diff options
Diffstat (limited to 'etc/compile-by-zinc/femulDisplayScheduled0.log')
-rw-r--r-- | etc/compile-by-zinc/femulDisplayScheduled0.log | 404 |
1 files changed, 192 insertions, 212 deletions
diff --git a/etc/compile-by-zinc/femulDisplayScheduled0.log b/etc/compile-by-zinc/femulDisplayScheduled0.log index 4bcf8b04c..a5e399038 100644 --- a/etc/compile-by-zinc/femulDisplayScheduled0.log +++ b/etc/compile-by-zinc/femulDisplayScheduled0.log @@ -1,213 +1,193 @@ +asm ( // Convention is low_reg:high_reg -"mov %[rx11], %[arg0]\t\n" -FIXME: lea for x48 = x11 * 0x13 -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r12], %rdx\t\n" -"mov %[rx17], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x55_tmp = x48 * x17 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r10], %[r9]\t\n" // bucket: x50_high + x52_high + x54_high + x56_high -"mov %[r11], %[r8]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low -"mov %[rx9], %[arg0]\t\n" -FIXME: lea for x47 = x9 * 0x13 -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r7], %rdx\t\n" -"mov %[rx19], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x53_tmp = x47 * x19 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r9]\t\n" // bucket: x50_high + x52_high + x54_high + x56_high -"add %[r10], %[r8]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low -"mov %[rx7], %[arg0]\t\n" -FIXME: lea for x46 = x7 * 0x13 -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r6], %rdx\t\n" -"mov %[rx18], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x51_tmp = x46 * x18 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r9]\t\n" // bucket: x50_high + x52_high + x54_high + x56_high -"adc %[r11], %[r8]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx5], %rdx\t\n" -"mov %[rx13], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x20_tmp = x5 * x13 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r9]\t\n" // bucket: x50_high + x52_high + x54_high + x56_high -"adc %[r10], %[r8]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low -"mov %[rx10], %[arg0]\t\n" -FIXME: lea for x45 = x10 * 0x13 -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r6], %rdx\t\n" -"mov %[rx15], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x49_tmp = x45 * x15 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r9]\t\n" // bucket: x50_high + x52_high + x54_high + x56_high -"adc %[r11], %[r8]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low -"adcx $0, %[r9]\t\n" // bucket: x50_high + x52_high + x54_high + x56_high -"and $0x7ffffffffffff, %[r8]\t\n" // x70 = x56_low & 0x7ffffffffffff -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r12], %rdx\t\n" -"mov %[rx19], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x61_tmp = x48 * x19 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -r9 <- SHR r8:r9, 0x33; // x69 = x56_low:x56_high >> 0x33 -"mov %[r9], %[r4]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low -"mov %[r11], %[r5]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high -"add %[r10], %[r4]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r7], %rdx\t\n" -"mov %[rx18], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x59_tmp = x47 * x18 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r5]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high -"adc %[r11], %[r4]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r6], %rdx\t\n" -"mov %[rx17], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x57_tmp = x45 * x17 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r5]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high -"adc %[r10], %[r4]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx5], %rdx\t\n" -"mov %[rx15], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x21_tmp = x5 * x15 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r5]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high -"adc %[r11], %[r4]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx7], %rdx\t\n" -"mov %[rx13], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x22_tmp = x7 * x13 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r5]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high -"adc %[r10], %[r4]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low -"adcx $0, %[r5]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high -"and $0x7ffffffffffff, %[r4]\t\n" // x73 = x71_low & 0x7ffffffffffff -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r12], %rdx\t\n" -"mov %[rx18], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x65_tmp = x48 * x18 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -r5 <- SHR r4:r5, 0x33; // x72 = x71_low:x71_high >> 0x33 -"mov %[r5], %[r9]\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low -"mov %[r10], %[r12]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high -"add %[r11], %[r9]\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r6], %rdx\t\n" -"mov %[rx19], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x63_tmp = x45 * x19 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r12]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high -"adc %[r10], %[r9]\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx7], %rdx\t\n" -"mov %[rx15], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x27_tmp = x7 * x15 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r12]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high -"adc %[r11], %[r9]\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx5], %rdx\t\n" -"mov %[rx17], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x24_tmp = x5 * x17 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r12]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high -"adc %[r10], %[r9]\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx9], %rdx\t\n" -"mov %[rx13], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x25_tmp = x9 * x13 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r12]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high -"adc %[r11], %[r9]\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low -"adcx $0, %[r12]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high -"and $0x7ffffffffffff, %[r9]\t\n" // x76 = x74_low & 0x7ffffffffffff -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[r6], %rdx\t\n" -"mov %[rx18], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x67_tmp = x45 * x18 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -r12 <- SHR r9:r12, 0x33; // x75 = x74_low:x74_high >> 0x33 -"mov %[r12], %[r5]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low -"mov %[r11], %[r6]\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high -"add %[r10], %[r5]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx9], %rdx\t\n" -"mov %[rx15], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x34_tmp = x9 * x15 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r6]\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high -"adc %[r11], %[r5]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx7], %rdx\t\n" -"mov %[rx17], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x32_tmp = x7 * x17 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r6]\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high -"adc %[r10], %[r5]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx5], %rdx\t\n" -"mov %[rx19], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x29_tmp = x5 * x19 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r6]\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high -"adc %[r11], %[r5]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx11], %rdx\t\n" -"mov %[rx13], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x30_tmp = x11 * x13 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r6]\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high -"adc %[r10], %[r5]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low -"adcx $0, %[r6]\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high -"and $0x7ffffffffffff, %[r5]\t\n" // x79 = x77_low & 0x7ffffffffffff -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx9], %rdx\t\n" -"mov %[rx17], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x43_tmp = x9 * x17 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -r6 <- SHR r5:r6, 0x33; // x78 = x77_low:x77_high >> 0x33 -"mov %[r6], %[r7]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low -"mov %[r10], %[r12]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high -"add %[r11], %[r7]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx7], %rdx\t\n" -"mov %[rx19], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x41_tmp = x7 * x19 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r12]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high -"adc %[r10], %[r7]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx11], %rdx\t\n" -"mov %[rx15], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x39_tmp = x11 * x15 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r12]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high -"adc %[r11], %[r7]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx5], %rdx\t\n" -"mov %[rx18], %[arg0]\t\n" -"mulx %[arg0], %[r11], %[r10]\t\n" // x36_tmp = x5 * x18 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r11], %[r12]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high -"adc %[r10], %[r7]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low -"mov %rdx, %[rdx_backup]\t\n" // XXX: How do I specify that a particular register should be rdx? -"mov %[rx10], %rdx\t\n" -"mov %[rx13], %[arg0]\t\n" -"mulx %[arg0], %[r10], %[r11]\t\n" // x37_tmp = x10 * x13 -"mov %[rdx_backup], %rdx\t\n" // XXX: How do I specify that a particular register should be rdx? -"adx %[r10], %[r12]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high -"adc %[r11], %[r7]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low -"adcx $0, %[r12]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high -"and $0x7ffffffffffff, %[r7]\t\n" // x82 = x80_low & 0x7ffffffffffff -r12 <- SHR r7:r12, 0x33; // x81 = x80_low:x80_high >> 0x33 -FIXME: lea for x83 = x81 * 0x13 -"adx %[r12], %[r8]\t\n" // bucket: x84 = x70 + x83 -r8 <- SHR r8, 0x33; // x85 = x70 >> 0x33 -"mov %[r8], %[r12]\t\n" -"and $0x7ffffffffffff, %[r12]\t\n" // x86 = x84 & 0x7ffffffffffff -"adx %[r4], %[r8]\t\n" // bucket: x87 = x85 + x73 -r8 <- SHR r8, 0x33; // x88 = x85 >> 0x33 -"mov %[r8], %[r4]\t\n" -"and $0x7ffffffffffff, %[r4]\t\n" // x89 = x87 & 0x7ffffffffffff -"adx %[r9], %[r8]\t\n" // bucket: x90 = x88 + x76 +// FIXME: lea for x48 = x11 * 0x13 +"mov $0x13, %rdx\t\n" +"mov %[mx11], %[rx86]\t\n" +"mulx %[rx86], %rdx, %[rx86]\t\n" // x48 = x11 * 0x13 +"mov %[rx86], %rdx\t\n" +"mov %[mx17], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x55_tmp = x48 * x17 +"mov %r13, %r12\t\n" // bucket: x50_high + x52_high + x54_high + x56_high +"mov %r14, %[rx90]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low +// FIXME: lea for x47 = x9 * 0x13 +"mov $0x13, %rdx\t\n" +"mov %[mx9], %[rx82]\t\n" +"mulx %[rx82], %rdx, %[rx82]\t\n" // x47 = x9 * 0x13 +"mov %[rx82], %rdx\t\n" +"mov %[mx19], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x53_tmp = x47 * x19 +"adx %r14, %r12\t\n" // bucket: x50_high + x52_high + x54_high + x56_high +"add %r13, %[rx90]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low +// FIXME: lea for x46 = x7 * 0x13 +"mov $0x13, %rdx\t\n" +"mov %[mx7], %r9\t\n" +"mulx %r9, %rdx, %r9\t\n" // x46 = x7 * 0x13 +"mov %r9, %rdx\t\n" +"mov %[mx18], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x51_tmp = x46 * x18 +"adx %r13, %r12\t\n" // bucket: x50_high + x52_high + x54_high + x56_high +"adc %r14, %[rx90]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low +"mov %[mx5], %rdx\t\n" +"mov %[mx13], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x20_tmp = x5 * x13 +"adx %r14, %r12\t\n" // bucket: x50_high + x52_high + x54_high + x56_high +"adc %r13, %[rx90]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low +// FIXME: lea for x45 = x10 * 0x13 +"mov $0x13, %rdx\t\n" +"mov %[mx10], %r9\t\n" +"mulx %r9, %rdx, %r9\t\n" // x45 = x10 * 0x13 +"mov %r9, %rdx\t\n" +"mov %[mx15], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x49_tmp = x45 * x15 +"adx %r13, %r12\t\n" // bucket: x50_high + x52_high + x54_high + x56_high +"adc %r14, %[rx90]\t\n" // bucket: x50_low + x52_low + x54_low + x56_low +"adcx $0, %r12\t\n" // bucket: x50_high + x52_high + x54_high + x56_high +"and $0x7ffffffffffff, %[rx90]\t\n" // x70 = x56_low & 0x7ffffffffffff +"mov %[rx86], %rdx\t\n" +"mov %[mx19], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x61_tmp = x48 * x19 +"mov %[rx90], %rdx\t\n" +"mov %[rx90], %r12\t\n" +"mov %rdx, %[rx90]\t\n" +"shrd $0x33, %r12, %[rx90]\t\n" // x69 = x56_low:x56_high >> 0x33 +"mov %r12, %[rx89]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low +"mov %r14, %[rx79]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high +"add %r13, %[rx89]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low +"mov %[rx82], %rdx\t\n" +"mov %[mx18], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x59_tmp = x47 * x18 +"adx %r13, %[rx79]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high +"adc %r14, %[rx89]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low +"mov %r9, %rdx\t\n" +"mov %[mx17], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x57_tmp = x45 * x17 +"adx %r14, %[rx79]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high +"adc %r13, %[rx89]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low +"mov %[mx5], %rdx\t\n" +"mov %[mx15], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x21_tmp = x5 * x15 +"adx %r13, %[rx79]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high +"adc %r14, %[rx89]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low +"mov %[mx7], %rdx\t\n" +"mov %[mx13], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x22_tmp = x7 * x13 +"adx %r14, %[rx79]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high +"adc %r13, %[rx89]\t\n" // bucket: x23_low + x58_low + x60_low + x62_low + x71_low +"adcx $0, %[rx79]\t\n" // bucket: x23_high + x58_high + x60_high + x62_high + x71_high +"and $0x7ffffffffffff, %[rx89]\t\n" // x73 = x71_low & 0x7ffffffffffff +"mov %[rx86], %rdx\t\n" +"mov %[mx18], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x65_tmp = x48 * x18 +"mov %[rx89], %rdx\t\n" +"mov %[rx89], %[rx79]\t\n" +"mov %rdx, %[rx89]\t\n" +"shrd $0x33, %[rx79], %[rx89]\t\n" // x72 = x71_low:x71_high >> 0x33 +"mov %[rx79], %r12\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low +"mov %r13, %[rx86]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high +"add %r14, %r12\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low +"mov %r9, %rdx\t\n" +"mov %[mx19], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x63_tmp = x45 * x19 +"adx %r14, %[rx86]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high +"adc %r13, %r12\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low +"mov %[mx7], %rdx\t\n" +"mov %[mx15], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x27_tmp = x7 * x15 +"adx %r13, %[rx86]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high +"adc %r14, %r12\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low +"mov %[mx5], %rdx\t\n" +"mov %[mx17], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x24_tmp = x5 * x17 +"adx %r14, %[rx86]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high +"adc %r13, %r12\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low +"mov %[mx9], %rdx\t\n" +"mov %[mx13], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x25_tmp = x9 * x13 +"adx %r13, %[rx86]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high +"adc %r14, %r12\t\n" // bucket: x26_low + x28_low + x64_low + x66_low + x74_low +"adcx $0, %[rx86]\t\n" // bucket: x26_high + x28_high + x64_high + x66_high + x74_high +"and $0x7ffffffffffff, %r12\t\n" // x76 = x74_low & 0x7ffffffffffff +"mov %r9, %rdx\t\n" +"mov %[mx18], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x67_tmp = x45 * x18 +"mov %r12, %rdx\t\n" +"mov %r12, %[rx86]\t\n" +"mov %rdx, %r12\t\n" +"shrd $0x33, %[rx86], %r12\t\n" // x75 = x74_low:x74_high >> 0x33 +"mov %[rx86], %[rx79]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low +"mov %r14, %r9\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high +"add %r13, %[rx79]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low +"mov %[mx9], %rdx\t\n" +"mov %[mx15], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x34_tmp = x9 * x15 +"adx %r13, %r9\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high +"adc %r14, %[rx79]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low +"mov %[mx7], %rdx\t\n" +"mov %[mx17], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x32_tmp = x7 * x17 +"adx %r14, %r9\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high +"adc %r13, %[rx79]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low +"mov %[mx5], %rdx\t\n" +"mov %[mx19], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x29_tmp = x5 * x19 +"adx %r13, %r9\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high +"adc %r14, %[rx79]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low +"mov %[mx11], %rdx\t\n" +"mov %[mx13], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x30_tmp = x11 * x13 +"adx %r14, %r9\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high +"adc %r13, %[rx79]\t\n" // bucket: x31_low + x33_low + x35_low + x68_low + x77_low +"adcx $0, %r9\t\n" // bucket: x31_high + x33_high + x35_high + x68_high + x77_high +"and $0x7ffffffffffff, %[rx79]\t\n" // x79 = x77_low & 0x7ffffffffffff +"mov %[mx9], %rdx\t\n" +"mov %[mx17], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x43_tmp = x9 * x17 +"mov %[rx79], %rdx\t\n" +"mov %[rx79], %r9\t\n" +"mov %rdx, %[rx79]\t\n" +"shrd $0x33, %r9, %[rx79]\t\n" // x78 = x77_low:x77_high >> 0x33 +"mov %r9, %[rx82]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low +"mov %r13, %[rx86]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high +"add %r14, %[rx82]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low +"mov %[mx7], %rdx\t\n" +"mov %[mx19], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x41_tmp = x7 * x19 +"adx %r14, %[rx86]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high +"adc %r13, %[rx82]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low +"mov %[mx11], %rdx\t\n" +"mov %[mx15], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x39_tmp = x11 * x15 +"adx %r13, %[rx86]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high +"adc %r14, %[rx82]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low +"mov %[mx5], %rdx\t\n" +"mov %[mx18], %r14\t\n" +"mulx %r14, %r14, %r13\t\n" // x36_tmp = x5 * x18 +"adx %r14, %[rx86]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high +"adc %r13, %[rx82]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low +"mov %[mx10], %rdx\t\n" +"mov %[mx13], %r13\t\n" +"mulx %r13, %r13, %r14\t\n" // x37_tmp = x10 * x13 +"adx %r13, %[rx86]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high +"adc %r14, %[rx82]\t\n" // bucket: x38_low + x40_low + x42_low + x44_low + x80_low +"adcx $0, %[rx86]\t\n" // bucket: x38_high + x40_high + x42_high + x44_high + x80_high +"and $0x7ffffffffffff, %[rx82]\t\n" // x82 = x80_low & 0x7ffffffffffff +"mov %[rx82], %rdx\t\n" +"mov %[rx82], %[rx86]\t\n" +"mov %rdx, %[rx82]\t\n" +"shrd $0x33, %[rx86], %[rx82]\t\n" // x81 = x80_low:x80_high >> 0x33 +// FIXME: lea for x83 = x81 * 0x13 +"mov $0x13, %rdx\t\n" +"mulx %[rx86], %rdx, %[rx86]\t\n" // x83 = x81 * 0x13 +"adx %[rx86], %[rx90]\t\n" // bucket: x84 = x70 + x83 +"shr $0x33, %[rx90]\t\n" // x85 = x70 >> 0x33 +"mov %[rx90], %[rx86]\t\n" +"and $0x7ffffffffffff, %[rx86]\t\n" // x86 = x84 & 0x7ffffffffffff +"adx %[rx89], %[rx90]\t\n" // bucket: x87 = x85 + x73 +"shr $0x33, %[rx90]\t\n" // x88 = x85 >> 0x33 +"mov %[rx90], %[rx89]\t\n" +"and $0x7ffffffffffff, %[rx89]\t\n" // x89 = x87 & 0x7ffffffffffff +"adx %r12, %[rx90]\t\n" // bucket: x90 = x88 + x76 +: [rx89] "=&r" (x89), [rx79] "=&r" (x79), [rx82] "=&r" (x82), [rx90] "=&r" (x90), [rx86] "=&r" (x86) +: [mx9] "m" (x9), [mx7] "m" (x7), [mx5] "m" (x5), [mx10] "m" (x10), [mx11] "m" (x11), [mx13] "m" (x13), [mx15] "m" (x15), [mx17] "m" (x17), [mx18] "m" (x18), [mx19] "m" (x19) +: "cc", "rdx", "r7", "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15" +); |