SPEC CPU(R)2026 Floating Point Rate Result Maginfra Co., Ltd. QR8228-D3 (AMD EPYC 9335) CPU2026 License: 9087 Test date: Jul-2026 Test sponsor: Maginfra Co., Ltd. Hardware availability: Dec-2025 Tested by: Maginfra Co., Ltd. Software availability: Jan-2026 Base Base Base Peak Peak Peak Benchmarks Copies Run Time Rate Copies Run Time Rate ---------------- ------- --------- --------- ------- --------- --------- 709.cactus_r 128 98.7 1110 * 128 98.4 1120 S 709.cactus_r 128 99.1 1110 S 128 98.6 1110 * 709.cactus_r 128 98.4 1120 S 128 99.0 1110 S 722.palm_r 128 369 458 S 128 369 458 * 722.palm_r 128 369 458 * 128 368 459 S 722.palm_r 128 371 456 S 128 373 453 S 731.astcenc_r 128 205 523 * 128 205 524 S 731.astcenc_r 128 206 523 S 128 205 524 * 731.astcenc_r 128 205 524 S 128 205 523 S 736.ocio_r 128 229 489 S 128 230 488 S 736.ocio_r 128 229 488 * 128 229 490 * 736.ocio_r 128 229 488 S 128 228 490 S 737.gmsh_r 128 180 326 S 128 181 325 S 737.gmsh_r 128 180 326 * 128 180 326 * 737.gmsh_r 128 181 325 S 128 180 326 S 748.flightdm_r 128 159 577 S 128 159 577 S 748.flightdm_r 128 160 573 * 128 160 573 * 748.flightdm_r 128 160 573 S 128 160 573 S 749.fotonik3d_r 128 579 256 S 64 290 255 S 749.fotonik3d_r 128 580 255 S 64 289 256 S 749.fotonik3d_r 128 580 255 * 64 290 255 * 765.roms_r 128 681 296 S 128 674 299 S 765.roms_r 128 680 296 * 128 675 299 S 765.roms_r 128 680 297 S 128 674 299 * 766.femflow_r 128 175 1080 * 128 175 1080 * 766.femflow_r 128 175 1080 S 128 175 1080 S 766.femflow_r 128 176 1070 S 128 176 1070 S 767.nest_r 128 225 451 S 128 225 451 S 767.nest_r 128 224 453 * 128 224 453 * 767.nest_r 128 224 454 S 128 224 454 S 772.marian_r 128 173 1170 * 128 173 1170 * 772.marian_r 128 171 1180 S 128 171 1180 S 772.marian_r 128 176 1150 S 128 176 1150 S 782.lbm_r 128 245 300 S 128 242 304 S 782.lbm_r 128 245 300 * 128 245 300 * 782.lbm_r 128 244 301 S 128 246 299 S ================================================================================== 709.cactus_r 128 98.7 1110 * 128 98.6 1110 * 722.palm_r 128 369 458 * 128 369 458 * 731.astcenc_r 128 205 523 * 128 205 524 * 736.ocio_r 128 229 488 * 128 229 490 * 737.gmsh_r 128 180 326 * 128 180 326 * 748.flightdm_r 128 160 573 * 128 160 573 * 749.fotonik3d_r 128 580 255 * 64 290 255 * 765.roms_r 128 680 296 * 128 674 299 * 766.femflow_r 128 175 1080 * 128 175 1080 * 767.nest_r 128 224 453 * 128 224 453 * 772.marian_r 128 173 1170 * 128 173 1170 * 782.lbm_r 128 245 300 * 128 245 300 * SPECrate(R)2026_fp_base 511 SPECrate(R)2026_fp_peak 511 HARDWARE -------- CPU Name: AMD EPYC 9335 Max MHz: 4400 Nominal: 3000 Enabled: 64 cores, 2 chips, 2 threads/core Orderable: 1,2 chips Cache L1: 32 KB I + 48 KB D on chip per core L2: 1 MB I+D on chip per core L3: 128 MB I+D on chip per chip, 32 MB shared / 8 cores Other: None Memory: 1536 GB (24 x 64 GB 2Rx4 PC5-6400B-R) Storage: 1 x 4 TB NVME SSD Cooling: Air Other: None SOFTWARE -------- OS: SUSE Linux Enterprise Server 15 SP7 6.4.0-150700.51-default Compiler: C/C++/Fortran: Version 5.1.0 of AOCC Compiler Category: Vendor Firmware: Version 01.14.00 released Dec-2025 File System: xfs System State: Run level 3 (multi-user) Base Pointers: 64-bit Peak Pointers: 64-bit Other: None Power Management: BIOS and OS set to prefer performance at the cost of additional power usage. Compiler Notes -------------- The AMD64 AOCC Compiler Suite is available at http://developer.amd.com/amd-aocc/ Submit Notes ------------ The config file option 'submit' was used. 