Summit
E112111
Summit is a high-performance supercomputer at Oak Ridge National Laboratory that was once the world’s fastest, designed for large-scale scientific and artificial intelligence research.
All labels observed (1)
| Label | Occurrences |
|---|---|
| Summit canonical | 2 |
Statements (50)
| Predicate | Object |
|---|---|
| instanceOf |
high-performance computing system
ⓘ
supercomputer ⓘ |
| applicationDomain |
astrophysics
ⓘ
climate modeling ⓘ computational chemistry ⓘ deep learning ⓘ genomics ⓘ materials science ⓘ nuclear physics ⓘ |
| architecture |
IBM Power Systems
ⓘ
surface form:
IBM AC922
|
| commissioned | 2018 ⓘ |
| country |
United States of America
ⓘ
surface form:
United States
|
| CPUVendor | IBM ⓘ |
| designedFor |
artificial intelligence research
ⓘ
data-intensive computing ⓘ large-scale scientific research ⓘ machine learning workloads ⓘ |
| energyEfficiency | ranked on Green500 list ⓘ |
| firstDeploymentYear | 2018 ⓘ |
| fundedBy |
Office of Science
ⓘ
surface form:
U.S. Department of Energy Office of Science
|
| GPUVendor |
NVIDIA Corporation
ⓘ
surface form:
NVIDIA
|
| heldTitle | world’s fastest supercomputer ⓘ |
| heldTitleFrom | June 2018 ⓘ |
| heldTitleUntil | June 2020 ⓘ |
| hostInstitution | Oak Ridge Leadership Computing Facility ⓘ |
| interconnect |
Mellanox Technologies
ⓘ
surface form:
Mellanox EDR InfiniBand
|
| LINPACKPerformance | over 140 petaflops ⓘ |
| locatedIn |
Oak Ridge National Laboratory
ⓘ
Oak Ridge, Tennessee ⓘ |
| managedBy | U.S. Department of Energy ⓘ |
| manufacturer | IBM ⓘ |
| memoryType |
DDR4
ⓘ
High Bandwidth Memory (HBM2 on GPUs) ⓘ |
| operator | Oak Ridge National Laboratory ⓘ |
| OS | Linux ⓘ |
| peakPerformance | approximately 200,000 teraflops ⓘ |
| predecessor |
Cray XK7
ⓘ
surface form:
Titan supercomputer
|
| rankedNumberOneOn | TOP500 list ⓘ |
| storageSystem |
GPFS (IBM Spectrum Scale)
ⓘ
surface form:
IBM Spectrum Scale (GPFS)
|
| storageType | parallel file system ⓘ |
| successor | Frontier supercomputer ⓘ |
| supportsPrecision |
double precision floating point
ⓘ
mixed precision computing ⓘ single precision floating point ⓘ |
| surpassedBy | Fugaku ⓘ |
| systemType | heterogeneous CPU-GPU system ⓘ |
| topSpeed | approximately 200 petaflops (double-precision) ⓘ |
| usedFor | COVID-19 related simulations and drug discovery research ⓘ |
| usesAccelerator |
NVIDIA Tesla data center GPUs
ⓘ
surface form:
NVIDIA Tesla V100 GPU
|
| usesProcessor |
IBM Power Systems
ⓘ
surface form:
IBM POWER9 CPU
|
Referenced by (2)
Full triples — surface form annotated when it differs from this entity's canonical label.