OpenCores
URL https://opencores.org/ocsvn/ft816float/ft816float/trunk

Subversion Repositories ft816float

Compare Revisions

  • This comparison shows the changes necessary to convert path 
    /
    from Rev 3 to Rev 4
    ↔ Reverse comparison
  •

Rev 3 → Rev 4

/ft816float/trunk/software/FAC1ToString.asm
0,0 → 1,572
; ============================================================================
; FAC1ToString.asm
; __
; \\__/ o\ (C) 2014 Robert Finch, Stratford
; \ __ / All rights reserved.
; \/_// robfinch<remove>@finitron.ca
; ||
;
;
; This source file is free software: you can redistribute it and/or modify
; it under the terms of the GNU Lesser General Public License as published
; by the Free Software Foundation, either version 3 of the License, or
; (at your option) any later version.
;
; This source file is distributed in the hope that it will be useful,
; but WITHOUT ANY WARRANTY; without even the implied warranty of
; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
; GNU General Public License for more details.
;
; You should have received a copy of the GNU General Public License
; along with this program. If not, see <http://www.gnu.org/licenses/>.
;
; ============================================================================
;
; This code is a heavily modified version of the floating point to string
; conversion routine which is a part of Lee Davison's EhBASIC.
;
Cvaral = $95 ; current var address low byte
Cvarah = Cvaral+1 ; current var address high byte
numexp = $A8 ; string to float number exponent count
expcnt = $AA ; string to float exponent count
Sendl = $BA ; BASIC pointer temp low byte
Sendh = $BB ; BASIC pointer temp low byte
 
Decss = $3A0 ; number to decimal string start
Decssp1 = Decss+1 ; number to decimal string start
FP_ADD EQU 1
FP_SUB EQU 2
FP_MUL EQU 3
FP_DIV EQU 4
FP_FIX2FLT EQU 5
FP_FLT2FIX EQU 6
FP_ABS EQU 7
FP_NEG EQU 16
FP_SWAP EQU 17
FIXED_MUL EQU $83
FIXED_ADD EQU $81
FIXED_SUB EQU $82
;parameter FIXED_DIV = 8'h84;
;parameter FIXED_ABS = 8'h87;
;parameter FIXED_NEG = 8'h90;
FP_CMDREG EQU $FEA20E
FP_STATREG EQU $FEA20E
FAC1 EQU $FEA200
FAC1_5 EQU $FEA200
FAC1_4 EQU $FEA202
FAC1_3 EQU $FEA204
FAC1_2 EQU $FEA206
FAC1_1 EQU $FEA208
FAC1_msw EQU $FEA208
FAC1_e EQU $FEA20A
FAC2 EQU $FEA210
 
CPU W65C816S
NDX 16
MEM 16
 
public FAC1ToString:
 
; The first chunk of code determines if the number is positive or negative
; and spits out the appropriate sign. Next it takes the absolute value of
; the accumulator so following code only has to deal with positive numbers.
 
LDY #$00 ; set index = 1
LDA FAC1_msw ; test FAC1 sign (b15) (Can't use BIT)
BPL .0002 ; branch if +ve
LDA #'-' ; else character = "-"
STA Decss,Y ; save leading character (" " or "-")
LDA #FP_NEG ; make the FAC positive
JSR FPCommandWait
BRA .0001
.0002:
LDA #$20 ; character = " " (assume +ve)
STA Decss,Y
.0001:
STY Sendl ; save index
 
; This little bit of code check for a zero exponent which indicates a
; value of zero.
 
LDA FAC1_e ; get FAC1 exponent
TAX
BNE LAB_2989 ; branch if FAC1<>0
; exponent was $00 so FAC1 is 0
LDA #'0' ; set character = "0"
BRL LAB_2A89 ; save last character, [EOT] and exit
 
; This loop attempts to make small values more significant, so that there are
; fewer leading zeros in the value. (The exponent is decremented so that it
; corresponds). Because of the potential for extremely small values looping is
; limited. The problem is the 16 bit exponent can allow for much smaller
; values than an 8 bit exponent would and we don't want to loop for thousands
; of iterations in order to display a value that's almost zero.
 