'numactl' was used to bind copies to the cores. See the configuration file for details. Operating System Notes ---------------------- 'ulimit -s unlimited' was used to set environment stack size limit 'ulimit -l 2097152' was used to set environment locked pages in memory limit runcpu command invoked through numactl i.e.: numactl --interleave=all runcpu To limit dirty cache to 8% of memory, 'sysctl -w vm.dirty_ratio=8' run as root. To limit swap usage to minimum necessary, 'sysctl -w vm.swappiness=1' run as root. To free node-local memory and avoid remote memory usage, 'sysctl -w vm.zone_reclaim_mode=1' run as root. To clear filesystem caches, 'sync; sysctl -w vm.drop_caches=3' run as root. To disable address space layout randomization (ASLR) to reduce run-to-run variability, 'sysctl -w kernel.randomize_va_space=0' run as root. To enable Transparent Hugepages (THP) for all allocations, 'echo always > /sys/kernel/mm/transparent_hugepage/enabled' and 'echo always > /sys/kernel/mm/transparent_hugepage/defrag' run as root. Environment Variables Notes --------------------------- Environment variables set by runcpu before the start of the run: LD_LIBRARY_PATH = "/home/CPU2026/amd_rate_aocc510_znver5_B_lib/lib:/home/CPU2026/amd_rate_ aocc510_znver5_B_lib/lib32:" MALLOC_CONF = "retain:true" General Notes ------------- Binaries were compiled on a system with 2x AMD EPYC Venice256 CPU + 2TiB Memory using Ubuntu 24.04 Platform Notes -------------- BIOS configuration: NUMA nodes per socket = NPS4 Determinism Slider = Power cTDP Control = Manual cTDP = 500 Package Power Limit Control = Manual Package Power Limit = 500 Sysinfo program /home/CPU2026/bin/sysinfo Rev: 779ab21020787073335a329f3a45e2cd running on localhost Sat Aug 1 01:12:34 2026 SUT (System Under Test) info as seen by some common utilities. ------------------------------------------------------------ Table of contents ------------------------------------------------------------ 1. uname -srvm 2. w 3. Username 4. ulimit -a 5. sysinfo process ancestry 6. /proc/cpuinfo 7. lscpu 8. numactl --hardware 9. /proc/meminfo 10. who -r 11. Systemd service manager version: systemd 254 (254.24+suse.148.g83b9060b6e) 12. Failed units, from systemctl list-units --state=failed 13. Services, from systemctl list-unit-files 14. Linux kernel boot-time arguments, from /proc/cmdline 15. cpupower frequency-info 16. tuned-adm active 17. sysctl 18. /sys/kernel/mm/transparent_hugepage 19. /sys/kernel/mm/transparent_hugepage/khugepaged 20. OS release 21. Disk information 22. /sys/devices/virtual/dmi/id 23. dmidecode 24. BIOS ------------------------------------------------------------ ------------------------------------------------------------ 1. uname -srvm Linux 6.4.0-150700.51-default #1 SMP PREEMPT_DYNAMIC Wed Apr 30 21:35:43 UTC 2025 (6930611) x86_64 ------------------------------------------------------------ 2. w 01:12:34 up 4:50, 1 user, load average: 68.22, 111.08, 119.63 USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT root tty1 - 20:22 4:49m 7.96s 6.96s python3 ./run_amd_rate_aocc510_znver5_B1.py ------------------------------------------------------------ 3. Username From environment variable $USER: root ------------------------------------------------------------ 4. ulimit -a core file size (blocks, -c) unlimited data seg size (kbytes, -d) unlimited scheduling priority (-e) 0 file size (blocks, -f) unlimited pending signals (-i) 6190465 max locked memory (kbytes, -l) 2048 max memory size (kbytes, -m) unlimited open files (-n) 1024 pipe size (512 bytes, -p) 8 POSIX message queues (bytes, -q) 819200 real-time priority (-r) 0 stack size (kbytes, -s) unlimited cpu time (seconds, -t) unlimited max user processes (-u) 6190465 virtual memory (kbytes, -v) unlimited file locks (-x) unlimited ------------------------------------------------------------ 5. sysinfo process ancestry /usr/lib/systemd/systemd --switched-root --system --deserialize=42 login -- root -bash python3 ./run_amd_rate_aocc510_znver5_B1.py runcpu --config amd_rate_aocc510_znver5_B1.cfg --tune all --reportable --iterations 3 fprate runcpu --configfile amd_rate_aocc510_znver5_B1.cfg --tune all --reportable --iterations 3 --nopower --runmode rate --tune base:peak --size test:train:refrate fprate --nopreenv --note-preenv --logfile $SPEC/tmp/CPU2026.043/templogs/preenv.fprate.043.0.log --lognum 043.0 --from_runcpu 2 specperl $SPEC/bin/sysinfo $SPEC = /home/CPU2026 ------------------------------------------------------------ 6. /proc/cpuinfo model name : AMD EPYC 9335 32-Core Processor vendor_id : AuthenticAMD cpu family : 26 model : 2 stepping : 1 microcode : 0xb00215a bugs : sysret_ss_attrs spectre_v1 spectre_v2 spec_store_bypass srso TLB size : 192 4K pages cpu cores : 32 siblings : 64 2 physical ids (chips) 128 processors (hardware threads) physical id 0: core ids 0-31 physical id 1: core ids 0-31 physical id 0: apicids 0-63 physical id 1: apicids 64-127 Caution: /proc/cpuinfo data regarding chips, cores, and threads is not necessarily reliable, especially for virtualized systems. Use the above data carefully. ------------------------------------------------------------ 7. lscpu From lscpu from util-linux 2.40.4: Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Address sizes: 52 bits physical, 57 bits virtual Byte Order: Little Endian CPU(s): 128 On-line CPU(s) list: 0-127 Vendor ID: AuthenticAMD Model name: AMD EPYC 9335 32-Core Processor CPU family: 26 Model: 2 Thread(s) per core: 2 Core(s) per socket: 32 Socket(s): 2 Stepping: 1 Frequency boost: enabled CPU(s) scaling MHz: 64% CPU max MHz: 3000.0000 CPU min MHz: 1500.0000 BogoMIPS: 5991.17 Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good amd_lbr_v2 nopl nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba perfmon_v2 ibrs ibpb stibp ibrs_enhanced vmmcall fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local user_shstk avx_vnni avx512_bf16 clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin cppc amd_ibpb_ret arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif x2avic v_spec_ctrl vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq la57 rdpid bus_lock_detect movdiri movdir64b overflow_recov succor smca fsrm avx512_vp2intersect flush_l1d debug_swap hv_inuse_wr_allowed srso_user_kernel_no amd_lbr_pmc_freeze Virtualization: AMD-V L1d cache: 3 MiB (64 instances) L1i cache: 2 MiB (64 instances) L2 cache: 64 MiB (64 instances) L3 cache: 256 MiB (8 instances) NUMA node(s): 8 NUMA node0 CPU(s): 0-7,64-71 NUMA node1 CPU(s): 8-15,72-79 NUMA node2 CPU(s): 16-23,80-87 NUMA node3 CPU(s): 24-31,88-95 NUMA node4 CPU(s): 32-39,96-103 NUMA node5 CPU(s): 40-47,104-111 NUMA node6 CPU(s): 48-55,112-119 NUMA node7 CPU(s): 56-63,120-127 Vulnerability Gather data sampling: Not affected Vulnerability Itlb multihit: Not affected Vulnerability L1tf: Not affected Vulnerability Mds: Not affected Vulnerability Meltdown: Not affected Vulnerability Mmio stale data: Not affected Vulnerability Reg file data sampling: Not affected Vulnerability Retbleed: Not affected Vulnerability Spec rstack overflow: Mitigation; IBPB on VMEXIT only Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization Vulnerability Spectre v2: Mitigation; Enhanced / Automatic IBRS; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected Vulnerability Srbds: Not affected Vulnerability Tsx async abort: Not affected From lscpu --cache: NAME ONE-SIZE ALL-SIZE WAYS TYPE LEVEL SETS PHY-LINE COHERENCY-SIZE L1d 48K 3M 12 Data 1 64 1 64 L1i 32K 2M 8 Instruction 1 64 1 64 L2 1M 64M 16 Unified 2 1024 1 64 L3 32M 256M 16 Unified 3 32768 1 64 ------------------------------------------------------------ 8. numactl --hardware NOTE: a numactl 'node' might or might not correspond to a physical chip. available: 8 nodes (0-7) node 0 cpus: 0-7,64-71 node 0 size: 193016 MB node 0 free: 192253 MB node 1 cpus: 8-15,72-79 node 1 size: 193531 MB node 1 free: 192806 MB node 2 cpus: 16-23,80-87 node 2 size: 193531 MB node 2 free: 192812 MB node 3 cpus: 24-31,88-95 node 3 size: 193531 MB node 3 free: 192840 MB node 4 cpus: 32-39,96-103 node 4 size: 193531 MB node 4 free: 192765 MB node 5 cpus: 40-47,104-111 node 5 size: 193531 MB node 5 free: 192792 MB node 6 cpus: 48-55,112-119 node 6 size: 193531 MB node 6 free: 192814 MB node 7 cpus: 56-63,120-127 node 7 size: 193438 MB node 7 free: 192679 MB node distances: node 0 1 2 3 4 5 6 7 0: 10 12 12 12 32 32 32 32 1: 12 10 12 12 32 32 32 32 2: 12 12 10 12 32 32 32 32 3: 12 12 12 10 32 32 32 32 4: 32 32 32 32 10 12 12 12 5: 32 32 32 32 12 10 12 12 6: 32 32 32 32 12 12 10 12 7: 32 32 32 32 12 12 12 10 ------------------------------------------------------------ 9. /proc/meminfo MemTotal: 1584787004 kB ------------------------------------------------------------ 10. who -r run-level 3 Jul 31 20:22 ------------------------------------------------------------ 11. Systemd service manager version: systemd 254 (254.24+suse.148.g83b9060b6e) Default Target Status multi-user degraded ------------------------------------------------------------ 12. Failed units, from systemctl list-units --state=failed UNIT LOAD ACTIVE SUB DESCRIPTION * logrotate.service loaded failed failed Rotate log files ------------------------------------------------------------ 13. Services, from systemctl list-unit-files STATE UNIT FILES enabled YaST2-Firstboot YaST2-Second-Stage apparmor auditd cron display-manager getty@ irqbalance issue-generator kbdsettings klog lvm2-monitor nscd nvmefc-boot-connections nvmf-autoconnect postfix purge-kernels rollback rsyslog smartd systemd-pstore enabled-runtime systemd-remount-fs disabled autofs autoyast-initscripts blk-availability boot-sysctl ca-certificates chrony-wait chronyd console-getty cups cups-browsed debug-shell ebtables exchange-bmc-os-info firewalld fsidd gpm grub2-once haveged hwloc-dump-hwdata ipmi ipmievd issue-add-ssh-keys kexec-load lunmask man-db-create multipathd nfs nfs-blkmap rpcbind rpmconfigcheck rsyncd serial-getty@ smartd_generate_opts snmpd snmptrapd sshd systemd-boot-check-no-failures systemd-confext systemd-network-generator systemd-sysext systemd-time-wait-sync systemd-timesyncd tuned udisks2 vncserver@ wicked wickedd-auto4 wickedd-dhcp4 wickedd-dhcp6 wickedd-nanny indirect systemd-userdbd wickedd ------------------------------------------------------------ 14. Linux kernel boot-time arguments, from /proc/cmdline BOOT_IMAGE=/boot/vmlinuz-6.4.0-150700.51-default root=UUID=b599d87c-474e-4c72-b781-36d0bcedf18d splash=silent mitigations=auto quiet security=apparmor ------------------------------------------------------------ 15. cpupower frequency-info analyzing CPU 57: current policy: frequency should be within 1.50 GHz and 3.00 GHz. The governor "ondemand" may decide which speed to use within