Better distinguish between platform vendor and device vendor

pull/352/merge
jsteube 8 years ago
parent be91174794
commit f5f9073f45

@ -995,13 +995,15 @@ struct __hc_device_param
// device specific attributes starting
char *device_name;
char *device_vendor;
char *device_name_chksum;
char *device_version;
char *driver_version;
bool opencl_v12;
cl_uint vendor_id;
cl_uint device_vendor_id;
cl_uint platform_vendor_id;
cl_kernel kernel1;
cl_kernel kernel12;

@ -1570,11 +1570,11 @@ void status_display ()
hm_device_val_to_str ((char *) utilization, HM_STR_BUF_SIZE, "%", hm_get_utilization_with_device_id (device_id));
hm_device_val_to_str ((char *) temperature, HM_STR_BUF_SIZE, "c", hm_get_temperature_with_device_id (device_id));
if (device_param->vendor_id == VENDOR_ID_AMD)
if (device_param->device_vendor_id == VENDOR_ID_AMD)
{
hm_device_val_to_str ((char *) fanspeed, HM_STR_BUF_SIZE, "%", hm_get_fanspeed_with_device_id (device_id));
}
else if (device_param->vendor_id == VENDOR_ID_NV)
else if (device_param->device_vendor_id == VENDOR_ID_NV)
{
hm_device_val_to_str ((char *) fanspeed, HM_STR_BUF_SIZE, "%", hm_get_fanspeed_with_device_id (device_id));
}
@ -2613,7 +2613,7 @@ static void run_kernel_bzero (hc_device_param_t *device_param, cl_mem buf, const
{
int rc = -1;
if (device_param->opencl_v12 && device_param->vendor_id == VENDOR_ID_AMD)
if (device_param->opencl_v12 && device_param->platform_vendor_id == VENDOR_ID_AMD)
{
// So far tested, amd is the only supporting this OpenCL 1.2 function without segfaulting
@ -13188,34 +13188,6 @@ int main (int argc, char **argv)
}
}
/**
* Detect if the user has both an Intel and and AMD OpenCL runtime and not filtered any of them
* In this case mark it for later use
*/
int has_intel = 0;
int has_amd = 0;
for (uint platform_id = 0; platform_id < platforms_cnt; platform_id++)
{
if ((opencl_platforms_filter & (1 << platform_id)) == 0) continue;
cl_platform_id platform = platforms[platform_id];
char platform_vendor[INFOSZ] = { 0 };
hc_clGetPlatformInfo (data.ocl, platform, CL_PLATFORM_VENDOR, sizeof (platform_vendor), platform_vendor, NULL);
if (strcmp (platform_vendor, CL_VENDOR_AMD) == 0)
{
has_intel = 1;
}
else if (strcmp (platform_vendor, CL_VENDOR_INTEL_SDK) == 0)
{
has_amd = 1;
}
}
/**
* OpenCL devices: simply push all devices from all platforms into the same device array
*/
@ -13244,39 +13216,39 @@ int main (int argc, char **argv)
// this causes trouble with vendor id based macros
// we'll assign generic to those without special optimization available
cl_uint vendor_id = 0;
cl_uint platform_vendor_id = 0;
if (strcmp (platform_vendor, CL_VENDOR_AMD) == 0)
{
vendor_id = VENDOR_ID_AMD;
platform_vendor_id = VENDOR_ID_AMD;
}
else if (strcmp (platform_vendor, CL_VENDOR_APPLE) == 0)
{
vendor_id = VENDOR_ID_APPLE;
platform_vendor_id = VENDOR_ID_APPLE;
}
else if (strcmp (platform_vendor, CL_VENDOR_INTEL_BEIGNET) == 0)
{
vendor_id = VENDOR_ID_INTEL_BEIGNET;
platform_vendor_id = VENDOR_ID_INTEL_BEIGNET;
}
