1. 25 Mar, 2009 2 commits
    • rtoy's avatar
      linux-os.lisp: · f0c421a2
      rtoy authored
      o Use unix:unix-uname to get the software version.  (I don't want to
        deal with broken Linux kernels or other implementation details in
        trying to read /proc/version, and I don't want to run uname -r.)
      
      sunos-os.lisp:
      o Use unix:unix-uname to get the software version instead of running
        uname -r.
      f0c421a2
    • rtoy's avatar
      Carl fixed the aliasing issue in e_rem_pio2.c sometime ago so we don't · 1ca43a1a
      rtoy authored
      need the -ffloat-store and -fno-strict-aliasing flags anymore.  The
      Linux x87 build correctly returns -0.9258790228548379d0 for (cos (expt
      2d0 120)).
      1ca43a1a
  2. 24 Mar, 2009 1 commit
  3. 20 Mar, 2009 1 commit
    • rtoy's avatar
      This big of code causes CMUCL to spend huge amounts of time trying to · cb674c89
      rtoy authored
      simplify the union of disjoint double-float types:
      
      (defun foo (arg)
        (declare (type double-double-float arg)
      	   (optimize (speed 3) (space 0)))
        (let* ((x arg)
      	 (xx 0w0)
      	 (k 0)
      	 (px (coerce k 'double-double-float))
      	 (qx 0w0))
          (declare (type double-double-float x xx px qx)
      	     (notinline poly-eval poly-eval-1))
          (setf x (- (- x (* px log2-c1)) (* px log2-c2)))
          (+ x 0d0)
          ))
      
      Fix it:
      
      code/type.lisp:
      o Rename SIMPLIFY-BIG-INTEGER-UNION to SIMPLIFY-BIG-UNION
      o SIMPLIFY-BIG-UNION takes an extra arg to specify the resulting type.
      o Modify SIMPLIFY-UNIONS to handle float unions as well as integer
        unions.
      
      general-info/release-20a.txt:
      o Update
      cb674c89
  4. 18 Mar, 2009 1 commit
    • rtoy's avatar
      code/irrat-dd.lisp: · 8bd5d96d
      rtoy authored
      o Declare POLY-EVAL and POLY-EVAL-1 as maybe-inline functions so they
        can be semi-inlined.
      
      code/irrat.lisp:
      o I don't think there's any real reason to make the local function
        REAL-EXPT inline.  It's very big and local function call is already
        good enough.
      
      compiler/float-tran.lisp:
      o Change some functions to be maybe-inline instead of inline to save
        on space for the the larger double-double functions:  SUB-DD,
        SUB-D-DD, SUB-DD-D, SPLIT.  Semi-inline should be good enough for
        these large functions, saving on cache memory, local function
        calling sequence should be good enough.
      8bd5d96d
  5. 16 Mar, 2009 1 commit
  6. 11 Mar, 2009 1 commit
  7. 03 Mar, 2009 1 commit
  8. 22 Feb, 2009 1 commit
  9. 09 Feb, 2009 3 commits
  10. 23 Jan, 2009 1 commit
  11. 22 Jan, 2009 1 commit
  12. 21 Jan, 2009 3 commits
  13. 20 Jan, 2009 4 commits
  14. 19 Jan, 2009 1 commit
  15. 13 Jan, 2009 1 commit
  16. 11 Jan, 2009 1 commit
    • rtoy's avatar
      Config.x86_common: · de39fd0a
      rtoy authored
      o Use += to append to variables instead of setting them.
      o Add common OS_SRC files.
      o Move rule for building e_rem_pio2.c to here.  Add CPPFLAGS too.
      
      Config.x86_common:
      o Remove common files from OS_SRC.
      o Remove rele for e_rem_pio2.c
      
      Config.x86_darwin:
      o Update to use Config.x86_common.
      
      GNUmakefile:
      o Use CPPFLAGS when building version.o
      de39fd0a
  17. 10 Jan, 2009 3 commits
  18. 09 Jan, 2009 6 commits
  19. 07 Jan, 2009 1 commit
  20. 06 Jan, 2009 6 commits