this range. boost state support: Supported: yes Active: yes ------------------------------------------------------------ 16. tuned-adm active It seems that tuned daemon is not running, preset profile is not activated. Preset profile: throughput-performance ------------------------------------------------------------ 17. sysctl kernel.numa_balancing 1 kernel.randomize_va_space 0 vm.compaction_proactiveness 20 vm.dirty_background_bytes 0 vm.dirty_background_ratio 10 vm.dirty_bytes 0 vm.dirty_expire_centisecs 3000 vm.dirty_ratio 8 vm.dirty_writeback_centisecs 500 vm.dirtytime_expire_seconds 43200 vm.extfrag_threshold 500 vm.min_unmapped_ratio 1 vm.nr_hugepages 0 vm.nr_hugepages_mempolicy 0 vm.nr_overcommit_hugepages 0 vm.swappiness 1 vm.watermark_boost_factor 15000 vm.watermark_scale_factor 10 vm.zone_reclaim_mode 1 ------------------------------------------------------------ 18. /sys/kernel/mm/transparent_hugepage defrag [always] defer defer+madvise madvise never enabled [always] madvise never hpage_pmd_size 2097152 shmem_enabled always within_size advise [never] deny force ------------------------------------------------------------ 19. /sys/kernel/mm/transparent_hugepage/khugepaged alloc_sleep_millisecs 60000 defrag 1 max_ptes_none 511 max_ptes_shared 256 max_ptes_swap 64 pages_to_scan 4096 scan_sleep_millisecs 10000 ------------------------------------------------------------ 20. OS release From /etc/*-release /etc/*-version os-release SUSE Linux Enterprise Server 15 SP7 ------------------------------------------------------------ 21. Disk information SPEC is set to: /home/CPU2026 Filesystem Type Size Used Avail Use% Mounted on /dev/nvme0n1p2 xfs 929G 74G 855G 8% / ------------------------------------------------------------ 22. /sys/devices/virtual/dmi/id Vendor: Maginfra Product: QR8228-D3 Product Family: Not specified Serial: 00000000 ------------------------------------------------------------ 23. dmidecode Additional information from dmidecode 3.6 follows. WARNING: Use caution when you interpret this section. The 'dmidecode' program reads system data which is "intended to allow hardware to be accurately determined", but the intent may not be met, as there are frequent changes to hardware, firmware, and the "DMTF SMBIOS" standard. Memory: 12x Samsung M321R8GA0PB2-CCPEC 64 GB 2 rank 6400 11x Samsung M321R8GA0PB2-CCPPC 64 GB 2 rank 6400 1x Samsung M321R8GA0PB2-CCPWC 64 GB 2 rank 6400 ------------------------------------------------------------ 24. BIOS (This section combines info from /sys/devices and dmidecode.) BIOS Vendor: American Megatrends International, LLC. BIOS Version: 01.14.00 BIOS Date: 12/25/2025 Compiler Version Notes ---------------------- ============================================================================== C | 782.lbm_r(base, peak) ------------------------------------------------------------------------------ AMD clang version 17.0.6 (CLANG: AOCC_5.1.0-Build#1994 2025_12_23) Target: x86_64-unknown-linux-gnu Thread model: posix InstalledDir: /opt/AMD/aocc/aocc-compiler-5.1.0/bin ------------------------------------------------------------------------------ ============================================================================== C++ | 731.astcenc_r(base, peak) 736.ocio_r(base, peak) | 748.flightdm_r(base, peak) 766.femflow_r(base, peak) | 767.nest_r(base, peak) 772.marian_r(base, peak) ------------------------------------------------------------------------------ AMD clang version 17.0.6 (CLANG: AOCC_5.1.0-Build#1994 2025_12_23) Target: x86_64-unknown-linux-gnu Thread