|
Intel Hatch: Intel Server D50DNP1SB (Xeon Platinum |
SPEChpc 2021_tny_base = 60.4 |
|
SPEChpc 2021_tny_peak = 64.5 |
| hpc2021 License: | 13 | Test Date: | Apr-2025 |
|---|---|---|---|
| Test Sponsor: | Intel | Hardware Availability: | Jan-2023 |
| Tested by: | Intel | Software Availability: | Mar-2025 |
Benchmark result graphs are available in the PDF report.
Results Table
| Benchmark | Base | Peak | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Model | Ranks | Thrds/Rnk | Seconds | Ratio | Seconds | Ratio | Seconds | Ratio | Model | Ranks | Thrds/Rnk | Seconds | Ratio | Seconds | Ratio | Seconds | Ratio | |
| SPEChpc 2021_tny_base | 60.4 | |||||||||||||||||
| SPEChpc 2021_tny_peak | 64.5 | |||||||||||||||||
| Results appear in the order in which they were run. Bold underlined text indicates a median measurement. | ||||||||||||||||||
| 505.lbm_t | TGT | 8 | 1 | 13.6 | 165 | 13.1 | 171 | 13.3 | 170 | TGT | 8 | 1 | 12.7 | 178 | 12.7 | 177 | 12.6 | 178 |
| 513.soma_t | TGT | 8 | 1 | 22.9 | 162 | 23.0 | 161 | 23.0 | 161 | TGT | 8 | 1 | 22.9 | 162 | 23.0 | 161 | 23.0 | 161 |
| 518.tealeaf_t | TGT | 8 | 1 | 38.1 | 43.3 | 38.0 | 43.4 | 37.5 | 44.0 | TGT | 8 | 1 | 37.1 | 44.5 | 36.4 | 45.4 | 36.6 | 45.1 |
| 519.clvleaf_t | TGT | 8 | 1 | 27.5 | 60.0 | 27.9 | 59.2 | 27.6 | 59.8 | TGT | 8 | 1 | 27.5 | 60.0 | 27.9 | 59.2 | 27.6 | 59.8 |
| 521.miniswp_t | TGT | 8 | 1 | 52.3 | 30.6 | 52.7 | 30.3 | 52.7 | 30.3 | TGT | 8 | 1 | 52.6 | 30.4 | 53.0 | 30.2 | 53.5 | 29.9 |
| 528.pot3d_t | TGT | 8 | 1 | 34.1 | 62.3 | 34.4 | 61.8 | 34.3 | 61.9 | TGT | 8 | 1 | 32.7 | 65.0 | 32.8 | 64.9 | 32.6 | 65.1 |
| 532.sph_exa_t | TGT | 8 | 1 | 66.3 | 29.4 | 68.1 | 28.7 | 69.5 | 28.1 | TGT | 8 | 1 | 70.1 | 27.8 | 70.2 | 27.8 | 68.3 | 28.6 |
| 534.hpgmgfv_t | TGT | 8 | 1 | 58.9 | 19.9 | 62.8 | 18.7 | 59.4 | 19.8 | TGT | 8 | 1 | 37.0 | 31.8 | 36.3 | 32.4 | 36.4 | 32.3 |
| 535.weather_t | TGT | 8 | 1 | 22.7 | 142 | 22.8 | 142 | 22.8 | 142 | TGT | 8 | 1 | 22.7 | 142 | 22.8 | 142 | 22.8 | 142 |
| Hardware Summary | |
|---|---|
| Type of System: | Homogenous Cluster |
| Compute Node: | Intel Server D50DNP1SB (Xeon Platinum 8480+) |
| Interconnect: | Mellanox HDR |
| Compute Nodes Used: | 1 |
| Total Chips: | 2 |
| Total Cores: | 112 |
| Total Threads: | 224 |
| Total Memory: | 1 TB |
| Total Accelerators: | 4 |
| Max. Peak Threads: | 1 |
| Software Summary | |
|---|---|
| Compiler: | Intel oneAPI Compiler 2025.1.0 |
| MPI Library: | Intel MPI Library 2021.15 for Linux OS |
| Other MPI Info: | None |
| Other Software: | None |
| Base Parallel Model: | TGT |
| Base Ranks Run: | 8 |
| Base Threads Run: | 1 |
| Peak Parallel Models: | TGT |
| Minimum Peak Ranks: | 8 |
| Maximum Peak Ranks: | 8 |
| Max. Peak Threads: | 1 |
| Min. Peak Threads: | 1 |
Node Description: Intel Server D50DNP1SB (Xeon Platinum 8480+)
| Hardware | |
|---|---|
| Number of nodes: | 1 |
| Uses of the node: | Compute |
| Vendor: | Intel |
| Model: | Intel Server D50DNP1SB (2 x Intel Xeon Platinum 8480+, 2.0GHz) |
| CPU Name: | Intel Xeon Platinum 8480+ |
| CPU(s) orderable: | 1, 2 chips |
| Chips enabled: | 2 |
| Cores enabled: | 112 |
| Cores per chip: | 56 |
| Threads per core: | 2 |
| CPU Characteristics: | Turbo Boost Technology up to 3.8 GHz |
| CPU MHz: | 2000 |
| Primary Cache: | 32 KB I + 48 KB D on chip per core |
| Secondary Cache: | 2 MB I+D on chip per core |
| L3 Cache: | 105 MB I+D on chip per chip |
| Other Cache: | None |
