Vous avez reçu un message "Your GitLab account has been locked ..." ? Pas d'inquiétude : lisez cet article https://docs.gricad-pages.univ-grenoble-alpes.fr/help/unlock/

Commit e265153e authored by ramdsc's avatar ramdsc Committed by Millian Poquet
Browse files

Minimized data from JSON files

parent 9a52d42c
......@@ -11,18 +11,29 @@ Fundamental architecture approach:
![Architecture model](./doc/architecture_model.png)
- **Platform**: root element for SimGrid platform descriptions.
- **Main zone**: top-level zone with *Full* routing; interconnects the master host with the set of clusters.
- **Master zone**: contains and isolates the master host.
- **Master host**: processes communication with the decision system and job submissions.
- **Config zone**: holds user defined properties associated to node and processor types.
- **Clusters**: groups of heterogeneous nodes in the data centre.
- **Router**: provides connectivity with the master host and thus job allocation.
- **Backbone**: intra-cluster communications.
- **Nodes**: computational resources integrating one or more types of processors.
- **Processors**: computing devices, can be CPUs, GPUs, MICs or whichever defined.
- **Cores**: individual computing units inside a processor, homogeneous within the processor.
- **Memory**: amount of storage shared by all processors in the node.
- **Platform**: root element for SimGrid platform descriptions.
- **Main zone**: top-level zone with *Full* routing; interconnects the master host with the set of clusters.
- **Master zone**: contains and isolates the master host.
- **Master host**: processes communication with the decision system and job submissions.
- **Config zone**: holds user defined properties associated to node and processor types.
- **Clusters**: groups of heterogeneous nodes in the data centre.
- **Router**: provides connectivity with the master host and thus job allocation.
- **Backbone**: intra-cluster communications.
- **Nodes**: computational resources integrating one or more types of processors.
- **Processors**: computing devices, can be CPUs, GPUs, MICs or whichever defined.
- **Cores**: individual computing units inside a processor, homogeneous within the processor.
- **Memory**: amount of storage shared by all processors in the node.
Energy and computational power are expressed per core for fine grain analysis. Cores are SimGrid *hosts*, so they are considered as individual resources.
Network types apply to both the node up / down links, the cluster backbones and the global links in the main zone.
......@@ -30,9 +41,9 @@ Network types apply to both the node up / down links, the cluster backbones and
## Usage
You may define:
- Node types in `node_types.json`
- Processor types in `processor_types.json`
- Network types in `network_types.json`
- Node types in `node_types.json`
- Processor types in `processor_types.json`
- Network types in `network_types.json`
Create a JSON file with the platform description, `example.json` shows a possible template.
......
{
"_comments": "Latency values obtained from http://www.crucial.com/usa/en/memory-performance-speed-latency",
"DDR3-1333": {
"latency": "13.50ns"
},
"DDR3-1600": {
"latency": "13.75ns"
},
"DDR4-1866": {
"latency": "13.93ns"
},
"DDR4-2133": {
"latency": "14.06ns"
},
"DDR4-2400": {
"latency": "14.17ns"
},
"DDR4-2666": {
"latency": "13.50ns"
}
}
{
"Example Network": {
"bandwidth": "10.0Gbps",
"latency": "0.5us"
"bandwidth": "10.0Gbps"
},
"InfiniBand FDR x4": {
"bandwidth": "54.54Gbps",
"latency": "0.7us"
"bandwidth": "54.54Gbps"
},
"Gigabit Ethernet 1000BASE-T": {
"badwidth": "1.0Gbps"
}
}
{
"Example Node": {
"id": "example_node",
"Superserver SYS-6019P-MT": {
"id": "sys6019p",
"processors": [
{ "type": "CPU", "model": "Example CPU", "number": "2" },
{ "type": "GPU", "model": "Example GPU", "number": "1" }
{ "type": "Xeon Silver 4114", "number": "2" }
],
"memory_gib": "128"
},
