[Bug testsuite/108190] New: Various testsuite asm fails

[Bug testsuite/108190] New: Various testsuite asm fails

[nightstrike at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108190

            Bug ID: 108190
           Summary: Various testsuite asm fails
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: testsuite
          Assignee: unassigned at gcc dot gnu.org
          Reporter: nightstrike at gmail dot com
  Target Milestone: ---

Created attachment 54134
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54134&action=edit
asm output 1

The following are a collection of failures from running PR54700 tests on a
linux to mingw cross compiler under Wine 7.20:

FAIL: g++.target/i386/sse4_1-pr54700-1.C   scan-assembler-not pcmpgt[bdq]
FAIL: g++.target/i386/sse4_1-pr54700-1.C   scan-assembler-times pblendvb 2
FAIL: g++.target/i386/sse4_1-pr54700-1.C   scan-assembler-times blendvps 4
FAIL: g++.target/i386/sse4_1-pr54700-1.C   scan-assembler-times blendvpd 4
FAIL: g++.target/i386/avx-pr54700-1.C   scan-assembler-not vpcmpgt[bdq]
FAIL: g++.target/i386/avx-pr54700-1.C   scan-assembler-times vpblendvb 2
FAIL: g++.target/i386/avx-pr54700-1.C   scan-assembler-times vblendvps 4
FAIL: g++.target/i386/avx-pr54700-1.C   scan-assembler-times vblendvpd 4
FAIL: g++.target/i386/avx2-pr54700-1.C   scan-assembler-not vpcmpgt[bdq]
FAIL: g++.target/i386/avx2-pr54700-1.C   scan-assembler-times vpblendvb 2
FAIL: g++.target/i386/avx2-pr54700-1.C   scan-assembler-times vblendvps 4
FAIL: g++.target/i386/avx2-pr54700-1.C   scan-assembler-times vblendvpd 4
FAIL: g++.target/i386/avx512vl-pr54700-1a.C   scan-assembler-times vpblendvb 2
FAIL: g++.target/i386/avx512vl-pr54700-1b.C   scan-assembler-times pblendvb 2


I don't know if it's helpful, but these are the /proc/cpuinfo flags on the cpus
running under wine on an Intel(R) Xeon(R) Gold 6230:

flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb
rdtscp lm constant_tsc art arch
_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni
pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm
pcid dca sse4_1 sse4_2 x2apic movbe
 popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch
cpuid_fault epb cat_l3 cdp_l3 invpcid_single ssbd mba ibrs ibpb stibp
ibrs_enhanced fsgsbase tsc_adjust bmi1
hle avx2 smep bmi2 erms invpcid cqm mpx rdt_a avx512f avx512dq rdseed adx smap
clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1
xsaves cqm_llc cqm_occup_llc cqm_
mbm_total cqm_mbm_local dtherm ida arat pln pts hwp hwp_act_window hwp_epp
hwp_pkg_req pku ospke avx512_vnni md_clear flush_l1d arch_capabilities

[Bug testsuite/108190] Various testsuite asm fails

[nightstrike at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108190

--- Comment #1 from nightstrike <nightstrike at gmail dot com> ---
Created attachment 54135
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54135&action=edit
avx-pr54700-1.s

[Bug testsuite/108190] g++.target/i386/*pr54700*.C testcases fail on x86_64-mingw

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108190

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
This is most likely because of the ABI difference between windows and elf
targets. __v16qi is passed via memory location on windows but on elf targets it
is passed via regsiter ...

[Bug testsuite/108190] g++.target/i386/*pr54700*.C testcases fail on x86_64-mingw

[nightstrike at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108190

--- Comment #3 from nightstrike <nightstrike at gmail dot com> ---
Created attachment 54136
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54136&action=edit
avx2-pr54700-1.s

[Bug testsuite/108190] g++.target/i386/*pr54700*.C testcases fail on x86_64-mingw

[nightstrike at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108190

--- Comment #4 from nightstrike <nightstrike at gmail dot com> ---
Created attachment 54137
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54137&action=edit
avx512vl-pr54700-1a.s

[Bug testsuite/108190] g++.target/i386/*pr54700*.C testcases fail on x86_64-mingw

[nightstrike at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108190

--- Comment #5 from nightstrike <nightstrike at gmail dot com> ---
Created attachment 54138
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54138&action=edit
avx512vl-pr54700-1b.s

[Bug testsuite/108190] g++.target/i386/*pr54700*.C testcases fail on x86_64-mingw

[glisse at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108190

--- Comment #6 from Marc Glisse <glisse at gcc dot gnu.org> ---
Indeed, this looks like a common issue (at least with the x86 backend): the
memory load is combined with the comparison before we try combining the
comparison with the blend, and this last combination is then rejected because
it expects a register, not memory. So either we are too eager in merging loads
with instructions, or we reject instructions too early when we could still fix
the operands with an extra load.