llvmpipe: export the tgsi translation code to a common layer

the llvmpipe tgsi translation is a lot more complete than what was in
gallivm so replacing the latter with the former. this is needed since
the draw llvm paths will use the same code. effectively the proven
llvmpipe code becomes gallivm.

1 files changed, 1325 insertions, 0 deletions

diff --git a/src/gallium/auxiliary/gallivm/lp_bld_arit.c b/src/gallium/auxiliary/gallivm/lp_bld_arit.c
new file mode 100644
index 00000000000..54b31befe6d
--- /dev/null
+++ b/src/gallium/auxiliary/gallivm/lp_bld_arit.c
@@ -0,0 +1,1325 @@
+/**************************************************************************
+ *
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+
+/**
+ * @file
+ * Helper
+ *
+ * LLVM IR doesn't support all basic arithmetic operations we care about (most
+ * notably min/max and saturated operations), and it is often necessary to
+ * resort machine-specific intrinsics directly. The functions here hide all
+ * these implementation details from the other modules.
+ *
+ * We also do simple expressions simplification here. Reasons are:
+ * - it is very easy given we have all necessary information readily available
+ * - LLVM optimization passes fail to simplify several vector expressions
+ * - We often know value constraints which the optimization passes have no way
+ *   of knowing, such as when source arguments are known to be in [0, 1] range.
+ *
+ * @author Jose Fonseca <[email protected]>
+ */
+
+
+#include "util/u_memory.h"
+#include "util/u_debug.h"
+#include "util/u_math.h"
+#include "util/u_string.h"
+#include "util/u_cpu_detect.h"
+
+#include "lp_bld_type.h"
+#include "lp_bld_const.h"
+#include "lp_bld_intr.h"
+#include "lp_bld_logic.h"
+#include "lp_bld_pack.h"
+#include "lp_bld_debug.h"
+#include "lp_bld_arit.h"
+
+
+/**
+ * Generate min(a, b)
+ * No checks for special case values of a or b = 1 or 0 are done.
+ */
+static LLVMValueRef
+lp_build_min_simple(struct lp_build_context *bld,
+                    LLVMValueRef a,
+                    LLVMValueRef b)
+{
+   const struct lp_type type = bld->type;
+   const char *intrinsic = NULL;
+   LLVMValueRef cond;
+
+   /* TODO: optimize the constant case */
+
+   if(type.width * type.length == 128) {
+      if(type.floating) {
+         if(type.width == 32 && util_cpu_caps.has_sse)
+            intrinsic = "llvm.x86.sse.min.ps";
+         if(type.width == 64 && util_cpu_caps.has_sse2)
+            intrinsic = "llvm.x86.sse2.min.pd";
+      }
+      else {
+         if(type.width == 8 && !type.sign && util_cpu_caps.has_sse2)
+            intrinsic = "llvm.x86.sse2.pminu.b";
+         if(type.width == 8 && type.sign && util_cpu_caps.has_sse4_1)
+            intrinsic = "llvm.x86.sse41.pminsb";
+         if(type.width == 16 && !type.sign && util_cpu_caps.has_sse4_1)
+            intrinsic = "llvm.x86.sse41.pminuw";
+         if(type.width == 16 && type.sign && util_cpu_caps.has_sse2)
+            intrinsic = "llvm.x86.sse2.pmins.w";
+         if(type.width == 32 && !type.sign && util_cpu_caps.has_sse4_1)
+            intrinsic = "llvm.x86.sse41.pminud";
+         if(type.width == 32 && type.sign && util_cpu_caps.has_sse4_1)
+            intrinsic = "llvm.x86.sse41.pminsd";
+      }
+   }
+
+   if(intrinsic)
+      return lp_build_intrinsic_binary(bld->builder, intrinsic, lp_build_vec_type(bld->type), a, b);
+
+   cond = lp_build_cmp(bld, PIPE_FUNC_LESS, a, b);
+   return lp_build_select(bld, cond, a, b);
+}
+
+
+/**
+ * Generate max(a, b)
+ * No checks for special case values of a or b = 1 or 0 are done.
+ */
+static LLVMValueRef
+lp_build_max_simple(struct lp_build_context *bld,
+                    LLVMValueRef a,
+                    LLVMValueRef b)
+{
+   const struct lp_type type = bld->type;
+   const char *intrinsic = NULL;
+   LLVMValueRef cond;
+
+   /* TODO: optimize the constant case */
+
+   if(type.width * type.length == 128) {
+      if(type.floating) {
+         if(type.width == 32 && util_cpu_caps.has_sse)
+            intrinsic = "llvm.x86.sse.max.ps";
+         if(type.width == 64 && util_cpu_caps.has_sse2)
+            intrinsic = "llvm.x86.sse2.max.pd";
+      }
+      else {
+         if(type.width == 8 && !type.sign && util_cpu_caps.has_sse2)
+            intrinsic = "llvm.x86.sse2.pmaxu.b";
+         if(type.width == 8 && type.sign && util_cpu_caps.has_sse4_1)
+            intrinsic = "llvm.x86.sse41.pmaxsb";
+         if(type.width == 16 && !type.sign && util_cpu_caps.has_sse4_1)
+            intrinsic = "llvm.x86.sse41.pmaxuw";
+         if(type.width == 16 && type.sign && util_cpu_caps.has_sse2)
+            intrinsic = "llvm.x86.sse2.pmaxs.w";
+         if(type.width == 32 && !type.sign && util_cpu_caps.has_sse4_1)
+            intrinsic = "llvm.x86.sse41.pmaxud";
+         if(type.width == 32 && type.sign && util_cpu_caps.has_sse4_1)
+            intrinsic = "llvm.x86.sse41.pmaxsd";
+      }
+   }
+
+   if(intrinsic)
+      return lp_build_intrinsic_binary(bld->builder, intrinsic, lp_build_vec_type(bld->type), a, b);
+
+   cond = lp_build_cmp(bld, PIPE_FUNC_GREATER, a, b);
+   return lp_build_select(bld, cond, a, b);
+}
+
+
+/**
+ * Generate 1 - a, or ~a depending on bld->type.
+ */
+LLVMValueRef
+lp_build_comp(struct lp_build_context *bld,
+              LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+
+   if(a == bld->one)
+      return bld->zero;
+   if(a == bld->zero)
+      return bld->one;
+
+   if(type.norm && !type.floating && !type.fixed && !type.sign) {
+      if(LLVMIsConstant(a))
+         return LLVMConstNot(a);
+      else
+         return LLVMBuildNot(bld->builder, a, "");
+   }
+
+   if(LLVMIsConstant(a))
+      return LLVMConstSub(bld->one, a);
+   else
+      return LLVMBuildSub(bld->builder, bld->one, a, "");
+}
+
+
+/**
+ * Generate a + b
+ */
+LLVMValueRef
+lp_build_add(struct lp_build_context *bld,
+             LLVMValueRef a,
+             LLVMValueRef b)
+{
+   const struct lp_type type = bld->type;
+   LLVMValueRef res;
+
+   if(a == bld->zero)
+      return b;
+   if(b == bld->zero)
+      return a;
+   if(a == bld->undef || b == bld->undef)
+      return bld->undef;
+
+   if(bld->type.norm) {
+      const char *intrinsic = NULL;
+
+      if(a == bld->one || b == bld->one)
+        return bld->one;
+
+      if(util_cpu_caps.has_sse2 &&
+         type.width * type.length == 128 &&
+         !type.floating && !type.fixed) {
+         if(type.width == 8)
+            intrinsic = type.sign ? "llvm.x86.sse2.padds.b" : "llvm.x86.sse2.paddus.b";
+         if(type.width == 16)
+            intrinsic = type.sign ? "llvm.x86.sse2.padds.w" : "llvm.x86.sse2.paddus.w";
+      }
+   
+      if(intrinsic)
+         return lp_build_intrinsic_binary(bld->builder, intrinsic, lp_build_vec_type(bld->type), a, b);
+   }
+
+   if(LLVMIsConstant(a) && LLVMIsConstant(b))
+      res = LLVMConstAdd(a, b);
+   else
+      res = LLVMBuildAdd(bld->builder, a, b, "");
+
+   /* clamp to ceiling of 1.0 */
+   if(bld->type.norm && (bld->type.floating || bld->type.fixed))
+      res = lp_build_min_simple(bld, res, bld->one);
+
+   /* XXX clamp to floor of -1 or 0??? */
+
+   return res;
+}
+
+
+/**
+ * Generate a - b
+ */
+LLVMValueRef
+lp_build_sub(struct lp_build_context *bld,
+             LLVMValueRef a,
+             LLVMValueRef b)
+{
+   const struct lp_type type = bld->type;
+   LLVMValueRef res;
+
+   if(b == bld->zero)
+      return a;
+   if(a == bld->undef || b == bld->undef)
+      return bld->undef;
+   if(a == b)
+      return bld->zero;
+
+   if(bld->type.norm) {
+      const char *intrinsic = NULL;
+
+      if(b == bld->one)
+        return bld->zero;
+
+      if(util_cpu_caps.has_sse2 &&
+         type.width * type.length == 128 &&
+         !type.floating && !type.fixed) {
+         if(type.width == 8)
+            intrinsic = type.sign ? "llvm.x86.sse2.psubs.b" : "llvm.x86.sse2.psubus.b";
+         if(type.width == 16)
+            intrinsic = type.sign ? "llvm.x86.sse2.psubs.w" : "llvm.x86.sse2.psubus.w";
+      }
+   
+      if(intrinsic)
+         return lp_build_intrinsic_binary(bld->builder, intrinsic, lp_build_vec_type(bld->type), a, b);
+   }
+
+   if(LLVMIsConstant(a) && LLVMIsConstant(b))
+      res = LLVMConstSub(a, b);
+   else
+      res = LLVMBuildSub(bld->builder, a, b, "");
+
+   if(bld->type.norm && (bld->type.floating || bld->type.fixed))
+      res = lp_build_max_simple(bld, res, bld->zero);
+
+   return res;
+}
+
+
+/**
+ * Normalized 8bit multiplication.
+ *
+ * - alpha plus one
+ *
+ *     makes the following approximation to the division (Sree)
+ *    
+ *       a*b/255 ~= (a*(b + 1)) >> 256
+ *    
+ *     which is the fastest method that satisfies the following OpenGL criteria
+ *    
+ *       0*0 = 0 and 255*255 = 255
+ *
+ * - geometric series
+ *
+ *     takes the geometric series approximation to the division
+ *
+ *       t/255 = (t >> 8) + (t >> 16) + (t >> 24) ..
+ *
+ *     in this case just the first two terms to fit in 16bit arithmetic
+ *
+ *       t/255 ~= (t + (t >> 8)) >> 8
+ *
+ *     note that just by itself it doesn't satisfies the OpenGL criteria, as
+ *     255*255 = 254, so the special case b = 255 must be accounted or roundoff
+ *     must be used
+ *
+ * - geometric series plus rounding
+ *
+ *     when using a geometric series division instead of truncating the result
+ *     use roundoff in the approximation (Jim Blinn)
+ *
+ *       t/255 ~= (t + (t >> 8) + 0x80) >> 8
+ *
+ *     achieving the exact results
+ *
+ * @sa Alvy Ray Smith, Image Compositing Fundamentals, Tech Memo 4, Aug 15, 1995, 
+ *     ftp://ftp.alvyray.com/Acrobat/4_Comp.pdf
+ * @sa Michael Herf, The "double blend trick", May 2000, 
+ *     http://www.stereopsis.com/doubleblend.html
+ */
+static LLVMValueRef
+lp_build_mul_u8n(LLVMBuilderRef builder,
+                 struct lp_type i16_type,
+                 LLVMValueRef a, LLVMValueRef b)
+{
+   LLVMValueRef c8;
+   LLVMValueRef ab;
+
+   c8 = lp_build_int_const_scalar(i16_type, 8);
+   
+#if 0
+   
+   /* a*b/255 ~= (a*(b + 1)) >> 256 */
+   b = LLVMBuildAdd(builder, b, lp_build_int_const_scalar(i16_type, 1), "");
+   ab = LLVMBuildMul(builder, a, b, "");
+
+#else
+   
+   /* ab/255 ~= (ab + (ab >> 8) + 0x80) >> 8 */
+   ab = LLVMBuildMul(builder, a, b, "");
+   ab = LLVMBuildAdd(builder, ab, LLVMBuildLShr(builder, ab, c8, ""), "");
+   ab = LLVMBuildAdd(builder, ab, lp_build_int_const_scalar(i16_type, 0x80), "");
+
+#endif
+   
+   ab = LLVMBuildLShr(builder, ab, c8, "");
+
+   return ab;
+}
+
+
+/**
+ * Generate a * b
+ */
+LLVMValueRef
+lp_build_mul(struct lp_build_context *bld,
+             LLVMValueRef a,
+             LLVMValueRef b)
+{
+   const struct lp_type type = bld->type;
+   LLVMValueRef shift;
+   LLVMValueRef res;
+
+   if(a == bld->zero)
+      return bld->zero;
+   if(a == bld->one)
+      return b;
+   if(b == bld->zero)
+      return bld->zero;
+   if(b == bld->one)
+      return a;
+   if(a == bld->undef || b == bld->undef)
+      return bld->undef;
+
+   if(!type.floating && !type.fixed && type.norm) {
+      if(type.width == 8) {
+         struct lp_type i16_type = lp_wider_type(type);
+         LLVMValueRef al, ah, bl, bh, abl, abh, ab;
+
+         lp_build_unpack2(bld->builder, type, i16_type, a, &al, &ah);
+         lp_build_unpack2(bld->builder, type, i16_type, b, &bl, &bh);
+
+         /* PMULLW, PSRLW, PADDW */
+         abl = lp_build_mul_u8n(bld->builder, i16_type, al, bl);
+         abh = lp_build_mul_u8n(bld->builder, i16_type, ah, bh);
+
+         ab = lp_build_pack2(bld->builder, i16_type, type, abl, abh);
+         
+         return ab;
+      }
+
+      /* FIXME */
+      assert(0);
+   }
+
+   if(type.fixed)
+      shift = lp_build_int_const_scalar(type, type.width/2);
+   else
+      shift = NULL;
+
+   if(LLVMIsConstant(a) && LLVMIsConstant(b)) {
+      res =  LLVMConstMul(a, b);
+      if(shift) {
+         if(type.sign)
+            res = LLVMConstAShr(res, shift);
+         else
+            res = LLVMConstLShr(res, shift);
+      }
+   }
+   else {
+      res = LLVMBuildMul(bld->builder, a, b, "");
+      if(shift) {
+         if(type.sign)
+            res = LLVMBuildAShr(bld->builder, res, shift, "");
+         else
+            res = LLVMBuildLShr(bld->builder, res, shift, "");
+      }
+   }
+
+   return res;
+}
+
+
+/**
+ * Small vector x scale multiplication optimization.
+ */
+LLVMValueRef
+lp_build_mul_imm(struct lp_build_context *bld,
+                 LLVMValueRef a,
+                 int b)
+{
+   LLVMValueRef factor;
+
+   if(b == 0)
+      return bld->zero;
+
+   if(b == 1)
+      return a;
+
+   if(b == -1)
+      return LLVMBuildNeg(bld->builder, a, "");
+
+   if(b == 2 && bld->type.floating)
+      return lp_build_add(bld, a, a);
+
+   if(util_is_pot(b)) {
+      unsigned shift = ffs(b) - 1;
+
+      if(bld->type.floating) {
+#if 0
+         /*
+          * Power of two multiplication by directly manipulating the mantissa.
+          *
+          * XXX: This might not be always faster, it will introduce a small error
+          * for multiplication by zero, and it will produce wrong results
+          * for Inf and NaN.
+          */
+         unsigned mantissa = lp_mantissa(bld->type);
+         factor = lp_build_int_const_scalar(bld->type, (unsigned long long)shift << mantissa);
+         a = LLVMBuildBitCast(bld->builder, a, lp_build_int_vec_type(bld->type), "");
+         a = LLVMBuildAdd(bld->builder, a, factor, "");
+         a = LLVMBuildBitCast(bld->builder, a, lp_build_vec_type(bld->type), "");
+         return a;
+#endif
+      }
+      else {
+         factor = lp_build_const_scalar(bld->type, shift);
+         return LLVMBuildShl(bld->builder, a, factor, "");
+      }
+   }
+
+   factor = lp_build_const_scalar(bld->type, (double)b);
+   return lp_build_mul(bld, a, factor);
+}
+
+
+/**
+ * Generate a / b
+ */
+LLVMValueRef
+lp_build_div(struct lp_build_context *bld,
+             LLVMValueRef a,
+             LLVMValueRef b)
+{
+   const struct lp_type type = bld->type;
+
+   if(a == bld->zero)
+      return bld->zero;
+   if(a == bld->one)
+      return lp_build_rcp(bld, b);
+   if(b == bld->zero)
+      return bld->undef;
+   if(b == bld->one)
+      return a;
+   if(a == bld->undef || b == bld->undef)
+      return bld->undef;
+
+   if(LLVMIsConstant(a) && LLVMIsConstant(b))
+      return LLVMConstFDiv(a, b);
+
+   if(util_cpu_caps.has_sse && type.width == 32 && type.length == 4)
+      return lp_build_mul(bld, a, lp_build_rcp(bld, b));
+
+   return LLVMBuildFDiv(bld->builder, a, b, "");
+}
+
+
+/**
+ * Linear interpolation.
+ *
+ * This also works for integer values with a few caveats.
+ *
+ * @sa http://www.stereopsis.com/doubleblend.html
+ */
+LLVMValueRef
+lp_build_lerp(struct lp_build_context *bld,
+              LLVMValueRef x,
+              LLVMValueRef v0,
+              LLVMValueRef v1)
+{
+   LLVMValueRef delta;
+   LLVMValueRef res;
+
+   delta = lp_build_sub(bld, v1, v0);
+
+   res = lp_build_mul(bld, x, delta);
+
+   res = lp_build_add(bld, v0, res);
+
+   if(bld->type.fixed)
+      /* XXX: This step is necessary for lerping 8bit colors stored on 16bits,
+       * but it will be wrong for other uses. Basically we need a more
+       * powerful lp_type, capable of further distinguishing the values
+       * interpretation from the value storage. */
+      res = LLVMBuildAnd(bld->builder, res, lp_build_int_const_scalar(bld->type, (1 << bld->type.width/2) - 1), "");
+
+   return res;
+}
+
+
+LLVMValueRef
+lp_build_lerp_2d(struct lp_build_context *bld,
+                 LLVMValueRef x,
+                 LLVMValueRef y,
+                 LLVMValueRef v00,
+                 LLVMValueRef v01,
+                 LLVMValueRef v10,
+                 LLVMValueRef v11)
+{
+   LLVMValueRef v0 = lp_build_lerp(bld, x, v00, v01);
+   LLVMValueRef v1 = lp_build_lerp(bld, x, v10, v11);
+   return lp_build_lerp(bld, y, v0, v1);
+}
+
+
+/**
+ * Generate min(a, b)
+ * Do checks for special cases.
+ */
+LLVMValueRef
+lp_build_min(struct lp_build_context *bld,
+             LLVMValueRef a,
+             LLVMValueRef b)
+{
+   if(a == bld->undef || b == bld->undef)
+      return bld->undef;
+
+   if(a == b)
+      return a;
+
+   if(bld->type.norm) {
+      if(a == bld->zero || b == bld->zero)
+         return bld->zero;
+      if(a == bld->one)
+         return b;
+      if(b == bld->one)
+         return a;
+   }
+
+   return lp_build_min_simple(bld, a, b);
+}
+
+
+/**
+ * Generate max(a, b)
+ * Do checks for special cases.
+ */
+LLVMValueRef
+lp_build_max(struct lp_build_context *bld,
+             LLVMValueRef a,
+             LLVMValueRef b)
+{
+   if(a == bld->undef || b == bld->undef)
+      return bld->undef;
+
+   if(a == b)
+      return a;
+
+   if(bld->type.norm) {
+      if(a == bld->one || b == bld->one)
+         return bld->one;
+      if(a == bld->zero)
+         return b;
+      if(b == bld->zero)
+         return a;
+   }
+
+   return lp_build_max_simple(bld, a, b);
+}
+
+
+/**
+ * Generate abs(a)
+ */
+LLVMValueRef
+lp_build_abs(struct lp_build_context *bld,
+             LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef vec_type = lp_build_vec_type(type);
+
+   if(!type.sign)
+      return a;
+
+   if(type.floating) {
+      /* Mask out the sign bit */
+      LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
+      unsigned long long absMask = ~(1ULL << (type.width - 1));
+      LLVMValueRef mask = lp_build_int_const_scalar(type, ((unsigned long long) absMask));
+      a = LLVMBuildBitCast(bld->builder, a, int_vec_type, "");
+      a = LLVMBuildAnd(bld->builder, a, mask, "");
+      a = LLVMBuildBitCast(bld->builder, a, vec_type, "");
+      return a;
+   }
+
+   if(type.width*type.length == 128 && util_cpu_caps.has_ssse3) {
+      switch(type.width) {
+      case 8:
+         return lp_build_intrinsic_unary(bld->builder, "llvm.x86.ssse3.pabs.b.128", vec_type, a);
+      case 16:
+         return lp_build_intrinsic_unary(bld->builder, "llvm.x86.ssse3.pabs.w.128", vec_type, a);
+      case 32:
+         return lp_build_intrinsic_unary(bld->builder, "llvm.x86.ssse3.pabs.d.128", vec_type, a);
+      }
+   }
+
+   return lp_build_max(bld, a, LLVMBuildNeg(bld->builder, a, ""));
+}
+
+
+LLVMValueRef
+lp_build_sgn(struct lp_build_context *bld,
+             LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef vec_type = lp_build_vec_type(type);
+   LLVMValueRef cond;
+   LLVMValueRef res;
+
+   /* Handle non-zero case */
+   if(!type.sign) {
+      /* if not zero then sign must be positive */
+      res = bld->one;
+   }
+   else if(type.floating) {
+      /* Take the sign bit and add it to 1 constant */
+      LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
+      LLVMValueRef mask = lp_build_int_const_scalar(type, (unsigned long long)1 << (type.width - 1));
+      LLVMValueRef sign;
+      LLVMValueRef one;
+      sign = LLVMBuildBitCast(bld->builder, a, int_vec_type, "");
+      sign = LLVMBuildAnd(bld->builder, sign, mask, "");
+      one = LLVMConstBitCast(bld->one, int_vec_type);
+      res = LLVMBuildOr(bld->builder, sign, one, "");
+      res = LLVMBuildBitCast(bld->builder, res, vec_type, "");
+   }
+   else
+   {
+      LLVMValueRef minus_one = lp_build_const_scalar(type, -1.0);
+      cond = lp_build_cmp(bld, PIPE_FUNC_GREATER, a, bld->zero);
+      res = lp_build_select(bld, cond, bld->one, minus_one);
+   }
+
+   /* Handle zero */
+   cond = lp_build_cmp(bld, PIPE_FUNC_EQUAL, a, bld->zero);
+   res = lp_build_select(bld, cond, bld->zero, bld->one);
+
+   return res;
+}
+
+
+enum lp_build_round_sse41_mode
+{
+   LP_BUILD_ROUND_SSE41_NEAREST = 0,
+   LP_BUILD_ROUND_SSE41_FLOOR = 1,
+   LP_BUILD_ROUND_SSE41_CEIL = 2,
+   LP_BUILD_ROUND_SSE41_TRUNCATE = 3
+};
+
+
+static INLINE LLVMValueRef
+lp_build_round_sse41(struct lp_build_context *bld,
+                     LLVMValueRef a,
+                     enum lp_build_round_sse41_mode mode)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef vec_type = lp_build_vec_type(type);
+   const char *intrinsic;
+
+   assert(type.floating);
+   assert(type.width*type.length == 128);
+   assert(lp_check_value(type, a));
+   assert(util_cpu_caps.has_sse4_1);
+
+   switch(type.width) {
+   case 32:
+      intrinsic = "llvm.x86.sse41.round.ps";
+      break;
+   case 64:
+      intrinsic = "llvm.x86.sse41.round.pd";
+      break;
+   default:
+      assert(0);
+      return bld->undef;
+   }
+
+   return lp_build_intrinsic_binary(bld->builder, intrinsic, vec_type, a,
+                                    LLVMConstInt(LLVMInt32Type(), mode, 0));
+}
+
+
+LLVMValueRef
+lp_build_trunc(struct lp_build_context *bld,
+               LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+
+   assert(type.floating);
+   assert(lp_check_value(type, a));
+
+   if(util_cpu_caps.has_sse4_1)
+      return lp_build_round_sse41(bld, a, LP_BUILD_ROUND_SSE41_TRUNCATE);
+   else {
+      LLVMTypeRef vec_type = lp_build_vec_type(type);
+      LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
+      LLVMValueRef res;
+      res = LLVMBuildFPToSI(bld->builder, a, int_vec_type, "");
+      res = LLVMBuildSIToFP(bld->builder, res, vec_type, "");
+      return res;
+   }
+}
+
+
+LLVMValueRef
+lp_build_round(struct lp_build_context *bld,
+               LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+
+   assert(type.floating);
+   assert(lp_check_value(type, a));
+
+   if(util_cpu_caps.has_sse4_1)
+      return lp_build_round_sse41(bld, a, LP_BUILD_ROUND_SSE41_NEAREST);
+   else {
+      LLVMTypeRef vec_type = lp_build_vec_type(type);
+      LLVMValueRef res;
+      res = lp_build_iround(bld, a);
+      res = LLVMBuildSIToFP(bld->builder, res, vec_type, "");
+      return res;
+   }
+}
+
+
+LLVMValueRef
+lp_build_floor(struct lp_build_context *bld,
+               LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+
+   assert(type.floating);
+
+   if(util_cpu_caps.has_sse4_1)
+      return lp_build_round_sse41(bld, a, LP_BUILD_ROUND_SSE41_FLOOR);
+   else {
+      LLVMTypeRef vec_type = lp_build_vec_type(type);
+      LLVMValueRef res;
+      res = lp_build_ifloor(bld, a);
+      res = LLVMBuildSIToFP(bld->builder, res, vec_type, "");
+      return res;
+   }
+}
+
+
+LLVMValueRef
+lp_build_ceil(struct lp_build_context *bld,
+              LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+
+   assert(type.floating);
+   assert(lp_check_value(type, a));
+
+   if(util_cpu_caps.has_sse4_1)
+      return lp_build_round_sse41(bld, a, LP_BUILD_ROUND_SSE41_CEIL);
+   else {
+      LLVMTypeRef vec_type = lp_build_vec_type(type);
+      LLVMValueRef res;
+      res = lp_build_iceil(bld, a);
+      res = LLVMBuildSIToFP(bld->builder, res, vec_type, "");
+      return res;
+   }
+}
+
+
+/**
+ * Convert to integer, through whichever rounding method that's fastest,
+ * typically truncating to zero.
