Tim_Buxton@UM.CC.UMICH.EDU (04/21/89)
We need a workaround, or better, SGI implementation of rectread and rectwrite commands, since they are the fastest way to do the graphics moves we require. Both commands are documented as working for our G-level machine, but having the temerity to USE them brings a brisk rebuff at run-time - command not implemented. A call to the hotline produced little - yes they know of the problem, yes it will be fixed in the (alas, indefinite) future. We have a prerelease version of fortran 3.2, in which it is still not fixed. This must be bugging users by the thousands, judging by how slowly our workarounds have worked so far, and how fast the commands work on GTX machines of our acquaintance. Thanks in advance for any help we can get here. I have already found this forum very helpful, especially the apropos scripts. - Tim Buxton, OptiMetrics, Inc. Tim_Buxton@um.cc.umich.edu
archer@elysium.SGI.COM (Archer Sully) (04/29/89)
In article <4294299@um.cc.umich.edu>, Tim_Buxton@UM.CC.UMICH.EDU writes: > > We need a workaround, or better, SGI implementation of rectread > and rectwrite commands, since they are the fastest way to do the > graphics moves we require. Both commands are documented as working > for our G-level machine, but having the temerity to USE them brings > a brisk rebuff at run-time - command not implemented. A call to > the hotline produced little - yes they know of the problem, yes > it will be fixed in the (alas, indefinite) future. We have a > prerelease version of fortran 3.2, in which it is still not fixed. > This must be bugging users by the thousands, judging by how > slowly our workarounds have worked so far, and how fast the commands > work on GTX machines of our acquaintance. Thanks in advance for > any help we can get here. I have already found this forum > very helpful, especially the apropos scripts. > > - Tim Buxton, OptiMetrics, Inc. > Tim_Buxton@um.cc.umich.edu rectread/write will be implemented in terms of readpixels/writepixels in the next release of the operating system (3.2). You probably didn't get the new gl libraries with your fortran, which would explain why the fix doesn't appear to be there. Be warned: the rectread/write that has been implemented will be just as slow as what you have done yourself. The reasons for this are tied to the architecture of the G series. There is no software fix for the slowness of these operations, and the hardware fix is the GT. Best, Archer Sully archer@sgi.com "life is short, and full of stuff" -- Lux Interior --