"System x iDataPlex dx360 M4": {
"id": "dx360m4",
"processors": [
{ "type": "CPU", "model": "Xeon Sandy Bridge E5-2670", "number": "2" }
],
"memory_gib": "64"
},
"System x iDataPlex dx360 M3": {
"id": "dx360m3",
"processors": [
{ "type": "GPU", "model": "Tesla M2090", "number": "2" }
],
"memory_gib": "64"
},
"POWER 730 Express": {
"id": "p730ex",
"processors": [
{ "type": "CPU", "model": "POWER 7", "number": "2" }
],
"memory_gib": "64"
"memory": {
"type": "DDR4-2400",
"capacity_gib": "32"
}
}
}
{
"CPU": {
"Example CPU": {
"id": "example_cpu",
"nb_cores": "1",
"_metadata": {
"clock": "2GHz",
"tdp": "85W",
"vector": "SSE",
"vector_length_bits": "128"
},
"core_attributes": {
"speed": "10.0Gf, 9.0Gf"
},
"core_properties": [
{ "id": "sleep_pstates", "value": "2:3:4"},
{ "id": "watt_per_state", "value": "4.0:8.0, 4.0:7.2, 1.0:1.0, 3.0:3.0, 3.0:3.0" },
{ "id": "watt_off", "value": "2.0" }
]
},
"Xeon Sandy Bridge E5-2670": {
"id": "xsbe52670",
"nb_cores": "8",
"_metadata": {
"clock": "2.6GHz",
"tdp": "115W",
"vector": "AVX",
"vector_length_bits": "128"
},
"core_attributes": {
"speed": "20.4Gf, 18.36Gf, 16.32Gf, 14.28Gf, 12.24Gf, 0.0Gf, 0.0Gf, 0.0Gf"
},
"core_properties": [
{ "id": "sleep_pstates", "value": "5:6:7"},
{ "id": "watt_per_state", "value": "5.0:14.375, 5.0:12.9375, 5.0:11.5, 5.0:10.0625, 5.0:8.625, 2.0:2.0, 7.0:7.0, 6.0:6.0" },
{ "id": "watt_off", "value": "2.0" }
]
},
"POWER 7": {
"id": "power7",
"nb_cores": "8",
"_metadata": {
"clock": "3.0GHz",
"tdp": "180W",
"vector": "VMX and VSX",
"vector_length_bits": "128"
},
"core_attributes": {
"speed": "24.0Gf, 21.6Gf, 19.2Gf, 16.8Gf, 14.4Gf, 0.0Gf, 0.0Gf, 0.0Gf"
},
"core_properties": [
{ "id": "sleep_pstates", "value": "5:6:7"},
{ "id": "watt_per_state", "value": "8.0:22.5, 8.0:20.25, 8.0:18.0, 8.0:15.75, 8.0:13.5, 3.0:3.0, 11.0:11.0, 9.0:9.0" },
{ "id": "watt_off", "value": "3.0" }
]
}
"Xeon E5540": {
"_arch": "Nehalem",
"id": "xe5540",
"type": "CPU",
"nb_cores": "4",
"clock_ghz": "2.53",
"tdp_watts": "80",
"dpflops_per_cycle": "4"
},
"GPU": {
"Example GPU": {
"id": "example_gpu",
"nb_cores": "2",
"_metadata": {
"clock": "500MHz",
"tdp": "120W"
},
"core_attributes": {
"speed": "1.0Gf, 0.9Gf"
},
"core_properties": [
{ "id": "sleep_pstates", "value": "2:3:4"},
{ "id": "watt_per_state", "value": "0.2:0.5, 0.2:0.45, 0.05:0.05, 0.25:0.25, 0.22:0.22" },
{ "id": "watt_off", "value": "0.05" }
]
},
"Tesla M2090": {
"id": "teslam2090",
"nb_cores": "512",
"_metadata": {
"clock": "650MHz",
"tdp": "250W"
},
"core_attributes": {
"speed": "1.3Gf, 1.17Gf, 1.04Gf, 0.91Gf, 0.78Gf, 0.0Gf, 0.0Gf, 0.0Gf"
},
"core_properties": [
{ "id": "sleep_pstates", "value": "5:6:7"},
{ "id": "watt_per_state", "value": "0.2:0.5, 0.2:0.45, 0.2:0.4, 0.2:0.35, 0.2:0.3, 0.05:0.05, 0.25:0.25, 0.22:0.22" },
{ "id": "watt_off", "value": "0.05" }
]
}
"Xeon E5-4657L v2": {
"_arch": "Ivy Bridge",
"id": "xe54657Lv2",
"type": "CPU",
"nb_cores": "12",
"clock_ghz": "2.4",
"tdp_watts": "115",
"dpflops_per_cycle": "8"
},
"Xeon E5-4669 v3": {
"_arch": "Haswell",
"id": "xe54669v3",
"type": "CPU",
"nb_cores": "18",
"clock_ghz": "2.1",
"tdp_watts": "135",
"dpflops_per_cycle": "16"
},
"Xeon Silver 4114": {
"_arch": "Skylake",
"id": "xs4114",
"type": "CPU",
"nb_cores": "10",
"clock_ghz": "2.2",
"tdp_watts": "85",
"dpflops_per_cycle": "16"
},
"Xeon Platinum 8180": {
"_arch": "Skylake",
"id": "xp8180",
"type": "CPU",
"nb_cores": "28",
"clock_ghz": "3.8",
"tdp_watts": "205",
"dpflops_per_cycle": "32"
},
"Xeon Phi 7290F": {
"_arch": "Knights Landing",
"id": "xphi7290f",
"type": "MIC",
"nb_cores": "72",
"clock_ghz": "1.5",
"tdp_watts": "260",
"dpflops_per_cycle": "32"
}
}
{
"clusters": [
{
"nodes": [
{ "type": "Superserver SYS-6019P-MT", "number": "5" }
],
"cluster_network": {
"type": "Gigabit Ethernet 1000BASE-T",
"latency": "100us"
}
}
],
"dc_network": {
"type": "Gigabit Ethernet 1000BASE-T",
"latency": "150us"
}
}
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment