art%mwvms@MITRE.ARPA.UUCP (06/03/86)
-------- >We are planning a conversion from VMS 3.7 to 4.0, with future conversions >taking us to the current VMS version. We support many RSX/IAS tasks >including the TKB associated with VMS 3.0 (don't ask why, we must be >able to continue using it). I have heard that AME does not support as >many executive functions as are supported in VMS 3.7 but have not been >able to find a specific list of changes from 3.7 to 4.0 (the SPD lists >changes from a real RSX to AME but not what we need). Any info on this, >or any other pitfalls of going to 4.0 (user-written drivers etc.) would >be greatly appreciated. I agree completely with the comments by the RSX SIG person on the history and goals of the AME, and will not repeat them. We found that when we upgraded from the VMS 3.7 AME to the VMS 4.0 AME that we did not run into problems with the calls to the executive. We were able to continue running the IAS TKB. Most, if not all, AME images run 5-20 times slower under VMS 4.0 AME. It was sufficiently worse that it did result in the transformation of many things to native mode. It even resulted in the upgrading of several PDP's to VAXes. I have no benchmark on the second version of the AME. We however have no driver experience. * *---Art * *Arthur T. McClinton Jr. ARPA: ART@MITRE *Mitre Corporation Phone: 703-883-6356 *1820 Dolley Madison Blvd Internal Mitre: ART@MWVMS or M10319@MWVM *McLean, Va. 22102 DCS: MCCLINTON *
LEICHTER-JERRY@YALE.ARPA.UUCP (06/06/86)
... We found that when we upgraded from the VMS 3.7 AME to the VMS 4.0 AME that we did not run into problems with the calls to the executive. We were able to continue running the IAS TKB. Most, if not all, AME images run 5-20 times slower under VMS 4.0 AME. It was sufficiently worse that it did result in the transformation of many things to native mode. It even resulted in the upgrading of several PDP's to VAXes. I have no benchmark on the second version of the AME. ... * *---Art The first version of the new RSX AME - V1.0, or did they start it out at V2.0? - had severe performance problems, mainly in the I/O system. (This had to do with the way RSX device names were translated back and forth to VMS device names; the AME ended up doing many, many logical name translations.) I believe this may even have been discussed in the release notes - though it could have been in the release notes for the "next" version, when the problem was fixed. Make sure you have a current version of the product.... -- Jerry -------