Apple-2gs-SW/ORCALib
1
0
Fork 0
mirror of https://github.com/byteworksinc/ORCALib.git synced 2025-02-07 04:30:31 +00:00
Code Issues Projects Releases Wiki Activity
ORCALib/fpextra.asm

117 lines
2.5 KiB
NASM
Raw Normal View History

Add new routines to reduce the precision of floating-point numbers. These will now be used to implement explicit floating-point casts.
2021-03-06 22:21:52 -06:00
keep obj/fpextra
mcopy fpextra.macros
****************************************************************
*
* FPextra - extra floating-point routines
*
* This code provides routines dealing with floating-point
* numbers that are used only by ORCA/C, supplementing the
* ones in SysFloat.
*
****************************************************************
*
fpextra private dummy segment
end
****************************************************************
*
* ~SinglePrecision - limit fp value to single precision & range
*
* Inputs:
* extended-format real on stack
*
****************************************************************
*
~SinglePrecision start
tsc
clc
Standardize indentation using spaces. Most files already used spaces, but three used tabs for indentation. These have been converted to use spaces. This allows the files to be displayed with proper formatting in modern editors and on GitHub. It also removes any dependency on SysTabs settings when assembling them. The spacing in fpextra.asm was also modified to use standard column positions. There are no non-whitespace changes in this commit.
2022-06-25 18:27:20 -05:00
adc #4
ldy #0
Add new routines to reduce the precision of floating-point numbers. These will now be used to implement explicit floating-point casts.
2021-03-06 22:21:52 -06:00
phy
pha
phy
pha
phy
pha
phy
pha
FX2S
FS2X
rtl
end
****************************************************************
*
* ~DoublePrecision - limit fp value to double precision & range
*
* Inputs:
* extended-format real on stack
*
****************************************************************
*
~DoublePrecision start
tsc
clc
Standardize indentation using spaces. Most files already used spaces, but three used tabs for indentation. These have been converted to use spaces. This allows the files to be displayed with proper formatting in modern editors and on GitHub. It also removes any dependency on SysTabs settings when assembling them. The spacing in fpextra.asm was also modified to use standard column positions. There are no non-whitespace changes in this commit.
2022-06-25 18:27:20 -05:00
adc #4
ldy #0
Add new routines to reduce the precision of floating-point numbers. These will now be used to implement explicit floating-point casts.
2021-03-06 22:21:52 -06:00
phy
pha
phy
pha
phy
pha
phy
pha
FX2D
FD2X
rtl
end
****************************************************************
*
* ~CompPrecision - limit fp value to comp precision & range
*
* Inputs:
* extended-format real on stack
*
Work around bug in FX2C and FX2L. These SANE operations can sometimes return incorrect values for certain negative integers such as -2147483648 and -53021371269120 (numbers with at least 16 low-order zero bits in their two's-complement representation). To work around this, we now avoid calling FX2C or FX2L on negative numbers, generally by saving and restoring the sign separately. These workarounds are used in several of the new <math.h> rounding functions, and also for code that converts floating-point values to comp or long long. There are some places in SysFloat that should be patched similarly, so we may still hit this problem in certain situations until that is done.
2021-11-27 22:16:54 -06:00
* Note: This avoids calling FX2C on negative numbers,
* because it is buggy for certain values.
*
Add new routines to reduce the precision of floating-point numbers. These will now be used to implement explicit floating-point casts.
2021-03-06 22:21:52 -06:00
****************************************************************
*
~CompPrecision start
Fix rounding issues introduced by SANE bug workarounds. The lrint functions could give the wrong result for negative numbers in upward/downward rounding modes. Casts to comp could also have different rounding behavior.
2021-11-30 20:19:57 -06:00
tsc round to integer
clc
Standardize indentation using spaces. Most files already used spaces, but three used tabs for indentation. These have been converted to use spaces. This allows the files to be displayed with proper formatting in modern editors and on GitHub. It also removes any dependency on SysTabs settings when assembling them. The spacing in fpextra.asm was also modified to use standard column positions. There are no non-whitespace changes in this commit.
2022-06-25 18:27:20 -05:00
adc #4
pea 0
Fix rounding issues introduced by SANE bug workarounds. The lrint functions could give the wrong result for negative numbers in upward/downward rounding modes. Casts to comp could also have different rounding behavior.