else if (strcmp (platform_vendor, CL_VENDOR_INTEL_SDK) == 0)
{
vendor_id = VENDOR_ID_INTEL_SDK;
platform_vendor_id = VENDOR_ID_INTEL_SDK;
}
else if (strcmp (platform_vendor, CL_VENDOR_MESA) == 0)
{
vendor_id = VENDOR_ID_MESA;
platform_vendor_id = VENDOR_ID_MESA;
}
else if (strcmp (platform_vendor, CL_VENDOR_NV) == 0)
{
vendor_id = VENDOR_ID_NV;
platform_vendor_id = VENDOR_ID_NV;
}
else if (strcmp (platform_vendor, CL_VENDOR_POCL) == 0)
{
vendor_id = VENDOR_ID_POCL;
platform_vendor_id = VENDOR_ID_POCL;
}
else
{
vendor_id = VENDOR_ID_GENERIC;
platform_vendor_id = VENDOR_ID_GENERIC;
}
for (uint platform_devices_id = 0; platform_devices_id < platform_devices_cnt; platform_devices_id++)
@ -13287,7 +13259,7 @@ int main (int argc, char **argv)
hc_device_param_t *device_param = &data.devices_param[device_id];
device_param->vendor_id = vendor_id;
device_param->platform_vendor_id = platform_vendor_id;
device_param->device = platform_devices[platform_devices_id];
@ -13315,6 +13287,16 @@ int main (int argc, char **argv)
device_param->device_name = device_name;
// device_vendor
hc_clGetDeviceInfo (data.ocl, device_param->device, CL_DEVICE_VENDOR, 0, NULL, &param_value_size);
char *device_vendor = (char *) mymalloc (param_value_size);
hc_clGetDeviceInfo (data.ocl, device_param->device, CL_DEVICE_VENDOR, param_value_size, device_vendor, NULL);
device_param->device_vendor = device_vendor;
// tuning db
tuning_db_entry_t *tuningdb_entry = tuning_db_search (tuning_db, device_param, attack_mode, hash_mode);
@ -13504,16 +13486,16 @@ int main (int argc, char **argv)
device_param->skipped = 1;
}
// if we have both intel and amd opencl runtime we want they share the same cpu
// so both virtual cpu share the same resources and run at 50%
// however, intel has better control over their own hardware so it makes sense
// if we have an Intel CPU and an not Intel OpenCL runtime that's a tricky situation
// both platforms will support CPU device types and therefore both will use 50% of the physical resources
// however, intel has better simd control over their own hardware so it makes sense
// to give them full control over their own hardware
if (device_type & CL_DEVICE_TYPE_CPU)
{
if ((has_intel == 1) && (has_amd == 1) && (vendor_id == VENDOR_ID_AMD))
if ((device_param->device_vendor_id == VENDOR_ID_INTEL_SDK) && (device_param->platform_vendor_id != VENDOR_ID_INTEL_SDK))
{
if (data.quiet == 0) log_info ("Device #%u: WARNING: Not its native platform vendor", device_id + 1);
if (data.quiet == 0) log_info ("Device #%u: WARNING: not native intel opencl platform", device_id + 1);
device_param->skipped = 1;
}
@ -13539,9 +13521,9 @@ int main (int argc, char **argv)
char *device_name_chksum = (char *) mymalloc (INFOSZ);
#if __x86_64__
snprintf (device_name_chksum, INFOSZ - 1, "%u-%u-%u-%s-%s-%s-%u", 64, device_param->vendor_id, device_param->vector_width, device_param->device_name, device_param->device_version, device_param->driver_version, COMPTIME);
