summaryrefslogtreecommitdiffstats
path: root/src/mesa/drivers/dri
diff options
context:
space:
mode:
authorKenneth Graunke <[email protected]>2016-01-27 12:21:04 -0800
committerKenneth Graunke <[email protected]>2016-04-04 11:35:16 -0700
commit65fbc43d54403905e3eaea02372b5a364dc1d773 (patch)
tree612a74b4914f251f97f63fb1b46ef252b59c2228 /src/mesa/drivers/dri
parent3aa51e02d6659c8f2e34a4153ae140a4f90cc51b (diff)
i965: Add an INTEL_PRECISE_TRIG=1 option to fix SIN/COS output range.
The SIN and COS instructions on Intel hardware can produce values slightly outside of the [-1.0, 1.0] range for a small set of values. Obviously, this can break everyone's expectations about trig functions. According to an internal presentation, the COS instruction can produce a value up to 1.000027 for inputs in the range (0.08296, 0.09888). One suggested workaround is to multiply by 0.99997, scaling down the amplitude slightly. Apparently this also minimizes the error function, reducing the maximum error from 0.00006 to about 0.00003. When enabled, fixes 16 dEQP precision tests dEQP-GLES31.functional.shaders.builtin_functions.precision. {cos,sin}.{highp,mediump}_compute.{scalar,vec2,vec4,vec4}. at the cost of making every sin and cos call more expensive (about twice the number of cycles on recent hardware). Enabling this option has been shown to reduce GPUTest Volplosion performance by about 10%. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Matt Turner <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
Diffstat (limited to 'src/mesa/drivers/dri')
-rw-r--r--src/mesa/drivers/dri/i965/brw_compiler.c2
-rw-r--r--src/mesa/drivers/dri/i965/brw_compiler.h6
-rw-r--r--src/mesa/drivers/dri/i965/brw_fs_nir.cpp16
-rw-r--r--src/mesa/drivers/dri/i965/brw_vec4_nir.cpp16
4 files changed, 36 insertions, 4 deletions
diff --git a/src/mesa/drivers/dri/i965/brw_compiler.c b/src/mesa/drivers/dri/i965/brw_compiler.c
index 3da6aac2cbf..6509267a52e 100644
--- a/src/mesa/drivers/dri/i965/brw_compiler.c
+++ b/src/mesa/drivers/dri/i965/brw_compiler.c
@@ -147,6 +147,8 @@ brw_compiler_create(void *mem_ctx, const struct brw_device_info *devinfo)
brw_fs_alloc_reg_sets(compiler);
brw_vec4_alloc_reg_set(compiler);
+ compiler->precise_trig = env_var_as_boolean("INTEL_PRECISE_TRIG", false);
+
compiler->scalar_stage[MESA_SHADER_VERTEX] =
devinfo->gen >= 8 && !(INTEL_DEBUG & DEBUG_VEC4VS);
compiler->scalar_stage[MESA_SHADER_TESS_CTRL] = false;
diff --git a/src/mesa/drivers/dri/i965/brw_compiler.h b/src/mesa/drivers/dri/i965/brw_compiler.h
index 27a95a3c661..231e0001d54 100644
--- a/src/mesa/drivers/dri/i965/brw_compiler.h
+++ b/src/mesa/drivers/dri/i965/brw_compiler.h
@@ -92,6 +92,12 @@ struct brw_compiler {
bool scalar_stage[MESA_SHADER_STAGES];
struct gl_shader_compiler_options glsl_compiler_options[MESA_SHADER_STAGES];
+
+ /**
+ * Apply workarounds for SIN and COS output range problems.
+ * This can negatively impact performance.
+ */
+ bool precise_trig;
};
diff --git a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
index 7839428c52e..5cca91ec5b4 100644
--- a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
@@ -775,12 +775,24 @@ fs_visitor::nir_emit_alu(const fs_builder &bld, nir_alu_instr *instr)
break;
case nir_op_fsin:
- inst = bld.emit(SHADER_OPCODE_SIN, result, op[0]);
+ if (!compiler->precise_trig) {
+ inst = bld.emit(SHADER_OPCODE_SIN, result, op[0]);
+ } else {
+ fs_reg tmp = bld.vgrf(BRW_REGISTER_TYPE_F);
+ inst = bld.emit(SHADER_OPCODE_SIN, tmp, op[0]);
+ inst = bld.MUL(result, tmp, brw_imm_f(0.99997));
+ }
inst->saturate = instr->dest.saturate;
break;
case nir_op_fcos:
- inst = bld.emit(SHADER_OPCODE_COS, result, op[0]);
+ if (!compiler->precise_trig) {
+ inst = bld.emit(SHADER_OPCODE_COS, result, op[0]);
+ } else {
+ fs_reg tmp = bld.vgrf(BRW_REGISTER_TYPE_F);
+ inst = bld.emit(SHADER_OPCODE_COS, tmp, op[0]);
+ inst = bld.MUL(result, tmp, brw_imm_f(0.99997));
+ }
inst->saturate = instr->dest.saturate;
break;
diff --git a/src/mesa/drivers/dri/i965/brw_vec4_nir.cpp b/src/mesa/drivers/dri/i965/brw_vec4_nir.cpp
index ee6929b16a2..6c8fd06fb5e 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4_nir.cpp
+++ b/src/mesa/drivers/dri/i965/brw_vec4_nir.cpp
@@ -1101,12 +1101,24 @@ vec4_visitor::nir_emit_alu(nir_alu_instr *instr)
break;
case nir_op_fsin:
- inst = emit_math(SHADER_OPCODE_SIN, dst, op[0]);
+ if (!compiler->precise_trig) {
+ inst = emit_math(SHADER_OPCODE_SIN, dst, op[0]);
+ } else {
+ src_reg tmp = src_reg(this, glsl_type::vec4_type);
+ inst = emit_math(SHADER_OPCODE_SIN, dst_reg(tmp), op[0]);
+ inst = emit(MUL(dst, tmp, brw_imm_f(0.99997)));
+ }
inst->saturate = instr->dest.saturate;
break;
case nir_op_fcos:
- inst = emit_math(SHADER_OPCODE_COS, dst, op[0]);
+ if (!compiler->precise_trig) {
+ inst = emit_math(SHADER_OPCODE_COS, dst, op[0]);
+ } else {
+ src_reg tmp = src_reg(this, glsl_type::vec4_type);
+ inst = emit_math(SHADER_OPCODE_COS, dst_reg(tmp), op[0]);
+ inst = emit(MUL(dst, tmp, brw_imm_f(0.99997)));
+ }
inst->saturate = instr->dest.saturate;
break;