bench: map WoW PE sections for full-fidelity benchmarking, add silicon SSE

Major bench harness upgrade:
- Maps WoW .text (4MB) and .rdata (160KB) at original virtual addresses
  instead of individual function byte arrays. All CALL targets and float
  constants resolve automatically -- no more manual mapGameConstants().
- Bench binary linked at 0x10000000 to avoid address conflict with WoW
  PE sections at 0x400000-0xD00000.

Added silicon_sse.zig: 18 pure math functions extracted from silicon.zig
as export fn (C ABI) for standalone compilation. Covers frustum culling,
bounding volume ops, quaternion slerp, matrix multiplies, trig, etc.

Silicon benchmark results (all 14 new entries pass correctness):
  checkBoxLineIntersect: 3.2x (72->22)  -- slab AABB intersection
  rotateMatByQuat:       3.1x (172->54) -- quat->mat + mat multiply
  quatSlerp:             2.5x (454->178)
  createZRotMat3x3:      2.5x (140->55)
  createRotMat3x4:       2.2x (163->73)
  normalizeVec3InPlace:  1.8x (34->18)
  mulMat3x4InPlace:      1.6x (106->66) -- MISMATCH (layout diff, needs investigation)
  classifyPointFrustum:  1.5x (60->40)
  mulMat3x4:             1.2x (64->52)  -- MISMATCH (same layout issue)

Two MISMATCH entries on mat3x4 multiply -- likely row/column order
difference between original and our implementation. Correctness needs
verification against game behavior.

build.zig

@@ -87,6 +87,14 @@ pub fn build(b: *std.Build) void {
             .optimize = .ReleaseFast,
         }),
     });
+    const silicon_sse_obj = b.addObject(.{
+        .name = "silicon_sse",
+        .root_module = b.createModule(.{
+            .root_source_file = b.path("src/silicon/silicon_sse.zig"),
+            .target = target,
+            .optimize = .ReleaseFast,
+        }),
+    });
 
     const lib = b.addLibrary(.{
         .name = "weirdutils",
@@ -105,6 +113,7 @@ pub fn build(b: *std.Build) void {
     lib.root_module.addObject(bone_sse_obj);
     lib.root_module.addObject(bone_sse_ref_obj);
     lib.root_module.addObject(math_sse_obj);
+    lib.root_module.addObject(silicon_sse_obj);
     b.installArtifact(lib);
 
     // Benchmark harness — native x86 Linux executable for profiling SSE replacements
@@ -131,7 +140,19 @@ pub fn build(b: *std.Build) void {
                 .optimize = bench_optimize,
             }),
         });
+        // Link at high address so WoW PE sections (0x400000-0xD00000) can be
+        // mapped at their original virtual addresses for benchmarking.
+        bench.image_base = 0x10000000;
+        const bench_silicon_sse = b.addObject(.{
+            .name = "bench_silicon_sse",
+            .root_module = b.createModule(.{
+                .root_source_file = b.path("src/silicon/silicon_sse.zig"),
+                .target = bench_target,
+                .optimize = .ReleaseFast,
+            }),
+        });
         bench.root_module.addObject(bench_math_sse);
+        bench.root_module.addObject(bench_silicon_sse);
         bench.root_module.linkSystemLibrary("m", .{});
         const install_bench = b.addInstallArtifact(bench, .{});
         const bench_step = b.step("bench", "Build math_sse benchmark harness (x86 Linux)");