- May 01, 2007
-
-
rtoy authored
start have changed in compiler/x86/parms.lisp.
-
- Apr 30, 2007
-
-
rtoy authored
-
- Apr 20, 2007
-
-
rtoy authored
COMPLEX-DOUBLE-FLOAT-VALUE vops. This reduces the number of moves needed.
-
- Apr 19, 2007
- Apr 14, 2007
-
-
rtoy authored
to be ok now. o Allow descriptor arg for COMPLEX-DOUBLE-DOUBLE-FLOAT-VALUE. This can significanlty reduce the number of loads from memory.
-
- Apr 12, 2007
-
-
rtoy authored
instruction instead of the STDF instruction. (How could this have worked?)
-
rtoy authored
rid of quite a few uneeded loads from memory because we'd load the whole complex double-double into registers and select just the real or imaginary part. Now we just load the two double-floats we want.
-
rtoy authored
rid of quite a few uneeded loads from memory because we'd the whole complex double-double into registers and select just the real or imaginary part. Now we just load the 2 double-floats we want.
-
rtoy authored
o Was not use the FP arg for the stack when storing the arg to the stack. o Was storing the wrong element to the stack (real-hi instead of real-lo).
-
- Apr 07, 2007
- Mar 31, 2007
-
-
rtoy authored
-
rtoy authored
o Turn off :load-if for the real arg because the generator doesn't handle the case if the real arg is on the stack. o Enable :load-if for the result because the code (already) handled it. o The wrong TN was being used for the imaginary part of the real part. o The wrong TNs were being used for the arg registers. Use the right TN values.
-
- Mar 30, 2007
-
-
rtoy authored
-
- Mar 28, 2007
-
-
rtoy authored
-
rtoy authored
includes a character. Update accordingly. Noted by Madhu on cmucl-imp 2007/03/27, forwarded from a mail on sbcl-devel, from Eric Marsden.
-
rtoy authored
o Fix extra paren for double-double-stack-sc-number case, and don't use with-nfp which doesn't exist on x86. o Add support for complex-double-double-stack-sc-number, which was missing for x86.
-
rtoy authored
and the result are both on the stack. Fix this by turning off the :LOAD-IF stuff so that we really have registers.
-
rtoy authored
double-double-stack-sc-number and complex-double-double-stack-sc-number. o The complex-double-double-stack-sc-number case was not computed the complex double-double correctly. Fix it.
-
fgilham authored
-
- Mar 27, 2007
-
-
rtoy authored
-
rtoy authored
correctly if the two parts had different signs, which does happen.
-
rtoy authored
MAKE/DOUBLE-DOUBLE-FLOAT.
-
rtoy authored
MAKE-COMPLEX-DOUBLE-DOUBLE-FLOAT and fix up the vop to use the correct TN numbers when the result is on the stack.
-
rtoy authored
was using the wrong TN numbers for the components of the complex double-double.
-
rtoy authored
complex-double-double-reg-sc-number should be 2, not 1. o Oops. Forgot to implement complex-double-double-stack-sc-number.
-
- Mar 26, 2007
-
-
rtoy authored
MAKE/DOUBLE-DOUBLE-FLOAT, so turn it off. Turn it off for MAKE-COMPLEX-DOUBLE-DOUBLE-FLOAT and COMPLEX-DOUBLE-DOUBLE-FLOAT-VALUE too, just in case.
-
- Mar 23, 2007
-
-
rtoy authored
silly bug that (float <neg bignum> 1w0) was returning a positive result instead of negative. o Rename DOUBLE-DOUBLE-FLOAT to DOUBLE-DOUBLE-FLOAT-FROM-BITS to match other functions and to make it separate from the same function in the KERNEL package. o Modify DOUBLE-DOUBLE-FLOAT-FROM-BITS to look at the bignum and extract out the pieces better. Look for a pattern <53 bits> <zeroes> <53 bits> where <zeroes> is a set of consecutive zero bits. We use the 2 53-bit sections to create the double-double-float. o Make DOUBLE-DOUBLE-FLOAT-FROM-BITS honor the sign. o Add BIGNUM-FLOAT-DIGITS to figure out how many bits of the bignum should be used to create the float. o Call BIGNUM-FLOAT-DIGITS from BIGNUM-TO-FLOAT instead of using FLOAT-FORMAT-DIGITS.
-
- Mar 22, 2007
-
-
rtoy authored
-
- Mar 21, 2007
- Mar 20, 2007
-
-
rtoy authored
exception. This was happening when the faulting FP instruction was in the delay slot of a branch. The PC was incremented to the branch target, and hence had the wrong information. The FPQ structure has the right information. code/sparc-svr4-vm.lisp: o Add FPQ structure so we can access the information about the faulting FP instruction. compiler/sparc/float.lisp: o Extract the correct information about the faulting FP instruction instead of from the PC.
-
rtoy authored
-
- Mar 05, 2007
-
-
rtoy authored
Simple tests with rest args and some dynamic-extent closures indicates that this work. Compiling CMUCL with dynamic-extent enabled appears to work as well. I conclude that the implementation here is probably correct. (Yes, I know dynamic-extent is currently disabled, but let's make ppc complete.) compiler/ppc/alloc.lisp: o Implement the vops %DYNAMIC-EXTENT-START and %DYNAMIC-EXTENT-END. o Update the LIST-OR-LIST*, MAKE-CLOSURE, and FIXED-ALLOC vops to support dynamic-extent, by passing ALLOCATION and WITH-FIXED-ALLOCATION macros the extra :STACK-P arg. compiler/ppc/call.lisp: o Update LISTIFY-REST-ARGS vop to support dynamic-extent, by passing the ALLOCATION macro the extra :STACK-P arg. compiler/ppc/macros.lisp: o Update the ALLOCATION macro to support the :STACK-P arg.
-
- Mar 03, 2007
-
-
rtoy authored
o Implement stack clearing stuff for ppc. compiler/ppc/call.lisp: o Implement stack clearing for xep-allocate-frame and allocate-frame, mostly cargo-culting the sparc version.
-
- Feb 21, 2007
-
-
rtoy authored
-
- Feb 03, 2007
-
-
rtoy authored
(defun f (x y) (declare (fixnum x y)) (expt x y)) caused an error because we weren't handling member types correctly.
-
- Feb 01, 2007
-
-
rtoy authored
-