+ */
+LLVMValueRef
+lp_build_itrunc(struct lp_build_context *bld,
+                LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
+
+   assert(type.floating);
+   assert(lp_check_value(type, a));
+
+   return LLVMBuildFPToSI(bld->builder, a, int_vec_type, "");
+}
+
+
+LLVMValueRef
+lp_build_iround(struct lp_build_context *bld,
+                LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
+   LLVMValueRef res;
+
+   assert(type.floating);
+   assert(lp_check_value(type, a));
+
+   if(util_cpu_caps.has_sse4_1) {
+      res = lp_build_round_sse41(bld, a, LP_BUILD_ROUND_SSE41_NEAREST);
+   }
+   else {
+      LLVMTypeRef vec_type = lp_build_vec_type(type);
+      LLVMValueRef mask = lp_build_int_const_scalar(type, (unsigned long long)1 << (type.width - 1));
+      LLVMValueRef sign;
+      LLVMValueRef half;
+
+      /* get sign bit */
+      sign = LLVMBuildBitCast(bld->builder, a, int_vec_type, "");
+      sign = LLVMBuildAnd(bld->builder, sign, mask, "");
+
+      /* sign * 0.5 */
+      half = lp_build_const_scalar(type, 0.5);
+      half = LLVMBuildBitCast(bld->builder, half, int_vec_type, "");
+      half = LLVMBuildOr(bld->builder, sign, half, "");
+      half = LLVMBuildBitCast(bld->builder, half, vec_type, "");
+
+      res = LLVMBuildAdd(bld->builder, a, half, "");
+   }
+
+   res = LLVMBuildFPToSI(bld->builder, res, int_vec_type, "");
+
+   return res;
+}
+
+
+/**
+ * Convert float[] to int[] with floor().
+ */
+LLVMValueRef
+lp_build_ifloor(struct lp_build_context *bld,
+                LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
+   LLVMValueRef res;
+
+   assert(type.floating);
+   assert(lp_check_value(type, a));
+
+   if(util_cpu_caps.has_sse4_1) {
+      res = lp_build_round_sse41(bld, a, LP_BUILD_ROUND_SSE41_FLOOR);
+   }
+   else {
+      /* Take the sign bit and add it to 1 constant */
+      LLVMTypeRef vec_type = lp_build_vec_type(type);
+      unsigned mantissa = lp_mantissa(type);
+      LLVMValueRef mask = lp_build_int_const_scalar(type, (unsigned long long)1 << (type.width - 1));
+      LLVMValueRef sign;
+      LLVMValueRef offset;
+
+      /* sign = a < 0 ? ~0 : 0 */
+      sign = LLVMBuildBitCast(bld->builder, a, int_vec_type, "");
+      sign = LLVMBuildAnd(bld->builder, sign, mask, "");
+      sign = LLVMBuildAShr(bld->builder, sign, lp_build_int_const_scalar(type, type.width - 1), "");
+      lp_build_name(sign, "floor.sign");
+
+      /* offset = -0.99999(9)f */
+      offset = lp_build_const_scalar(type, -(double)(((unsigned long long)1 << mantissa) - 1)/((unsigned long long)1 << mantissa));
+      offset = LLVMConstBitCast(offset, int_vec_type);
+
+      /* offset = a < 0 ? -0.99999(9)f : 0.0f */
+      offset = LLVMBuildAnd(bld->builder, offset, sign, "");
+      offset = LLVMBuildBitCast(bld->builder, offset, vec_type, "");
+      lp_build_name(offset, "floor.offset");
+
+      res = LLVMBuildAdd(bld->builder, a, offset, "");
+      lp_build_name(res, "floor.res");
+   }
+
+   res = LLVMBuildFPToSI(bld->builder, res, int_vec_type, "");
+   lp_build_name(res, "floor");
+
+   return res;
+}
+
+
+LLVMValueRef
+lp_build_iceil(struct lp_build_context *bld,
+               LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
+   LLVMValueRef res;
+
+   assert(type.floating);
+   assert(lp_check_value(type, a));
+
+   if(util_cpu_caps.has_sse4_1) {
+      res = lp_build_round_sse41(bld, a, LP_BUILD_ROUND_SSE41_CEIL);
+   }
+   else {
+      assert(0);
+      res = bld->undef;
+   }
+
+   res = LLVMBuildFPToSI(bld->builder, res, int_vec_type, "");
+
+   return res;
+}
+
+
+LLVMValueRef
+lp_build_sqrt(struct lp_build_context *bld,
+              LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef vec_type = lp_build_vec_type(type);
+   char intrinsic[32];
+
+   /* TODO: optimize the constant case */
+   /* TODO: optimize the constant case */
+
+   assert(type.floating);
+   util_snprintf(intrinsic, sizeof intrinsic, "llvm.sqrt.v%uf%u", type.length, type.width);
+
+   return lp_build_intrinsic_unary(bld->builder, intrinsic, vec_type, a);
+}
+
+
+LLVMValueRef
+lp_build_rcp(struct lp_build_context *bld,
+             LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+
+   if(a == bld->zero)
+      return bld->undef;
+   if(a == bld->one)
+      return bld->one;
+   if(a == bld->undef)
+      return bld->undef;
+
+   assert(type.floating);
+
+   if(LLVMIsConstant(a))
+      return LLVMConstFDiv(bld->one, a);
+
+   if(util_cpu_caps.has_sse && type.width == 32 && type.length == 4)
+      /* FIXME: improve precision */
+      return lp_build_intrinsic_unary(bld->builder, "llvm.x86.sse.rcp.ps", lp_build_vec_type(type), a);
+
+   return LLVMBuildFDiv(bld->builder, bld->one, a, "");
+}
+
+
+/**
+ * Generate 1/sqrt(a)
+ */
+LLVMValueRef