2021-11-30 20:19:57 -06:00
pha
FRINTX
Standardize indentation using spaces. Most files already used spaces, but three used tabs for indentation. These have been converted to use spaces. This allows the files to be displayed with proper formatting in modern editors and on GitHub. It also removes any dependency on SysTabs settings when assembling them. The spacing in fpextra.asm was also modified to use standard column positions. There are no non-whitespace changes in this commit.
2022-06-25 18:27:20 -05:00
lda 4+8,s
Work around bug in FX2C and FX2L. These SANE operations can sometimes return incorrect values for certain negative integers such as -2147483648 and -53021371269120 (numbers with at least 16 low-order zero bits in their two's-complement representation). To work around this, we now avoid calling FX2C or FX2L on negative numbers, generally by saving and restoring the sign separately. These workarounds are used in several of the new <math.h> rounding functions, and also for code that converts floating-point values to comp or long long. There are some places in SysFloat that should be patched similarly, so we may still hit this problem in certain situations until that is done.
2021-11-27 22:16:54 -06:00
pha save original sign
Standardize indentation using spaces. Most files already used spaces, but three used tabs for indentation. These have been converted to use spaces. This allows the files to be displayed with proper formatting in modern editors and on GitHub. It also removes any dependency on SysTabs settings when assembling them. The spacing in fpextra.asm was also modified to use standard column positions. There are no non-whitespace changes in this commit.
2022-06-25 18:27:20 -05:00
asl a force sign to positive
lsr a
sta 6+8,s
Work around bug in FX2C and FX2L. These SANE operations can sometimes return incorrect values for certain negative integers such as -2147483648 and -53021371269120 (numbers with at least 16 low-order zero bits in their two's-complement representation). To work around this, we now avoid calling FX2C or FX2L on negative numbers, generally by saving and restoring the sign separately. These workarounds are used in several of the new <math.h> rounding functions, and also for code that converts floating-point values to comp or long long. There are some places in SysFloat that should be patched similarly, so we may still hit this problem in certain situations until that is done.
2021-11-27 22:16:54 -06:00
tsc limit precision
Add new routines to reduce the precision of floating-point numbers. These will now be used to implement explicit floating-point casts.
2021-03-06 22:21:52 -06:00
clc
Standardize indentation using spaces. Most files already used spaces, but three used tabs for indentation. These have been converted to use spaces. This allows the files to be displayed with proper formatting in modern editors and on GitHub. It also removes any dependency on SysTabs settings when assembling them. The spacing in fpextra.asm was also modified to use standard column positions. There are no non-whitespace changes in this commit.
2022-06-25 18:27:20 -05:00
adc #6
ldy #0
Add new routines to reduce the precision of floating-point numbers. These will now be used to implement explicit floating-point casts.
2021-03-06 22:21:52 -06:00
phy
pha
phy
pha
phy
pha
phy
pha
FX2C
FC2X
Work around bug in FX2C and FX2L. These SANE operations can sometimes return incorrect values for certain negative integers such as -2147483648 and -53021371269120 (numbers with at least 16 low-order zero bits in their two's-complement representation). To work around this, we now avoid calling FX2C or FX2L on negative numbers, generally by saving and restoring the sign separately. These workarounds are used in several of the new <math.h> rounding functions, and also for code that converts floating-point values to comp or long long. There are some places in SysFloat that should be patched similarly, so we may still hit this problem in certain situations until that is done.
2021-11-27 22:16:54 -06:00
pla restore original sign
Standardize indentation using spaces. Most files already used spaces, but three used tabs for indentation. These have been converted to use spaces. This allows the files to be displayed with proper formatting in modern editors and on GitHub. It also removes any dependency on SysTabs settings when assembling them. The spacing in fpextra.asm was also modified to use standard column positions. There are no non-whitespace changes in this commit.
2022-06-25 18:27:20 -05:00
bpl ret
lda 4+8,s
ora #$8000
sta 4+8,s
Work around bug in FX2C and FX2L. These SANE operations can sometimes return incorrect values for certain negative integers such as -2147483648 and -53021371269120 (numbers with at least 16 low-order zero bits in their two's-complement representation). To work around this, we now avoid calling FX2C or FX2L on negative numbers, generally by saving and restoring the sign separately. These workarounds are used in several of the new <math.h> rounding functions, and also for code that converts floating-point values to comp or long long. There are some places in SysFloat that should be patched similarly, so we may still hit this problem in certain situations until that is done.
2021-11-27 22:16:54 -06:00
ret rtl
Add new routines to reduce the precision of floating-point numbers. These will now be used to implement explicit floating-point casts.
2021-03-06 22:21:52 -06:00
end
Reference in New Issue Copy Permalink