Meeting Jim Lewis, Chuck Swart, John Ries, David Bishop, Peter Ashenden Karl Eisnhofer Action items. Proposal Who Description FT9 David For fixed point literals, get info from NASA on accuracy of 512 bit literals. Precision is need not for constants. This is based on prior e-mail hasn't been able to contact person. String literals, and qualified expression should work. Closed no action needed New action item NONE Jim Provide description of using name subtype, qualified expressions, and literals to David. >From David package problem: comparing integer to fix point numbers that is has negative bounds. New Action FT9 David Write up problem of comparing integer with fixed point / real number which has only negative indexs. DVcon paper brought up a bug in package. Has been fixed. David working on user guide. FT20 John Check over to see what needs to be done. Not Done FT25 Chuck/John Rewrite to handle () expressions as selector. Assure that "others" requirement doesn't change. Issue: length of expression does not need to be known statically Not Done NONE Jim/Chuck Coordinate w/ Steve to hold vote on FT/ISAC issues. Make sure dependent things go out together (ie FT11 & FT20) Chuck has broken ISAC into three catagories, typos, minor changes, significate IRs. Chuck will put groups on to a web site. Issue who are the voting members? Jim nothing yet. NONE Charlie Check to see that FT has reasonable descriptions. Not Done. FT31 Peter Add FILE_FLUSH function Not Done. FT16 Peter Make it an error to reference WORK within a context declaration. Not Done. FT?? ALL Code review of package updates (FT8, FT5, Floating Point) on going, Jim posted request to reflecotr. ??? Jim, David usage model for floating point package Not done. FT11 John Update to allow aggregates w/o delay and all others have delay Not done FT7 Peter, Jim Update package and define in terms of VHPI Delta Delay in connecting for read, Delta Delay in connecting for write, forces match semantics for forces in VHPI, ... consistent path syntax 'instance_name, 'path_name, was documentation donated Not Done FT19 John Update based on process (ALL) and attribute changes 11-16-2004 Not Done FT24 John Restrict to std_logic_1164 and one-dimensional arrays of them John Record proof of targets being mutually exclusive wrt ?= Not Done CP002 David Make sure shift operators match resolutions at 7/18 meeting Explicitly define non-numeric shifts (sll, srl, rol, ror) for 1164 same as bit vectors including the value '0' Explicitly define numeric shifts (sla, sra) for slv in numeric_unsigned identical to unsigned Explicitly define all shifts for unsigned, signed in numeric_std sla sll sra srl -------------------------------------------------------------------------------------------- + sv | sll arg(arg'left-1:0)&0 arg(arg'left)&arg(arg'left:1) _0&arg(arg'left:1)_ - sv | sra (-cnt) srl(-cnt) sla(-cnt) sll(-cnt) + uv | sll arg(arg'left-1:0)&0 0&arg(arg'left:1) 0&arg(arg'left:1) - uv | sra (-cnt) srl(-cnt) sla(-cnt) sll(-cnt) For fixed point, these should be defined as above. Looked into packages. Will implement overload the ones for bit and create one for 1164 and don't define for floating point and fixed. Bit overloads will be in numeric_bit_unsigned (bit_vector type) and in numeric_bit for the arithmatic shifts, implicit rotate operators will be used. Nothing will be done for fix point and floating point types New Action xxx Jim Schedule meeting with VHPI to go through proposals and identify VHPI effects. Next Meeting Tenatively planed for March 14 and 28. Add number to action items to allow for easier track. Action items closed FT9 David For fixed point literals, get info from NASA on accuracy of 512 bit literals. Action items A1 FT20 John Check over to see what needs to be done. A2 FT25 Chuck/John Rewrite to handle () expressions as selector. Assure that "others" requirement doesn't change. Issue: length of expression does not need to be known statically A3 NONE Jim/Chuck Coordinate w/ Steve to hold vote on FT/ISAC issues. Make sure dependent things go out together (ie FT11 & FT20) A4 NONE Charlie Check to see that FT has reasonable descriptions. A5 FT31 Peter Add FILE_FLUSH function A6 FT16 Peter Make it an error to reference WORK within a context declaration. A7 FT16 Jim Mark done when Peter finishes action item. A8 FT?? ALL Code review of package updates (FT8, FT5, Floating Point) A9 ??? Jim, David usage model for floating point package A10 FT11 John Update to allow aggregates w/o delay and all others have delay A11 FT7 Peter, Jim Update package and define in terms of VHPI Delta Delay in connecting for read, Delta Delay in connecting for write, forces match semantics for forces in VHPI, ... consistent path syntax 'instance_name, 'path_name, was documentation donated A12 FT19 John Update based on process (ALL) and attribute changes 11-16-2004 A13 FT24 John Restrict to std_logic_1164 and one-dimensional arrays of them John Record proof of targets being mutually exclusive wrt ?= A14 CP002 David Make sure shift operators match resolutions at 7/18 meeting Explicitly define non-numeric shifts (sll, srl, rol, ror) for 1164 same as bit vectors including the value '0' Explicitly define numeric shifts (sla, sra) for slv in numeric_unsigned identical to unsigned Explicitly define all shifts for unsigned, signed in numeric_std sla sll sra srl -------------------------------------------------------------------------------------------- + sv | sll arg(arg'left-1:0)&0 arg(arg'left)&arg(arg'left:1) _0&arg(arg'left:1)_ - sv | sra (-cnt) srl(-cnt) sla(-cnt) sll(-cnt) + uv | sll arg(arg'left-1:0)&0 0&arg(arg'left:1) 0&arg(arg'left:1) - uv | sra (-cnt) srl(-cnt) sla(-cnt) sll(-cnt) For fixed point, these should be defined as above. A15 NONE Jim Provide description of using name subtype, qualified expressions, and literals to David. A16 FT9 David Write up problem of comparing integer with fixed point / real number which has only negative indexs. A17 NONE Jim Schedule meeting with VHPI to go through proposals and identify VHPI effects.