; FAC1 is some non zero value
LAB_2989
STY Sendl ; save off .Y
LDY #1639 ; max number of retries
LDA #$00 ; clear (number exponent count)
STA numexp
LOOP_MBMILLION:
CPX #$8000 ; compare FAC1 exponent with $8000 (>1.00000)
BCS LAB_299A ; branch if FAC1=>1
; FAC1<1
PEA A_MILLION ; multiply FAC * 1,000,000
JSR LOAD_FAC2 ;
PLA ; get rid of parameter
JSR FMUL
LDA numexp
SEC
SBC #6 ; set number exponent count (-6)
STA numexp
LDA FAC1_e
TAX
DEY
BPL LOOP_MBMILLION
 
LAB_299A
LDY Sendl ; get back .Y
 
; These two loops coerce the value of the FAC to be between 100,000 and
; 1,000,000. This gives a maximum of six digits before the decimal point
; in scientific notation.
 
; This loop divides by 10 until the value in the FAC is less than 1,000,000
;
LOOP_DB10:
PEA MAX_BEFORE_SCI ; set pointer low byte to 999999.4375 (max before sci note)
JSR LOAD_FAC2 ; compare FAC1 with (AY)
PLA ; get rid of parameter
LDA FP_CMDREG
BIT #$08 ; test equals bit
BNE LAB_29C3 ; exit if FAC1 = (AY)
BIT #$04 ; test greater than bit
BEQ LOOP_MB10 ; go do *10 if FAC1 < (AY)
 
LAB_29B9
JSR DivideByTen ; divide by 10
INC numexp ; increment number exponent count
BRA LOOP_DB10 ; go test again (branch always)
 
; This loop multiplies the value by 10 until it's greater than
; 100,000.
; FAC1 < (AY)
LOOP_MB10
PEA CONST_9375 ; set pointer to 99999.9375
JSR LOAD_FAC2 ; compare FAC1 with (AY)
PLA ; get rid of parameter
LDA FP_CMDREG
BIT #$08
BNE LAB_29B2 ; branch if FAC1 = (AY) (allow decimal places)
BIT #$04
BNE LAB_29C0 ; branch if FAC1 > (AY) (no decimal places)
; FAC1 <= (AY)
LAB_29B2
JSR MultiplyByTen ; multiply by 10
DEC numexp ; decrement number exponent count
BRA LOOP_MB10 ; go test again (branch always)
 
; now we have just the digits to do
 
LAB_29C0
; JSR AddPoint5 ; add 0.5 to FAC1 (round FAC1)
LAB_29C3
; JSR FloatToFixed ; convert FAC1 floating-to-fixed
LDX #$01 ; set default digits before dp = 1
LDA numexp ; get number exponent count
CLC ; clear carry for add
ADC #$07 ; up to 6 digits before point
BMI LAB_29D8 ; if -ve then 1 digit before dp
 
CMP #$08 ; A>=8 if n>=1E6
BCS LAB_29D9 ; branch if >= $08
 
; carry is clear
TAX ; copy to A
DEX ; take 1 from digit count
LDA #$02 ;.set exponent adjust
 
LAB_29D8
SEC ; set carry for subtract
LAB_29D9
SBC #$02 ; -2
STA expcnt ;.save exponent adjust
STX numexp ; save digits before dp count
TXA ; copy to A
BEQ LAB_29E4 ; branch if no digits before dp
 
BPL LAB_29F7 ; branch if digits before dp
 
LAB_29E4
LDY Sendl ; get output string index
LDA #'.' ; character "."
INY ; increment index
STA Decss,Y ; save to output string
TXA ;.
BEQ LAB_29F5 ;.
 
LDA #'0' ; character "0"
INY ; increment index
STA Decss,Y ; save to output string
LAB_29F5
STY Sendl ; save output string index
 