model: posix InstalledDir: /opt/AMD/aocc/aocc-compiler-5.1.0/bin ------------------------------------------------------------------------------ ============================================================================== C++, C | 709.cactus_r(base, peak) 737.gmsh_r(base, peak) ------------------------------------------------------------------------------ AMD clang version 17.0.6 (CLANG: AOCC_5.1.0-Build#1994 2025_12_23) Target: x86_64-unknown-linux-gnu Thread model: posix InstalledDir: /opt/AMD/aocc/aocc-compiler-5.1.0/bin ------------------------------------------------------------------------------ ============================================================================== Fortran | 722.palm_r(base, peak) 749.fotonik3d_r(base, peak) 765.roms_r(base, | peak) ------------------------------------------------------------------------------ AMD clang version 17.0.6 (CLANG: AOCC_5.1.0-Build#1994 2025_12_23) Target: x86_64-unknown-linux-gnu Thread model: posix InstalledDir: /opt/AMD/aocc/aocc-compiler-5.1.0/bin ------------------------------------------------------------------------------ Base Compiler Invocation ------------------------ C benchmarks: clang C++ benchmarks: clang++ Fortran benchmarks: flang Benchmarks using both C and C++: clang++ clang Base Portability Flags ---------------------- 736.ocio_r: -fno-finite-math-only 737.gmsh_r: -fno-fast-math 748.flightdm_r: -fno-reciprocal-math 767.nest_r: -fno-finite-math-only Base Optimization Flags ----------------------- C benchmarks: -m64 -std=c18 -Wl,-mllvm -Wl,-align-all-nofallthru-blocks=6 -Wl,-mllvm -Wl,-reduce-array-computations=3 -Wl,-mllvm -Wl,-ldist-scalar-expand -fenable-aggressive-gather -ffast-math -O3 -march=znver5 -fveclib=AMDLIBM -fno-PIE -no-pie -flto -fstruct-layout=7 -mllvm -unroll-threshold=50 -mllvm -inline-threshold=1000 -fremap-arrays -fstrip-mining -mllvm -reduce-array-computations=3 -zopt -lamdlibm -lamdalloc -lflang C++ benchmarks: -m64 -std=c++17 -Wl,-mllvm -Wl,-align-all-nofallthru-blocks=6 -Wl,-mllvm -Wl,-reduce-array-computations=3 -Wl,-mllvm -Wl,-x86-use-vzeroupper=false -Wl,-mllvm -Wl,-extra-inliner -ffast-math -O3 -march=znver5 -fveclib=AMDLIBM -flto -mllvm -unroll-threshold=100 -mllvm -loop-unswitch-threshold=200000 -mllvm -reduce-array-computations=3 -zopt -lamdlibm -lamdalloc -lflang Fortran benchmarks: -m64 -Mstandard -Wl,-mllvm -Wl,-align-all-nofallthru-blocks=6 -Wl,-mllvm -Wl,-reduce-array-computations=3 -Wl,-mllvm -Wl,-enable-X86-prefetching -Wl,-mllvm -Wl,-enable-aggressive-gather=true -Wl,-mllvm -Wl,-enable-masked-gather-sequence=false -ffast-math -O3 -march=znver5 -fveclib=AMDLIBM -flto -Mrecursive -funroll-loops -mllvm -lsr-in-nested-loop -mllvm -reduce-array-computations=3 -fepilog-vectorization-of-inductions -zopt -lamdlibm -lamdalloc -lflang Benchmarks using both C and C++: -m64 -std=c++17 -std=c18 -Wl,-mllvm -Wl,-align-all-nofallthru-blocks=6 -Wl,-mllvm -Wl,-reduce-array-computations=3 -Wl,-mllvm -Wl,-x86-use-vzeroupper=false -Wl,-mllvm -Wl,-extra-inliner -ffast-math -O3 -march=znver5 -fveclib=AMDLIBM -fno-PIE -no-pie -flto -fstruct-layout=7 -mllvm -unroll-threshold=50 -mllvm -inline-threshold=1000 -fremap-arrays -fstrip-mining -mllvm -reduce-array-computations=3 -zopt -mllvm -unroll-threshold=100 -mllvm -loop-unswitch-threshold=200000 -lamdlibm -lamdalloc -lflang Peak Compiler Invocation ------------------------ C benchmarks: clang C++ benchmarks: clang++ Fortran benchmarks: flang Benchmarks using both C and C++: clang++ clang Peak Portability Flags ---------------------- 736.ocio_r: -fno-finite-math-only 