- Nov 05, 2001
-
-
toy authored
31 bits. Modified by RLT to get rid of an extra branch instruction. o Update right shifts for (signed-byte 32) to get rid of an extra branch instruction too.
-
- Oct 12, 2001
-
-
toy authored
cross-compilation not to loop forever. From Douglas.
-
- May 18, 2001
-
-
toy authored
-
- Jan 19, 2001
-
-
dtc authored
o Some cases of where the shift was a known constant were missed.
-
- Jan 03, 2001
- Sep 26, 2000
-
-
dtc authored
port. o Add some new VOPS to handle special cases for ash where we know the sign of the shift and the result will fit into a good type. o Add deftransform for ash to call the these vops
-
- Feb 16, 2000
-
-
dtc authored
assembly/sparc/arith.lisp won't be used anymore on V8 or V9. o fast-ash now uses a conditional move instruction on the V9. Should help since it gets rid of a branch instruction, but I can't really tell. o The fast-* for signed and unsigned multiplication appropriately conditionalized to work on V8 and V9, as appropriate. (As discussed in my earlier patch for assembly/sparc/arith.lisp). o bignum::%digit-0-or-plusp uses a conditional move on the V9 o Some miscellaneous patches to the VOP guards so that the ones that require sparc-64 support are only done on sparc-64 platforms (of which there aren't any supported yet). o Added a bignum-floor vop for V8.
-
- Jan 17, 2000
-
-
dtc authored
-
- Jan 16, 2000
-
-
dtc authored
unsigned-nums. Including a union of both these primitive type restrictions results in a template type that can not be achieved by either of them. Problem noted by Raymond Toy.
-
- Dec 22, 1999
-
-
dtc authored
the :sparc-64 feature as the 64-bit state is not currently saved which would causes lossage; from Raymond Toy.
-
- Dec 08, 1999
-
-
dtc authored
select a VOP version they must have different names.
-
- Nov 19, 1999
-
-
dtc authored
Sparc V9; replace them with mulx for the Sparc V9. Now use only one of :sparc-v7, :sparc-v8, or :sparc-v9 when building to get the desired features. From Raymond Toy.
-
- Nov 11, 1999
-
-
dtc authored
port. Do not use the popc instruction which isn't implemented in hardware on an Ultrasparc. From Raymond Toy.
-
- Jun 22, 1999
-
-
dtc authored
o Enable the v9 version of emit-multiply. This is slightly slower than the v8 version, but doesn't use the deprecated instructions and registers. This approximately doubles the speed of bignum multiplies. o Add a v9 version of bignum::%floor. This approximately doubles the speed of bignum floors. No v8 version because the divide instruction might overflow.
-
- Jun 19, 1999
-
-
dtc authored
the V8 and V9; from Raymond Toy.
-
- Mar 21, 1998
-
-
dtc authored
-
- Oct 31, 1994
-
-
ram authored
-
- Feb 11, 1994
-
-
cvs2git authored
-
- Jan 13, 1993
-
-
cvs2git authored
-
- Aug 02, 1992
-
-
wlott authored
type to be tagged-num instead of fixnum.
-
- Jan 31, 1992
-
-
ram authored
-
- Jan 10, 1992
-
-
ram authored
must be FIXNUM (when with :SAFE VOPs need not be proven type, only asserted.) Added GENERIC-EQL/FIXNUM VOPs (with higher costs) to prevent spurious representation number-consing when a fixnum and a word-integer are compared (e.g. in ZEROP.)
-
- Apr 09, 1991
-
-
wlott authored
set, and deftransform the rest. Added VOPs for shift-towards-{start,end}.
-
- Dec 16, 1990
-
-
wlott authored
-
- Dec 15, 1990
-
-
wlott authored
-
- Dec 11, 1990
-
-
wlott authored
cost.
-
- Dec 03, 1990
-
-
wlott authored
instead of the raw value.
-
- Nov 30, 1990
-
-
wlott authored
-
- Nov 16, 1990
-
-
wlott authored
-
- Nov 03, 1990
-
-
wlott authored
-
- Nov 01, 1990
-
-
wlott authored
-
- Oct 02, 1990
-
-
wlott authored
for 4 args.
-
- Jul 26, 1990
-
-
wlott authored
-
- Jul 21, 1990
-
-
ram authored
version gets used when it applies.
-
- Jul 12, 1990
- Jul 10, 1990
-
-
ram authored
lifetime problems.
-
- Jul 03, 1990