LAB_29F7
LDX #'0' ; holds onto the digit value
 
; Now loop subtracting 100,000 as many times as we can. The value was coerced
; to be between 100,000 and 1,000,000. Count the number of times subtraction
; can be done successfully.
;
LAB_29FB
PEA CONST_100000
JSR LOAD_FAC2 ; load FAC2 with 100,000
PLA ; get rid of parameter
LDA FP_STATREG
BIT #$04 ; Is FAC1 > 100,000 ?
BEQ .0005 ; branch if not
LDA #FP_SWAP ; subtract is FAC2-FAC1!
JSR FPCommandWait;
LDA #FP_SUB ; subtract 100,000 from the mantissa.
JSR FPCommandWait
INX ; increment the value of the digit
BRA LAB_29FB ; try again
.0005:
TXA
LDY Sendl ; get output string index
INY ; increment output string index
TXA
STA Decss,Y ; save to output string
DEC numexp ; decrement # of characters before the dp
BNE LAB_2A3B ; branch if still characters to do
; else output the point
LDA #'.' ; character "."
INY ; increment output string index
STA Decss,Y ; save to output string
LAB_2A3B
STY Sendl ; save output string index
; We subtracted until the value was < 100,000 so multiply the
; remainder upwards to get the next digit.
JSR MultiplyByTen ; If not, multiply by 10
CPY #27 ; converted (+/- . incl)
BCC LAB_29F7
; now remove trailing zeroes
.RemoveTrailingZeros
LDA Decss,Y ; get character from output string
AND #$FF ; mask to a byte
DEY ; decrement output string index
CMP #'0' ; compare with "0"
BEQ .RemoveTrailingZeros ; loop until non "0" character found
 
CMP #'.' ; compare with "."
BEQ LAB_2A58 ; branch if was dp
 
; restore last character
INY ; increment output string index
LAB_2A58
LDA #'+' ; character "+"
LDX expcnt ; get exponent count
LBEQ LAB_2A8C ; if zero go set null terminator and exit
 
; exponent isn't zero so write exponent
BPL LAB_2A68 ; branch if exponent count +ve
 
LDA #$00 ; clear A
SEC ; set carry for subtract
SBC expcnt ; subtract exponent count adjust (convert -ve to +ve)
TAX ; copy exponent count to X
LDA #'-' ; character "-"
 
; We must keep moving forwards through the string because the acc is storing
; two bytes.
 
LAB_2A68
PHA
LDA #'E' ; character "E"
STA Decss+1,Y ; save exponent sign to output string
PLA
STA Decss+2,Y ; save to output string
TXA ; get exponent count back
 
; do highest exponent digit
STZ Sendl
LDX #'0'-1 ; one less than "0" character
SEC ; set carry for subtract
.0001:
INX ; count how many times we can subtract 10,000
SBC #10000
BCS .0001
ADC #10000
CPX #'0'
BEQ .0005
INC Sendl
PHA
TXA
STA Decss+3,Y
PLA
INY
; do the next exponent digit
.0005:
LDX #'0'-1
SEC
.0002:
INX
SBC #1000
BCS .0002
ADC #1000
LSR Sendl
BCS .00010
CPX #'0'
BEQ .0006
.00010:
INC Sendl
PHA
TXA
STA Decss+3,Y
PLA
INY
; and the next
.0006:
LDX #'0'-1
SEC
.0003:
INX
SBC #100
BCS .0003
ADC #100
LSR Sendl
BCS .00011
CPX #'0'
BEQ .0007
.00011:
INC Sendl
PHA
TXA
STA Decss+3,Y
PLA
INY
 
.0007:
LDX #'0'-1
SEC
.0004:
INX
SBC #10
BCS .0004
ADC #10
LSR Sendl
BCS .00012
CPX #'0'
BEQ .0008
.00012:
INC Sendl
PHA
TXA
STA Decss+3,Y
PLA
INY
 
.0008:
ADC #'0'
STA Decss+3,Y
LDA #$00 ; set null terminator
STA Decss+4,Y ; save to output string
RTS ; go set string pointer (AY) and exit (branch always)
 
LAB_2A89
STA Decss,Y ; save last character to output string
; set null terminator and exit
LAB_2A8C
LDA #$00 ; set null terminator
STA Decss+1,Y ; save after last character
 
LAB_2A91
; LDA #<Decssp1 ; set result string low pointer
; LDY #>Decssp1 ; set result string high pointer
RTS
 