| Memory: | 1 TB (16x64 GB DDR5 2Rx4 PC5-4800B-R) |
| Disk Subsystem: | 1 x 1 1TB NVMe M.2 INTEL SSDPELKX010T8 |
| Other Hardware: | None |
| Accel Count: | 4 |
| Accel Model: | Intel Data Center GPU Max 1550 |
| Accel Vendor: | Intel |
| Accel Type: | GPU |
| Accel Connection: | PCIe Gen5 x16 |
| Accel ECC enabled: | yes |
| Accel Description: | Intel Data Center GPU Max 1550 |
| Adapter: | Mellanox ConnectX-6 HDR |
| Number of Adapters: | 1 |
| Slot Type: | PCI-Express 4.0 x16 |
| Data Rate: | 200Gbit/s |
| Ports Used: | 1 |
| Interconnect Type: | Mellanox HDR |
| Software | |
|---|---|
| Accelerator Driver: | 25.05.32567 |
| Adapter: | Mellanox ConnectX-6 HDR |
| Adapter Firmware: | 20.38.1900 |
| Operating System: | SUSE Linux Enterprise Server 15 SP6 6.4.0-150600.23.42-default |
| Local File System: | lustre |
| Shared File System: | LUSTRE FS |
| System State: | Run level 5 |
| Other Software: | None |
Interconnect Description: Mellanox HDR
| Hardware | |
|---|---|
| Vendor: | Mellanox |
| Model: | Mellanox HDR |
| Switch Model: | Mellanox Technologies MT28908 Family InfiniBand Switch |
| Number of Switches: | 12 |
| Number of Ports: | 40 |
| Data Rate: | 200 Gbit/s |
| Firmware: | 20.38.1900 |
| Topology: | Fat-tree |
| Primary Use: | MPI Traffic, LustreFS traffic |
| Software |
|---|
Submit Notes
The config file option 'submit' was used.
General Notes
Environment variables set by runhpc before the start of the run: LIBOMPTARGET_LEVEL_ZERO_USE_IMMEDIATE_COMMAND_LIST = "all" I_MPI_FABRICS=shm:ofi I_MPI_OFFLOAD=1 I_MPI_OFFLOAD_CELL=tile I_MPI_OFFLOAD_TOPOLIB=level_zero I_MPI_OFFLOAD_CELL_LIST=0,1,2,3,4,5,6,7 For the following tests src.alt was used in PEAK: 518 521 528 532 534
Platform Notes
Device Vendor Intel Device Version OpenCL 3.0 NEO Driver Version 25.05.32567 Base clock 900MHz Max clock frequency 1600MHz Tiles 2 Slices per Tile 1 Max compute units per Tile 512 Sub-slices per slice 64 EUs per sub-slice 8 Threads per EU 8 Max work item dimensions 3 Max work item sizes 1024x1024x1024 Max work group size 1024 Preferred work group size multiple 32 Max sub-groups per work group 64 Sub-group sizes 16, 32 L1 Cache per EU 65536 L2 cache size 427819008 Global memory size 137438953472 Address bits 64, Little-Endian
Base Compiler Invocation
C benchmarks:
| mpiicc -cc=icx |
C++ benchmarks:
| mpiicpc -cxx=icpx |
Fortran benchmarks:
| mpiifort -fc=ifx |
Base Portability Flags
| 505.lbm_t: | -DUSE_MPI |
| 513.soma_t: | -DUSE_MPI -DSPEC_NO_VAR_ARRAY_REDUCE |
| 518.tealeaf_t: | -DUSE_MPI |
| 519.clvleaf_t: | -DUSE_MPI |
| 528.pot3d_t: | -DUSE_MPI |
| 535.weather_t: | -DUSE_MPI |
Base Optimization Flags
C benchmarks:
C++ benchmarks:
Fortran benchmarks:
Peak Compiler Invocation
C benchmarks:
| mpiicc -cc=icx |
C++ benchmarks:
| mpiicpc -cxx=icpx |
Fortran benchmarks:
| mpiifort -fc=ifx |
Peak Portability Flags
| 505.lbm_t: | -DUSE_MPI |
| 513.soma_t: | -DUSE_MPI -DSPEC_NO_VAR_ARRAY_REDUCE |
| 518.tealeaf_t: | -DUSE_MPI |
| 519.clvleaf_t: | -DUSE_MPI |
| 528.pot3d_t: | -DUSE_MPI |
| 535.weather_t: | -DUSE_MPI |
Peak Optimization Flags
C benchmarks:
C++ benchmarks:
Fortran benchmarks:
| 519.clvleaf_t: | basepeak = yes |
| 528.pot3d_t: | -DSPEC_COLLAPSE -O3 -xCORE-AVX512 -flto -mprefer-vector-width=512 -ffast-math -fiopenmp -fopenmp-targets=spir64_gen -ftarget-register-alloc-mode=pvc:auto -Xopenmp-target-backend '-device pvc -revision_id 0x2f' -DSPEC_ACCEL_AWARE_MPI -fopenmp-target-loopopt |
| 535.weather_t: | basepeak = yes |