Skip to content

ARM64 DYNAMIC_ARCH: add CortexA75/76 and restore VORTEX for DYNAMIC_LIST #5719

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
martin-frbg merged 1 commit into from
Mar 30, 2026
+19 −13
Merged

ARM64 DYNAMIC_ARCH: add CortexA75/76 and restore VORTEX for DYNAMIC_LIST #5719

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 19 additions & 13 deletions driver/others/dynamic_arm64.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,13 +38,7 @@
/*********************************************************************/

#include "common.h"

#ifndef _MSC_VER
#include <strings.h>
#else
#define strncasecmp _strnicmp
#endif

#if (defined OS_LINUX || defined OS_ANDROID)
#include <asm/hwcap.h>
#include <sys/auxv.h>
Expand Down Expand Up @@ -135,10 +129,16 @@ extern gotoblas_t gotoblas_ARMV9SME;
#else
#define gotoblas_ARMV9SME gotoblas_ARMV8
#endif
#ifdef DYN_VORTEX
extern gotoblas_t gotoblas_VORTEX;
#elif defined(DYN_NEOVERSEN1)
#define gotoblas_VORTEX gotoblas_NEOVERSEN1
#else
#define gotoblas_VORTEX gotoblas_ARMV8
#endif
#ifdef DYN_VORTEXM4
extern gotoblas_t gotoblas_VORTEXM4;
#else
#error "dont have vortexm4"
#define gotoblas_VORTEXM4 gotoblas_ARMV8
#endif
#ifdef DYN_CORTEXA55
Expand All @@ -151,7 +151,7 @@ extern gotoblas_t gotoblas_A64FX;
#else
#define gotoblas_A64FX gotoblas_ARMV8
#endif
#else
#else //not a user-specified dynamic_list
extern gotoblas_t gotoblas_CORTEXA53;
#define gotoblas_CORTEXA55 gotoblas_CORTEXA53
extern gotoblas_t gotoblas_CORTEXA57;
Expand All @@ -163,6 +163,7 @@ extern gotoblas_t gotoblas_THUNDERX2T99;
extern gotoblas_t gotoblas_TSV110;
extern gotoblas_t gotoblas_EMAG8180;
extern gotoblas_t gotoblas_NEOVERSEN1;
#define gotoblas_VORTEX gotoblas_NEOVERSEN1
#ifndef NO_SVE
extern gotoblas_t gotoblas_NEOVERSEV1;
extern gotoblas_t gotoblas_NEOVERSEN2;
Expand Down Expand Up @@ -198,7 +199,7 @@ extern void openblas_warning(int verbose, const char * msg);
#define FALLBACK_VERBOSE 1
#define NEOVERSEN1_FALLBACK "OpenBLAS : Your OS does not support SVE instructions. OpenBLAS is using Neoverse N1 kernels as a fallback, which may give poorer performance.\n"

#define NUM_CORETYPES 20
#define NUM_CORETYPES 21

/*
* In case asm/hwcap.h is outdated on the build system, make sure
Expand Down Expand Up @@ -238,6 +239,7 @@ static char *corename[] = {
"armv8sve",
"a64fx",
"armv9sme",
"vortex",
"vortexm4",
"unknown"
};
Expand All @@ -262,7 +264,8 @@ char *gotoblas_corename(void) {
if (gotoblas == &gotoblas_ARMV8SVE) return corename[16];
if (gotoblas == &gotoblas_A64FX) return corename[17];
if (gotoblas == &gotoblas_ARMV9SME) return corename[18];
if (gotoblas == &gotoblas_VORTEXM4) return corename[19];
if (gotoblas == &gotoblas_VORTEX) return corename[19];
if (gotoblas == &gotoblas_VORTEXM4) return corename[20];
return corename[NUM_CORETYPES];
}

Expand Down Expand Up @@ -301,7 +304,8 @@ static gotoblas_t *force_coretype(char *coretype) {
case 16: return (&gotoblas_ARMV8SVE);
case 17: return (&gotoblas_A64FX);
case 18: return (&gotoblas_ARMV9SME);
case 19: return (&gotoblas_VORTEXM4);
case 19: return (&gotoblas_VORTEX);
case 20: return (&gotoblas_VORTEXM4);
}
snprintf(message, 128, "Core not found: %s\n", coretype);
openblas_warning(1, message);
Expand All @@ -318,7 +322,7 @@ static gotoblas_t *get_coretype(void) {
return &gotoblas_VORTEXM4;
}
#endif
return &gotoblas_NEOVERSEN1;
return &gotoblas_VORTEX;
#endif

#if (!defined OS_LINUX && !defined OS_ANDROID)
Expand Down Expand Up @@ -403,6 +407,8 @@ static gotoblas_t *get_coretype(void) {
case 0xd08: // Cortex A72
return &gotoblas_CORTEXA72;
case 0xd09: // Cortex A73
case 0xd0a: // Cortex A75
case 0xd0b: // Cortex A76
return &gotoblas_CORTEXA73;
case 0xd0c: // Neoverse N1
return &gotoblas_NEOVERSEN1;
Expand Down Expand Up @@ -492,7 +498,7 @@ static gotoblas_t *get_coretype(void) {
break;
case 0x61: // Apple
if (support_sme1()) return &gotoblas_VORTEXM4;
return &gotoblas_NEOVERSEN1;
return &gotoblas_VORTEX;
break;
default:
snprintf(coremsg, 128, "Unknown CPU model - implementer %x part %x\n",implementer,part);
Expand Down
Loading