless, this code might come handy for some people.


                      Yet, at the moment it won't work very well for awk, as it would force an X11 dependency for awk. This will probably change, as soon as we get render nodes.
                      Ask and you shall receive?
                      http://lists.freedesktop.org/archive...er/047295.html
                      http://lists.freedesktop.org/archive...er/047296.html

                      Comment

                      Working...
                      X