737.gmsh_r: -fno-fast-math 748.flightdm_r: -fno-reciprocal-math 767.nest_r: -fno-finite-math-only Peak Optimization Flags ----------------------- C benchmarks: -m64 -std=c18 -Wl,-mllvm -Wl,-align-all-nofallthru-blocks=6 -Wl,-mllvm -Wl,-reduce-array-computations=3 -ffast-math -Ofast -march=znver5 -fveclib=AMDLIBM -flto -fstruct-layout=7 -mllvm -unroll-threshold=50 -fremap-arrays -fstrip-mining -mllvm -inline-threshold=1000 -mllvm -reduce-array-computations=3 -zopt -lamdlibm -lamdalloc C++ benchmarks: 731.astcenc_r: -m64 -std=c++17 -Wl,-mllvm -Wl,-align-all-nofallthru-blocks=6 -Wl,-mllvm -Wl,-reduce-array-computations=3 -Wl,-mllvm -Wl,-x86-use-vzeroupper=false -Wl,-mllvm -Wl,-extra-inliner -ffast-math -O3 -march=znver5 -fveclib=AMDLIBM -flto -mllvm -unroll-threshold=100 -mllvm -loop-unswitch-threshold=200000 -mllvm -reduce-array-computations=3 -zopt -lamdlibm -lamdalloc -lflang 736.ocio_r: Same as 731.astcenc_r 748.flightdm_r: basepeak = yes 766.femflow_r: basepeak = yes 767.nest_r: basepeak = yes 772.marian_r: basepeak = yes Fortran benchmarks: 722.palm_r: -m64 -Mstandard -Wl,-mllvm -Wl,-align-all-nofallthru-blocks=6 -Wl,-mllvm -Wl,-reduce-array-computations=3 -Wl,-mllvm -Wl,-enable-X86-prefetching -Wl,-mllvm -Wl,-enable-aggressive-gather=true -Wl,-mllvm -Wl,-enable-masked-gather-sequence=false -ffast-math -O3 -march=znver5 -fveclib=AMDLIBM -flto -Mrecursive -funroll-loops -mllvm -lsr-in-nested-loop -mllvm -reduce-array-computations=3 -fepilog-vectorization-of-inductions -zopt -lamdlibm -lamdalloc -lflang 749.fotonik3d_r: -m64 -Mstandard -Wl,-mllvm -Wl,-align-all-nofallthru-blocks=6 -Wl,-mllvm -Wl,-reduce-array-computations=3 -ffast-math -Ofast -march=znver5 -fveclib=AMDLIBM -flto -Mrecursive -mllvm -reduce-array-computations=3 -fepilog-vectorization-of-inductions -zopt -lamdlibm -lamdalloc -lflang 765.roms_r: Same as 749.fotonik3d_r Benchmarks using both C and C++: -m64 -std=c++17 -std=c18 -Wl,-mllvm -Wl,-align-all-nofallthru-blocks=6 -Wl,-mllvm -Wl,-reduce-array-computations=3 -Wl,-mllvm -Wl,-x86-use-vzeroupper=false -Wl,-mllvm -Wl,-extra-inliner -ffast-math -O3 -march=znver5 -fveclib=AMDLIBM -fno-PIE -no-pie -flto -fstruct-layout=7 -mllvm -unroll-threshold=50 -mllvm -inline-threshold=1000 -fremap-arrays -fstrip-mining -mllvm -reduce-array-computations=3 -zopt -mllvm -unroll-threshold=100 -mllvm -loop-unswitch-threshold=200000 -lamdlibm -lamdalloc -lflang The flags files that were used to format this result can be browsed at http://www.spec.org/cpu2026/results/flags/aocc-flags.2026-05-04.00.html http://www.spec.org/cpu2026/results/flags/Maginfra-Platform-Settings-amd-V1.1.html You can also download the XML flags sources by saving the following links: http://www.spec.org/cpu2026/results/flags/aocc-flags.2026-05-04.00.xml http://www.spec.org/cpu2026/results/flags/Maginfra-Platform-Settings-amd-V1.1.xml SPEC CPU and SPECrate are registered trademarks of the Standard Performance Evaluation Corporation. All other brand and product names appearing in this result are trademarks or registered trademarks of their respective holders. ---------------------------------------------------------------------------------------------------------------------------------- For questions about this result, please contact the tester. For other inquiries, please contact info@spec.org. Copyright 2026 Standard Performance Evaluation Corporation Tested with SPEC CPU(R)2026 v1.0.1 on 2026-08-01 01:12:33-0400. Report generated on 2026-09-23 13:08:11 by CPU2026 text formatter (aa18784a). Originally published on 2026-09-22.