sob@soma.UUCP (Stan Barber) (12/04/86)
This meeting, sponsored by the South West District Sales Office and MUS, was attended by more than 60 people from the area. Most were from Houston, but others came from Dallas, Austin and other areas. BOF sessions were held discussing graphics, data acquisition, communications/LANs, and Multiprocessors/Computation Accelerators. These were well attended. Some news coming out of those sessions: 1. X-Windows using the V10 protocol will not be released as a formal product, but will be available through the MUS Library. X-Windows will be released as a formal product once MIT releases a version using V11 of the X-protocol. X-Windows using V10 protocol should be available by the time of the MUS International meeting. 2. The VA-1 Vector Accerlator was discussed in detail. It features: a. 14 Mflops single precision (32-bit) floating point [1024 point complex FFT in 4.3 ms] b. 31 K element vector memory c. DMA access to all physical memory d. Overlapped DMA and math operations e. IEEE-754 floating point format f. Source code compatibility with AP-501 applications It is available for shipment at the end of this month. It still requires special coding and interlock considerations when programming to use the VA-1. 3. The memory buffer allocation problem on the EXOS-201 board is a top priority problem being addressed by the networking group. Once this problem is solved, NFS will be implemented. It was unclear whether this was going to be addressed by better software or by changing the hardware or some mix of the two. We also understood that a second ethernet vendor was being considered, but no annoucement was being made at this time. 4. The 5700 officially supports 4 CPU's, but it will actually do more. The reason that it has not been announced that it will do more is because the details surrounding bus saturation are not well understood. There was a question and answer session before the luncheon. Unfortunately, I had to miss that. If someone who attended would like to summarize that part of the meeting. Please do so and submit it. Following the luncheon, there were two user presentations. From Ford Aerospace, Judith Backman described the challenges in doing "Externally Driven Real-time Data Acquisition" for payload monitoring on the Space Shuttle. From Baylor College of Medicine, Dan Johnston (dan@soma.bcm.tmc.edu) described "Neuron Simulation" using the Masscomp. Frank Lebeda (frank@soma.bcm.tmc.edu) demonstrated the single neuron model the lab has been working on. OEM's were next with product presentations. Larry Combs of Process Control Technology, INC (512-341-5028) described the Process Control Operating System. The PCOS uses four processes and shared memory to acquire data and log it while doing conversions and displaying it. The system is built to be user-customized and is largely mouse-menu driven. Dave Mandelkern of V.I. Corperation (413-253-3482) presented the latest version of DataViews featuring DV-Tools, DV-Draw and DV-Source. These products allow easy creation and modification of graphical displays for process control, prototyping, and animation. Larry Stocker of IMPRES (512-346-1209) described the IMPRESS DS 6000, a system that allows easy digitization of blueprints and aperture drawings, easy on-line modifications of these drawings and reproduction of the original or the modified material on an attached plotter. Masscomp made presentations on Multiprocessing and the new data acquisition package called Lab Workbench. Also, Tom Aiken presented details on the HELIOS graphics system (a product that MASSCOMP OEM's from General Electric).