snprintf (device_name_chksum, INFOSZ - 1, "%u-%u-%u-%s-%s-%s-%u", 64, device_param->platform_vendor_id, device_param->vector_width, device_param->device_name, device_param->device_version, device_param->driver_version, COMPTIME);
#else
snprintf (device_name_chksum, INFOSZ - 1, "%u-%u-%u-%s-%s-%s-%u", 32, device_param->vendor_id, device_param->vector_width, device_param->device_name, device_param->device_version, device_param->driver_version, COMPTIME);
snprintf (device_name_chksum, INFOSZ - 1, "%u-%u-%u-%s-%s-%s-%u", 32, device_param->platform_vendor_id, device_param->vector_width, device_param->device_name, device_param->device_version, device_param->driver_version, COMPTIME);
#endif
uint device_name_digest[4] = { 0 };
@ -13563,7 +13545,7 @@ int main (int argc, char **argv)
if (device_type & CL_DEVICE_TYPE_GPU)
{
if (vendor_id == VENDOR_ID_AMD)
if (platform_vendor_id == VENDOR_ID_AMD)
{
cl_uint device_processor_cores = 0;
@ -13573,7 +13555,7 @@ int main (int argc, char **argv)
device_param->device_processor_cores = device_processor_cores;
}
else if (vendor_id == VENDOR_ID_NV)
else if (platform_vendor_id == VENDOR_ID_NV)
{
cl_uint kernel_exec_timeout = 0;
@ -13642,7 +13624,7 @@ int main (int argc, char **argv)
{
if (device_type & CL_DEVICE_TYPE_GPU)
{
if (vendor_id == VENDOR_ID_AMD)
if (platform_vendor_id == VENDOR_ID_AMD)
{
int catalyst_check = (force == 1) ? 0 : 1;
@ -13690,7 +13672,7 @@ int main (int argc, char **argv)
return (-1);
}
}
else if (vendor_id == VENDOR_ID_NV)
else if (platform_vendor_id == VENDOR_ID_NV)
{
if (device_param->kernel_exec_timeout != 0)
{
@ -13703,7 +13685,7 @@ int main (int argc, char **argv)
/* turns out pocl still creates segfaults (because of llvm)
if (device_type & CL_DEVICE_TYPE_CPU)
{
if (vendor_id == VENDOR_ID_AMD)
if (platform_vendor_id == VENDOR_ID_AMD)
{
if (force == 0)
{
@ -14083,14 +14065,14 @@ int main (int argc, char **argv)
const uint platform_devices_id = device_param->platform_devices_id;
#if defined(HAVE_NVML) || defined(HAVE_NVAPI)
if (device_param->vendor_id == VENDOR_ID_NV)
if (device_param->platform_vendor_id == VENDOR_ID_NV)
{
memcpy (&data.hm_device[device_id], &hm_adapters_nv[platform_devices_id], sizeof (hm_attrs_t));
}
#endif
#ifdef HAVE_ADL
if (device_param->vendor_id == VENDOR_ID_AMD)
if (device_param->platform_vendor_id == VENDOR_ID_AMD)
{
memcpy (&data.hm_device[device_id], &hm_adapters_amd[platform_devices_id], sizeof (hm_attrs_t));
}
@ -14278,22 +14260,22 @@ int main (int argc, char **argv)
if (hash_mode == 8900)
{
if (device_param->vendor_id == VENDOR_ID_AMD)
if (device_param->platform_vendor_id == VENDOR_ID_AMD)
{
tmto_start = 1;
}
else if (device_param->vendor_id == VENDOR_ID_NV)
else if (device_param->platform_vendor_id == VENDOR_ID_NV)
{
tmto_start = 2;
}
}
else if (hash_mode == 9300)
{
if (device_param->vendor_id == VENDOR_ID_AMD)
if (device_param->platform_vendor_id == VENDOR_ID_AMD)
{
tmto_start = 2;
}
else if (device_param->vendor_id == VENDOR_ID_NV)
else if (device_param->platform_vendor_id == VENDOR_ID_NV)
{
tmto_start = 2;
