diff options
| author | Christian Krinitsin <mail@krinitsin.com> | 2025-07-03 07:27:52 +0000 |
|---|---|---|
| committer | Christian Krinitsin <mail@krinitsin.com> | 2025-07-03 07:27:52 +0000 |
| commit | d0c85e36e4de67af628d54e9ab577cc3fad7796a (patch) | |
| tree | f8f784b0f04343b90516a338d6df81df3a85dfa2 /results/classifier/gemma3:12b/assembly/1841990 | |
| parent | 7f4364274750eb8cb39a3e7493132fca1c01232e (diff) | |
| download | emulator-bug-study-d0c85e36e4de67af628d54e9ab577cc3fad7796a.tar.gz emulator-bug-study-d0c85e36e4de67af628d54e9ab577cc3fad7796a.zip | |
add deepseek and gemma results
Diffstat (limited to 'results/classifier/gemma3:12b/assembly/1841990')
| -rw-r--r-- | results/classifier/gemma3:12b/assembly/1841990 | 39 |
1 files changed, 39 insertions, 0 deletions
diff --git a/results/classifier/gemma3:12b/assembly/1841990 b/results/classifier/gemma3:12b/assembly/1841990 new file mode 100644 index 00000000..edb7c3ce --- /dev/null +++ b/results/classifier/gemma3:12b/assembly/1841990 @@ -0,0 +1,39 @@ + +instruction 'denbcdq' misbehaving + +Instruction 'denbcdq' appears to have no effect. Test case attached. + +On ppc64le native: +-- +gcc -g -O -mcpu=power9 bcdcfsq.c test-denbcdq.c -o test-denbcdq +$ ./test-denbcdq +0x00000000000000000000000000000000 +0x0000000000000000000000000000000c +0x22080000000000000000000000000000 +$ ./test-denbcdq 1 +0x00000000000000000000000000000001 +0x0000000000000000000000000000001c +0x22080000000000000000000000000001 +$ ./test-denbcdq $(seq 0 99) +0x00000000000000000000000000000064 +0x0000000000000000000000000000100c +0x22080000000000000000000000000080 +-- + +With "qemu-ppc64le -cpu power9" +-- +$ qemu-ppc64le -cpu power9 -L [...] ./test-denbcdq +0x00000000000000000000000000000000 +0x0000000000000000000000000000000c +0x0000000000000000000000000000000c +$ qemu-ppc64le -cpu power9 -L [...] ./test-denbcdq 1 +0x00000000000000000000000000000001 +0x0000000000000000000000000000001c +0x0000000000000000000000000000001c +$ qemu-ppc64le -cpu power9 -L [...] ./test-denbcdq $(seq 100) +0x00000000000000000000000000000064 +0x0000000000000000000000000000100c +0x0000000000000000000000000000100c +-- + +I started looking at the code, but I got confused rather quickly. Could be related to endianness? I think denbcdq arrived on the scene before little-endian was a big deal. Maybe something to do with utilizing implicit floating-point register pairs... I don't think the right data is getting to helper_denbcdq, which would point back to the gen_fprp_ptr uses in dfp-impl.inc.c (GEN_DFP_T_FPR_I32_Rc). (Maybe?) \ No newline at end of file |