x64: conv: avoid overflows and add limit for huge spatial sizes (backport)

[tczeszun]

backport of #2627

(base) gta@DUT7314PVC:~/tczeszun$ LD_LIBRARY_PATH=. DNNL_VERBOSE=2 ./benchdnn  --conv --engine=gpu --dir=BWD_W --dt=bf16:bf16:bf16 --stag=axb --wtag=xba --dtag=axb --impl=jit -v5 mb1ic32iw134217732oc1ow134217730kw3pw0
create: --conv --engine=gpu --dir=BWD_W --dt=bf16:bf16:bf16 --stag=axb --wtag=xba --dtag=axb --impl=jit mb1ic32iw134217732oc1ow134217730kw3pw0
onednn_verbose,v1,info,oneDNN v3.7.0 (commit 10df79807cf134982920bbbdb02a0dab666d4ed3)
onednn_verbose,v1,info,cpu,runtime:OpenMP,nthr:208
onednn_verbose,v1,info,cpu,isa:Intel AVX-512 with float16, Intel DL Boost and bfloat16 support
onednn_verbose,v1,info,gpu,runtime:OpenCL
onednn_verbose,v1,info,gpu,engine,opencl device count:2
onednn_verbose,v1,info,gpu,engine,0,name:Intel(R) Data Center GPU Max 1550,driver_version:24.26.30049,binary_kernels:enabled
onednn_verbose,v1,info,gpu,engine,1,name:Intel(R) Data Center GPU Max 1550,driver_version:24.26.30049,binary_kernels:enabled
onednn_verbose,v1,info,graph,backend,0:dnnl_backend
onednn_verbose,v1,primitive,info,template:operation,engine,primitive,implementation,prop_kind,memory_descriptors,attributes,auxiliary,problem_desc,exec_time
onednn_verbose,v1,graph,info,template:operation,engine,partition_id,partition_kind,op_names,data_formats,logical_tensors,fpmath_mode,implementation,backend,exec_time
onednn_verbose,v1,primitive,create:cache_miss,gpu:0,convolution,jit:ir,backward_weights,src:bf16::blocked:acb::f0 wei:bf16::blocked:cba::f0 bia:undef::undef::: dst:bf16::blocked:acb::f0,,alg:convolution_direct,mb1_ic32oc1_iw134217732ow134217730kw3sw1dw0pw0,248.712
oneDNN implementation: jit:ir
[IMPL_FILTER] All implementations were skipped!
[IMPL_FILTER] All implementations were skipped!
onednn_verbose,v1,primitive,create:persistent_cache_hit,gpu:0,convolution,jit:ir,backward_weights,src:bf16::blocked:acb::f0 wei:bf16::blocked:cba::f0 bia:undef::undef::: dst:bf16::blocked:acb::f0,,alg:convolution_direct,mb1_ic32oc1_iw134217732ow134217730kw3sw1dw0pw0,24.6731
run: --conv --engine=gpu --dir=BWD_W --dt=bf16:bf16:bf16 --stag=axb --wtag=xba --dtag=axb --impl=jit mb1ic32iw134217732oc1ow134217730kw3pw0
onednn_verbose,v1,primitive,create:cache_miss,cpu,reorder,jit:uni,undef,src:f32::blocked:abc::f0 dst:bf16::blocked:acb::f0,,,1x1x134217730,0.334961
onednn_verbose,v1,primitive,exec,cpu,reorder,jit:uni,undef,src:f32::blocked:abc::f0 dst:bf16::blocked:acb::f0,,,1x1x134217730,40.6531
onednn_verbose,v1,primitive,create:cache_miss,cpu,reorder,simple:any,undef,src:f32::blocked:abc::f0 dst:bf16::blocked:acb::f0,,,1x32x134217732,0.106934
onednn_verbose,v1,primitive,exec,cpu,reorder,simple:any,undef,src:f32::blocked:abc::f0 dst:bf16::blocked:acb::f0,,,1x32x134217732,16760.3
onednn_verbose,v1,primitive,exec,gpu:0,convolution,jit:ir,backward_weights,src:bf16::blocked:acb::f0 wei:bf16::blocked:cba::f0 bia:undef::undef::: dst:bf16::blocked:acb::f0,,alg:convolution_direct,mb1_ic32oc1_iw134217732ow134217730kw3sw1dw0pw0,37.7659
onednn_verbose,v1,primitive,create:cache_miss,cpu,reorder,jit:uni,undef,src:bf16::blocked:cba::f0 dst:f32::blocked:abc::f0,,,1x32x3,0.313965
onednn_verbose,v1,primitive,exec,cpu,reorder,jit:uni,undef,src:bf16::blocked:cba::f0 dst:f32::blocked:abc::f0,,,1x32x3,0.019043
0:PASSED __REPRO: --conv --engine=gpu --dir=BWD_W --dt=bf16:bf16:bf16 --stag=axb --wtag=xba --dtag=axb --impl=jit mb1ic32iw134217732oc1ow134217730kw3pw0
tests:1 passed:1 skipped:0 mistrusted:0 unimplemented:0 invalid_arguments:0 failed:0 listed:0
total: 128.80s; fill: 36.24s (28%); compute_ref: 86.90s (67%); compare: 0.01s (0%);

[densamoilov]

I ran ci and nightly test set for this PR, the failures in conv look unrelated to this particular issue.