+lp_build_rsqrt(struct lp_build_context *bld,
+               LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+
+   assert(type.floating);
+
+   if(util_cpu_caps.has_sse && type.width == 32 && type.length == 4)
+      return lp_build_intrinsic_unary(bld->builder, "llvm.x86.sse.rsqrt.ps", lp_build_vec_type(type), a);
+
+   return lp_build_rcp(bld, lp_build_sqrt(bld, a));
+}
+
+
+/**
+ * Generate cos(a)
+ */
+LLVMValueRef
+lp_build_cos(struct lp_build_context *bld,
+              LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef vec_type = lp_build_vec_type(type);
+   char intrinsic[32];
+
+   /* TODO: optimize the constant case */
+
+   assert(type.floating);
+   util_snprintf(intrinsic, sizeof intrinsic, "llvm.cos.v%uf%u", type.length, type.width);
+
+   return lp_build_intrinsic_unary(bld->builder, intrinsic, vec_type, a);
+}
+
+
+/**
+ * Generate sin(a)
+ */
+LLVMValueRef
+lp_build_sin(struct lp_build_context *bld,
+              LLVMValueRef a)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef vec_type = lp_build_vec_type(type);
+   char intrinsic[32];
+
+   /* TODO: optimize the constant case */
+
+   assert(type.floating);
+   util_snprintf(intrinsic, sizeof intrinsic, "llvm.sin.v%uf%u", type.length, type.width);
+
+   return lp_build_intrinsic_unary(bld->builder, intrinsic, vec_type, a);
+}
+
+
+/**
+ * Generate pow(x, y)
+ */
+LLVMValueRef
+lp_build_pow(struct lp_build_context *bld,
+             LLVMValueRef x,
+             LLVMValueRef y)
+{
+   /* TODO: optimize the constant case */
+   if(LLVMIsConstant(x) && LLVMIsConstant(y))
+      debug_printf("%s: inefficient/imprecise constant arithmetic\n",
+                   __FUNCTION__);
+
+   return lp_build_exp2(bld, lp_build_mul(bld, lp_build_log2(bld, x), y));
+}
+
+
+/**
+ * Generate exp(x)
+ */
+LLVMValueRef
+lp_build_exp(struct lp_build_context *bld,
+             LLVMValueRef x)
+{
+   /* log2(e) = 1/log(2) */
+   LLVMValueRef log2e = lp_build_const_scalar(bld->type, 1.4426950408889634);
+
+   return lp_build_mul(bld, log2e, lp_build_exp2(bld, x));
+}
+
+
+/**
+ * Generate log(x)
+ */
+LLVMValueRef
+lp_build_log(struct lp_build_context *bld,
+             LLVMValueRef x)
+{
+   /* log(2) */
+   LLVMValueRef log2 = lp_build_const_scalar(bld->type, 0.69314718055994529);
+
+   return lp_build_mul(bld, log2, lp_build_exp2(bld, x));
+}
+
+
+#define EXP_POLY_DEGREE 3
+#define LOG_POLY_DEGREE 5
+
+
+/**
+ * Generate polynomial.
+ * Ex:  coeffs[0] + x * coeffs[1] + x^2 * coeffs[2].
+ */
+static LLVMValueRef
+lp_build_polynomial(struct lp_build_context *bld,
+                    LLVMValueRef x,
+                    const double *coeffs,
+                    unsigned num_coeffs)
+{
+   const struct lp_type type = bld->type;
+   LLVMValueRef res = NULL;
+   unsigned i;
+
+   /* TODO: optimize the constant case */
+   if(LLVMIsConstant(x))
+      debug_printf("%s: inefficient/imprecise constant arithmetic\n",
+                   __FUNCTION__);
+
+   for (i = num_coeffs; i--; ) {
+      LLVMValueRef coeff = lp_build_const_scalar(type, coeffs[i]);
+      if(res)
+         res = lp_build_add(bld, coeff, lp_build_mul(bld, x, res));
+      else
+         res = coeff;
+   }
+
+   if(res)
+      return res;
+   else
+      return bld->undef;
+}
+
+
+/**
+ * Minimax polynomial fit of 2**x, in range [-0.5, 0.5[
+ */
+const double lp_build_exp2_polynomial[] = {
+#if EXP_POLY_DEGREE == 5
+   9.9999994e-1, 6.9315308e-1, 2.4015361e-1, 5.5826318e-2, 8.9893397e-3, 1.8775767e-3
+#elif EXP_POLY_DEGREE == 4
+   1.0000026, 6.9300383e-1, 2.4144275e-1, 5.2011464e-2, 1.3534167e-2
+#elif EXP_POLY_DEGREE == 3
+   9.9992520e-1, 6.9583356e-1, 2.2606716e-1, 7.8024521e-2
+#elif EXP_POLY_DEGREE == 2
+   1.0017247, 6.5763628e-1, 3.3718944e-1
+#else
+#error
+#endif
+};
+
+
+void
+lp_build_exp2_approx(struct lp_build_context *bld,
+                     LLVMValueRef x,
+                     LLVMValueRef *p_exp2_int_part,
+                     LLVMValueRef *p_frac_part,
+                     LLVMValueRef *p_exp2)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef vec_type = lp_build_vec_type(type);
+   LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
+   LLVMValueRef ipart = NULL;
+   LLVMValueRef fpart = NULL;
+   LLVMValueRef expipart = NULL;
+   LLVMValueRef expfpart = NULL;
+   LLVMValueRef res = NULL;
+
+   if(p_exp2_int_part || p_frac_part || p_exp2) {
+      /* TODO: optimize the constant case */
+      if(LLVMIsConstant(x))
+         debug_printf("%s: inefficient/imprecise constant arithmetic\n",
+                      __FUNCTION__);
+
+      assert(type.floating && type.width == 32);
+
+      x = lp_build_min(bld, x, lp_build_const_scalar(type,  129.0));
+      x = lp_build_max(bld, x, lp_build_const_scalar(type, -126.99999));
+