LAB_25FB:
LDA #FP_SWAP
JSR FPCommandWait
LDY #0
TYX
.0002:
LDA (3,S),Y
STA FAC1,X
INY
INY
INX
INX
CPX #12
BNE .0002
LDA #FP_FIX2FLT
JSR FPCommandWait
FMUL:
LDA #FP_MUL
JMP FPCommandWait
LOAD_FAC2:
PHX
PHY
LDY #0
TYX
.0002:
LDA (7,s),Y
STA FAC2,X
INY
INY
INX
INX
CPX #12
BNE .0002
PLY
PLX
RTS
FloatToFixed:
LDA #FP_FLT2FIX
JMP FPCommandWait
AddPoint5:
PEA CONST_POINT5
JSR LOAD_FAC2
PLA
LDA #FP_ADD
JMP FPCommandWait
MultiplyByTen:
PEA TEN_AS_FLOAT
JSR LOAD_FAC2
PLA
LDA #FP_MUL
JMP FPCommandWait
public DivideByTen:
PEA TEN_AS_FLOAT
JSR LOAD_FAC2
PLA
JSR SwapFACs
LDA #FP_DIV
JMP FPCommandWait
SwapFACs:
LDA #FP_SWAP
 
; Issue a command to the FP unit and wait for it to complete
;
public FPCommandWait:
PHA
.0001:
LDA FP_STATREG ; get the status register
BIT #$80 ; check for busy bit
BNE .0001 ; if busy go back
PLA ; to pop acc
STA FP_CMDREG ; store the command
RTS
 
; Display the FAC1 as a hex number
;
public DispFAC1:
LDA FAC1_e
JSR DispWord
LDA FAC1_1
JSR DispWord
LDA FAC1_2
JSR DispWord
LDA FAC1_3
JSR DispWord
LDA FAC1_4
JSR DispWord
LDA FAC1_5
JSR DispWord
LDA #' '
JSR OutChar
RTS
;
; 1,000,000 as a floating point number
;
A_MILLION: ; $F4240
dw $0000
dw $0000
dw $0000
dW $0000
dw $7A12
dw $8013
 
CONST_100000:
;186A0
dw $0000
dw $0000
dw $0000
dw $0000
dw $61A8
dw $8010
; The constant 999999.4375 as hex
; 01.11_1010_0001_0001_1111_1011_1000_00000000000000000000000000
MAX_BEFORE_SCI:
dw $0000
dw $0000
dw $0000
dw $FB80
dw $7A11
dw $8013
 
TEN_AS_FLOAT:
dw $0000
dw $0000
dw $0000
dw $0000
dw $5000
dw $8003
 
; 99999.9375
; 01.10_0001_1010_0111_1111_1100_000000000000000000000000000000
;
CONST_9375:
dw $0000
dw $0000
dw $0000
dw $FC00
dw $61A7
dw $8010
 
; 0.5
CONST_POINT5:
dw $0000
dw $0000
dw $0000
dw $0000
dw $4000
dw $7FFF
 
; This table is used in converting numbers to ASCII.
 
LAB_2A9A
LAB_2A9B = LAB_2A9A+1
LAB_2A9C = LAB_2A9B+1
; .word $FFFF,$F21F,$494C,$589C,$0000
; .word $0000,$0163,$4578,$5D8A,$0000
; .word $FFFF,$FFDC,$790D,$903F,$0000
; .word $0000,$0003,$8D7E,$A4C6,$8000
; .word $FFFF,$FFFF,$A50C,$EF85,$C000
; .word $0000,$0000,$0918,$4E72,$A000
; .word $FFFF,$FFFF,$FF17,$2B5A,$F000
; .word $0000,$0000,$0017,$4876,$E800
; .word $FFFF,$FFFF,$FFFD,$ABF4,$1C00
; .word $0000,$0000,$0000,$3B9A,$CA00
; .word $FFFF,$FFFF,$FFFF,$FF67,$6980
; .word $0000,$0000,$0000,$05F5,$E100 ; 100000000
; .word $0000,$0000,$0098,$9680 ; 10000000
; .word $4240,$000F,$0000,$0000,$0000,$804E ; 1000000
.word $86A0,$0001,$0000,$0000,$0000,$804E ; 100000
.word $2710,$0000,$0000,$0000,$0000,$804E ; 10000
.word $03E8,$0000,$0000,$0000,$0000,$804E ; 1000
.word $0064,$0000,$0000,$0000,$0000,$804E ; 100
FIXED10:
.word $000A,$0000,$0000,$0000,$0000,$804E ; 10
.word $0001,$0000,$0000,$0000,$0000,$804E ; 1
 
MEM 16
NDX 16

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.