AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

r300: fix RC_OMOD_DIV_2 modifier

Browse Source 
Backend only recognizes the MUL a*0.5 pattern if there is an immediate
as one of the sources, however by then the source would we already
coverted to RC_FILE_NONE with constant half swizzles. So teach
peephole_omod to recognize this pattern as well.

RV530:
total instructions in shared programs: 128860 -> 128750 (-0.09%)
instructions in affected programs: 11942 -> 11832 (-0.92%)
helped: 106
HURT: 17
total presub in shared programs: 8739 -> 8736 (-0.03%)
presub in affected programs: 32 -> 29 (-9.38%)
helped: 3
HURT: 0
total omod in shared programs: 427 -> 1212 (183.84%)
omod in affected programs: 38 -> 823 (2065.79%)
helped: 0
HURT: 160
total temps in shared programs: 17544 -> 17554 (0.06%)
temps in affected programs: 70 -> 80 (14.29%)
helped: 0
HURT: 10
total lits in shared programs: 3153 -> 3159 (0.19%)
lits in affected programs: 9 -> 15 (66.67%)
helped: 0
HURT: 6
total cycles in shared programs: 191334 -> 191253 (-0.04%)
cycles in affected programs: 21240 -> 21159 (-0.38%)
helped: 101
HURT: 27

Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28784>
This commit is contained in:
Pavel Ondračka
2024-05-03 10:37:07 +02:00
parent d94d2a05b2
commit ae06e018fa
1 changed files with 45 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/gallium/drivers/r300/compiler/radeon_optimize.c
+45 -24
View File
@@ -707,9 +707,28 @@ static int peephole_mul_omod(
for (i = 0; i < 2; i++) {
unsigned int j;
if (inst_mul->U.I.SrcReg[i].File != RC_FILE_CONSTANT
&& inst_mul->U.I.SrcReg[i].File != RC_FILE_TEMPORARY) {
&& inst_mul->U.I.SrcReg[i].File != RC_FILE_TEMPORARY
&& inst_mul->U.I.SrcReg[i].File != RC_FILE_NONE) {
return 0;
}
/* The only relevant case with constant swizzles we should check for
* is multiply by one half.
*/
if (inst_mul->U.I.SrcReg[i].File == RC_FILE_NONE) {
for (j = 0; j < 4; j++) {
swz = GET_SWZ(inst_mul->U.I.SrcReg[i].Swizzle, j);
if (swz == RC_SWIZZLE_UNUSED) {
continue;
}
if (swz != RC_SWIZZLE_HALF) {
return 0;
} else {
omod_op = RC_OMOD_DIV_2;
}
}
}
if (inst_mul->U.I.SrcReg[i].File == RC_FILE_TEMPORARY) {
if (temp_index != -1) {
/* The instruction has two temp sources */
@@ -748,30 +767,32 @@ static int peephole_mul_omod(
}
}
if (!rc_src_reg_is_immediate(c, inst_mul->U.I.SrcReg[const_index].File,
inst_mul->U.I.SrcReg[const_index].Index)) {
return 0;
}
const_value = rc_get_constant_value(c,
inst_mul->U.I.SrcReg[const_index].Index,
inst_mul->U.I.SrcReg[const_index].Swizzle,
inst_mul->U.I.SrcReg[const_index].Negate,
chan);
if (omod_op == RC_OMOD_DISABLE) {
if (!rc_src_reg_is_immediate(c, inst_mul->U.I.SrcReg[const_index].File,
inst_mul->U.I.SrcReg[const_index].Index)) {
return 0;
}
const_value = rc_get_constant_value(c,
inst_mul->U.I.SrcReg[const_index].Index,
inst_mul->U.I.SrcReg[const_index].Swizzle,
inst_mul->U.I.SrcReg[const_index].Negate,
chan);
if (const_value == 2.0f) {
omod_op = RC_OMOD_MUL_2;
} else if (const_value == 4.0f) {
omod_op = RC_OMOD_MUL_4;
} else if (const_value == 8.0f) {
omod_op = RC_OMOD_MUL_8;
} else if (const_value == (1.0f / 2.0f)) {
omod_op = RC_OMOD_DIV_2;
} else if (const_value == (1.0f / 4.0f)) {
omod_op = RC_OMOD_DIV_4;
} else if (const_value == (1.0f / 8.0f)) {
omod_op = RC_OMOD_DIV_8;
} else {
return 0;
if (const_value == 2.0f) {
omod_op = RC_OMOD_MUL_2;
} else if (const_value == 4.0f) {
omod_op = RC_OMOD_MUL_4;
} else if (const_value == 8.0f) {
omod_op = RC_OMOD_MUL_8;
} else if (const_value == (1.0f / 2.0f)) {
omod_op = RC_OMOD_DIV_2;
} else if (const_value == (1.0f / 4.0f)) {
omod_op = RC_OMOD_DIV_4;
} else if (const_value == (1.0f / 8.0f)) {
omod_op = RC_OMOD_DIV_8;
} else {
return 0;
}
}
writer_list = rc_variable_list_get_writers_one_reader(var_list,