+      /* ipart = int(x - 0.5) */
+      ipart = LLVMBuildSub(bld->builder, x, lp_build_const_scalar(type, 0.5f), "");
+      ipart = LLVMBuildFPToSI(bld->builder, ipart, int_vec_type, "");
+
+      /* fpart = x - ipart */
+      fpart = LLVMBuildSIToFP(bld->builder, ipart, vec_type, "");
+      fpart = LLVMBuildSub(bld->builder, x, fpart, "");
+   }
+
+   if(p_exp2_int_part || p_exp2) {
+      /* expipart = (float) (1 << ipart) */
+      expipart = LLVMBuildAdd(bld->builder, ipart, lp_build_int_const_scalar(type, 127), "");
+      expipart = LLVMBuildShl(bld->builder, expipart, lp_build_int_const_scalar(type, 23), "");
+      expipart = LLVMBuildBitCast(bld->builder, expipart, vec_type, "");
+   }
+
+   if(p_exp2) {
+      expfpart = lp_build_polynomial(bld, fpart, lp_build_exp2_polynomial,
+                                     Elements(lp_build_exp2_polynomial));
+
+      res = LLVMBuildMul(bld->builder, expipart, expfpart, "");
+   }
+
+   if(p_exp2_int_part)
+      *p_exp2_int_part = expipart;
+
+   if(p_frac_part)
+      *p_frac_part = fpart;
+
+   if(p_exp2)
+      *p_exp2 = res;
+}
+
+
+LLVMValueRef
+lp_build_exp2(struct lp_build_context *bld,
+              LLVMValueRef x)
+{
+   LLVMValueRef res;
+   lp_build_exp2_approx(bld, x, NULL, NULL, &res);
+   return res;
+}
+
+
+/**
+ * Minimax polynomial fit of log2(x)/(x - 1), for x in range [1, 2[
+ * These coefficients can be generate with
+ * http://www.boost.org/doc/libs/1_36_0/libs/math/doc/sf_and_dist/html/math_toolkit/toolkit/internals2/minimax.html
+ */
+const double lp_build_log2_polynomial[] = {
+#if LOG_POLY_DEGREE == 6
+   3.11578814719469302614, -3.32419399085241980044, 2.59883907202499966007, -1.23152682416275988241, 0.318212422185251071475, -0.0344359067839062357313
+#elif LOG_POLY_DEGREE == 5
+   2.8882704548164776201, -2.52074962577807006663, 1.48116647521213171641, -0.465725644288844778798, 0.0596515482674574969533
+#elif LOG_POLY_DEGREE == 4
+   2.61761038894603480148, -1.75647175389045657003, 0.688243882994381274313, -0.107254423828329604454
+#elif LOG_POLY_DEGREE == 3
+   2.28330284476918490682, -1.04913055217340124191, 0.204446009836232697516
+#else
+#error
+#endif
+};
+
+
+/**
+ * See http://www.devmaster.net/forums/showthread.php?p=43580
+ */
+void
+lp_build_log2_approx(struct lp_build_context *bld,
+                     LLVMValueRef x,
+                     LLVMValueRef *p_exp,
+                     LLVMValueRef *p_floor_log2,
+                     LLVMValueRef *p_log2)
+{
+   const struct lp_type type = bld->type;
+   LLVMTypeRef vec_type = lp_build_vec_type(type);
+   LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
+
+   LLVMValueRef expmask = lp_build_int_const_scalar(type, 0x7f800000);
+   LLVMValueRef mantmask = lp_build_int_const_scalar(type, 0x007fffff);
+   LLVMValueRef one = LLVMConstBitCast(bld->one, int_vec_type);
+
+   LLVMValueRef i = NULL;
+   LLVMValueRef exp = NULL;
+   LLVMValueRef mant = NULL;
+   LLVMValueRef logexp = NULL;
+   LLVMValueRef logmant = NULL;
+   LLVMValueRef res = NULL;
+
+   if(p_exp || p_floor_log2 || p_log2) {
+      /* TODO: optimize the constant case */
+      if(LLVMIsConstant(x))
+         debug_printf("%s: inefficient/imprecise constant arithmetic\n",
+                      __FUNCTION__);
+
+      assert(type.floating && type.width == 32);
+
+      i = LLVMBuildBitCast(bld->builder, x, int_vec_type, "");
+
+      /* exp = (float) exponent(x) */
+      exp = LLVMBuildAnd(bld->builder, i, expmask, "");
+   }
+
+   if(p_floor_log2 || p_log2) {
+      logexp = LLVMBuildLShr(bld->builder, exp, lp_build_int_const_scalar(type, 23), "");
+      logexp = LLVMBuildSub(bld->builder, logexp, lp_build_int_const_scalar(type, 127), "");
+      logexp = LLVMBuildSIToFP(bld->builder, logexp, vec_type, "");
+   }
+
+   if(p_log2) {
+      /* mant = (float) mantissa(x) */
+      mant = LLVMBuildAnd(bld->builder, i, mantmask, "");
+      mant = LLVMBuildOr(bld->builder, mant, one, "");
+      mant = LLVMBuildBitCast(bld->builder, mant, vec_type, "");
+
+      logmant = lp_build_polynomial(bld, mant, lp_build_log2_polynomial,
+                                    Elements(lp_build_log2_polynomial));
+
+      /* This effectively increases the polynomial degree by one, but ensures that log2(1) == 0*/
+      logmant = LLVMBuildMul(bld->builder, logmant, LLVMBuildSub(bld->builder, mant, bld->one, ""), "");
+
+      res = LLVMBuildAdd(bld->builder, logmant, logexp, "");
+   }
+
+   if(p_exp)
+      *p_exp = exp;
+
+   if(p_floor_log2)
+      *p_floor_log2 = logexp;
+
+   if(p_log2)
+      *p_log2 = res;
+}
+
+
+LLVMValueRef
+lp_build_log2(struct lp_build_context *bld,
+              LLVMValueRef x)
+{
+   LLVMValueRef res;
+   lp_build_log2_approx(bld, x, NULL, NULL, &res);
+   return res;
+}