math/mpx-mul4-x86-sse2.S: Use the correct vector-multiply instruction.

Not sure why GNU as let me get away with that.

diff --git a/math/mpx-mul4-x86-sse2.S b/math/mpx-mul4-x86-sse2.S
index 5d8c5b97d2e72c4a1f1adde83432f232dbd4a401..0b57dbfd1e01a0fa6b0a3789cdddc71a9c62d230 100644 (file)
--- a/math/mpx-mul4-x86-sse2.S
+++ b/math/mpx-mul4-x86-sse2.S
@@ -64,7 +64,7 @@
 ///       0    v'_0    v'_1    v''_0   v''_1
 ///      16    v'_2    v'_3    v''_2   v''_3
 ///
-/// A `pmuludqd' instruction ignores the odd positions in its operands; thus,
+/// A `pmuludq' instruction ignores the odd positions in its operands; thus,
 /// it will act on (say) v'_0 and v''_0 in a single instruction.  Shifting
 /// this vector right by 4 bytes brings v'_1 and v''_1 into position.  We can
 /// multiply such a vector by a full 32-bit scalar to produce two 48-bit
@@ -81,7 +81,7 @@
 /// the register c0, for example, holds c'_0 (low half) and c''_0 (high
 /// half), and represents the value c_0 = c'_0 + c''_0 b; the carry registers
 /// collectively represent the value c_0 + c_1 B + c_2 B^2 + c_3 B^3.  The
-/// `pmuluqdq' instruction acting on a scalar operand (broadcast across all
+/// `pmuluqd' instruction acting on a scalar operand (broadcast across all
 /// lanes of its vector) and an operand in the expanded form above produces a
 /// result which can be added directly to the appropriate carry register.
 /// Following a pass of four multiplications, we perform some limited carry
@@ -118,19 +118,19 @@
        psrldq  \d3, 4                  // (s'_3, s''_2, s''_3, 0)
   .endif
   .ifnes "\d1", "nil"
-       pmuludqd \d1, \d0               // (r_i s'_1, r_i s''_1)
+       pmuludq \d1, \d0                // (r_i s'_1, r_i s''_1)
   .endif
   .ifnes "\d3", "nil"
-       pmuludqd \d3, \d0               // (r_i s'_3, r_i s''_3)
+       pmuludq \d3, \d0                // (r_i s'_3, r_i s''_3)
   .endif
   .ifnes "\d2", "nil"
     .ifnes "\d3", "nil"
-       pmuludqd \d2, \d0               // (r_i s'_2, r_i s''_2)
+       pmuludq \d2, \d0                // (r_i s'_2, r_i s''_2)
     .else
-       pmuludqd \d2, [\s + 16]
+       pmuludq \d2, [\s + 16]
     .endif
   .endif
-       pmuludqd \d0, [\s]              // (r_i s'_0, r_i s''_0)
+       pmuludq \d0, [\s]               // (r_i s'_0, r_i s''_0)
 .endm
 
 .macro accum   c0, c1, c2, c3