}
@ -14621,12 +14603,12 @@ int main (int argc, char **argv)
char build_opts_new[1024] = { 0 };
snprintf (build_opts_new, sizeof (build_opts_new) - 1, "%s -DVENDOR_ID=%u -DCUDA_ARCH=%d -DVECT_SIZE=%u -DDEVICE_TYPE=%u -DKERN_TYPE=%u -D_unroll -cl-std=CL1.1", build_opts, device_param->vendor_id, (device_param->sm_major * 100) + device_param->sm_minor, device_param->vector_width, (u32) device_param->device_type, kern_type);
snprintf (build_opts_new, sizeof (build_opts_new) - 1, "%s -DVENDOR_ID=%u -DCUDA_ARCH=%d -DVECT_SIZE=%u -DDEVICE_TYPE=%u -DKERN_TYPE=%u -D_unroll -cl-std=CL1.1", build_opts, device_param->platform_vendor_id, (device_param->sm_major * 100) + device_param->sm_minor, device_param->vector_width, (u32) device_param->device_type, kern_type);
strncpy (build_opts, build_opts_new, sizeof (build_opts) - 1);
if (device_param->vendor_id == VENDOR_ID_INTEL_SDK)
if (device_param->platform_vendor_id == VENDOR_ID_INTEL_SDK)
{
// we do vectorizing much better than the auto-vectorizer

@ -3062,7 +3062,7 @@ int hm_get_temperature_with_device_id (const uint device_id)
if ((data.devices_param[device_id].device_type & CL_DEVICE_TYPE_GPU) == 0) return -1;
#ifdef HAVE_ADL
if (data.devices_param[device_id].vendor_id == VENDOR_ID_AMD)
if (data.devices_param[device_id].device_vendor_id == VENDOR_ID_AMD)
{
if (data.hm_amd)
{
@ -3089,7 +3089,7 @@ int hm_get_temperature_with_device_id (const uint device_id)
#endif
#if defined(HAVE_NVML) || defined(HAVE_NVAPI)
if (data.devices_param[device_id].vendor_id == VENDOR_ID_NV)
if (data.devices_param[device_id].device_vendor_id == VENDOR_ID_NV)
{
#if defined(LINUX) && defined(HAVE_NVML)
int temperature = 0;
@ -3125,7 +3125,7 @@ int hm_get_fanspeed_with_device_id (const uint device_id)
if (data.hm_device[device_id].fan_supported == 1)
{
#ifdef HAVE_ADL
if (data.devices_param[device_id].vendor_id == VENDOR_ID_AMD)
if (data.devices_param[device_id].device_vendor_id == VENDOR_ID_AMD)
{
if (data.hm_amd)
{
@ -3158,7 +3158,7 @@ int hm_get_fanspeed_with_device_id (const uint device_id)
#endif // HAVE_ADL
#if defined(HAVE_NVML) || defined(HAVE_NVAPI)
if (data.devices_param[device_id].vendor_id == VENDOR_ID_NV)
if (data.devices_param[device_id].device_vendor_id == VENDOR_ID_NV)
{
#if defined(LINUX) && defined(HAVE_NVML)
int speed = 0;
@ -3190,7 +3190,7 @@ int hm_get_utilization_with_device_id (const uint device_id)
if ((data.devices_param[device_id].device_type & CL_DEVICE_TYPE_GPU) == 0) return -1;
#ifdef HAVE_ADL
if (data.devices_param[device_id].vendor_id == VENDOR_ID_AMD)
if (data.devices_param[device_id].device_vendor_id == VENDOR_ID_AMD)
{
if (data.hm_amd)
{
@ -3206,7 +3206,7 @@ int hm_get_utilization_with_device_id (const uint device_id)
#endif // HAVE_ADL
#if defined(HAVE_NVML) || defined(HAVE_NVAPI)
if (data.devices_param[device_id].vendor_id == VENDOR_ID_NV)
if (data.devices_param[device_id].device_vendor_id == VENDOR_ID_NV)
{
#if defined(LINUX) && defined(HAVE_NVML)
nvmlUtilization_t utilization;

Loading…
Cancel
Save