From WikiChip
Frontera - Supercomputers
| Edit Values | |
| Frontera | |
| General Info | |
| Sponsors | National Science Foundation (NSF) |
| Designers | Intel, Dell |
| Operators | Texas Advanced Computing Center |
| Introduction | 2019 |
| Peak FLOPS | 38,746 TFLOPS |
| Succession | |
Frontera is a the successor to Stampede 2, a planned academic supercomputer with a peak performance of 35-40 petaflops. Frontera is expected to deliver the 3x speedup in real-application performance over Blue Waters at about one-third the cost.
Contents
Overview
Announced in late August 2018, Frontera is expected to come online in 2019 and replace Stampede 2 as the fastest academic supercomputer in the United States. The system will be delivered by Dell EMC and incorporate ~16,000 Intel Xeon processors based on the Cascade Lake microarchitecture.
- Frontera - Dell C6420, Intel Xeon Platinum 8280 28C 2.7 GHz, Mellanox InfiniBand HDR
- Year: 2019 • Cores: 448,448 • Rmax: 23,516,400 GFLOPS • Rpeak: 38,745,907 GFLOPS
- • Texas Advanced Computing Center (TACC)
- • University of Texas at Austin
- Blue Waters - Cray XE/XK, AMD "Interlagos" CPU, NVIDIA K20X GPU, 13,000 PFLOPS
- • National Center for Supercomputing Applications (NCSA)
- • University of Illinois at Urbana-Champaign
- • National Science Foundation (NSF)
Models
- TACC
- Frontera • 2019
- Longhorn • 2019
- Stampede 2 • 2017
- Stampede KNL • 2016
- Lonestar 5 • 2015
- Maverick • 2014
- Stampede • 2012
- Lonestar 4 • 2011
- Ranger • 2008
- Lonestar 2 • 2006
- Wrangler 2 • 2005
- Wrangler • 2005
- Lonestar • 2004
TOP500
Texas Advanced Computing Center/Univer. of Texas URL: http://www.tacc.utexas.edu/ Segment Academic City: Austin Country/Region: United States System Year Vendor • Cores • Rmax (GFlop/s) • Rpeak (GFlop/s) *Frontera - PowerEdge C6420, Intel Xeon Platinum 8280 28C 2.7 GHz, Mellanox InfiniBand HDR-100 :2019 DELL • 448,448 • 23,516,400 • 38,745,907 *Stampede 2 - PowerEdge C6320P/C6420, Intel Xeon Phi 7250 68C 1.4 GHz/Platinum 8160, Intel Omni-Path :2017 DELL • 367,024 • 10,680,700 • 18,309,222 *Stampede - PowerEdge C8220, Intel Xeon E5-2680 8C 2.7 GHz, Infiniband FDR, Intel Xeon Phi SE10P :2012 DELL • 522,080 • 5,168,110 • 9,598,208 *Longhorn - IBM Power System AC922, IBM POWER9 20C 3.6 GHz, NVIDIA Tesla V100 SXM2, Infiniband EDR :2019 IBM • 37,440 • 2,288,000 • 2,867,405 *Lonestar 5 - Cray XC40, Intel Xeon E5-2690v3 12C 2.6 GHz, Aries interconnect :2015 Cray/HPE • 30,048 • 1,017,040 • 1,249,997 *Stampede-KNL - Intel S7200AP Cluster, Intel Xeon Phi 7250 68C 1.4 GHz, Intel Omni-Path :2016 Dell/Intel • 34,272 • 842,869 • 1,535,386 *Ranger - SunBlade x6420, AMD Opteron QC 2.3 GHz, Infiniband :2008 Sun Microsystems • 62,976 • 433,200 • 579,379.2 *Ranger - SunBlade x6420, AMD Opteron Quad 2.0 GHz, Infiniband :2008 Sun Microsystems • 62,976 • 326,000 • 503,808 *Lonestar 4 - Dell PowerEdge M610 Cluster, Intel Xeon 5680 3.3 GHz, Infiniband QDR :2011 DELL • 22,656 • 251,800 • 301,777 *Maverick - Cluster Platform SL250s Gen8, Intel Xeon E5-2680v2 10C 2.8 GHz, Infiniband FDR, Nvidia K40m :2014 HPE • 4,620 • 158,100 • 247,896 *Lonestar - PowerEdge 1955, 2.66 GHz, Infiniband :2007 DELL • 5,848 • 46,730 • 62,220 *Lonestar - PowerEdge 1955, 2.66 GHz, Infiniband :2006 DELL • 5,200 • 41,460 • 55,473.6 *Lonestar2 - PowerEdge 1855, 3.2 GHz, Infiniband :2006 DELL • 1,300 • 6,989 • 8,320 *Lonestar - PowerEdge 1750, Pentium 4 Xeon 3.2/3.06 GHz, Myrinet :2004 Dell/Cray • 1,024 • 4,152 • 6,338 *Lonestar - PowerEdge 1750, Pentium 4 Xeon 3.06 GHz, Myrinet :2003 Dell/Cray • 600 • 2,455 • 3,672 *Wrangler 2 - PowerEdge 1855, 3.2 GHz, Infiniband :2005 DELL • 400 • 2,141 • 2,560 *Wrangler - PowerEdge 1850, 3.2 GHz, Myrinet :2005 DELL • 256 • 1,269 • 1,638 *Longhorn - pSeries p690/p655 1.3 GHz :2002 IBM • 224 • 704.8 • 1,164.8
Bibliography
- Texas Advanced Computing Center (TACC). (August 29, 2018). Personal communication.
See also
Facts about "Frontera - Supercomputers"
| designer | Intel + and Dell + |
| introductory date | 2019 + |
| logo | |
| name | Frontera + |
| operator | Texas Advanced Computing Center + |
| peak flops (double-precision) | 3.8746e+16 FLOPS (38,746,000,000,000 KFLOPS, 38,746,000,000 MFLOPS, 38,746,000 GFLOPS, 38,746 TFLOPS, 38.746 PFLOPS, 0.0387 EFLOPS, 3.8746e-5 ZFLOPS) + |
| sponsor | National Science Foundation (NSF) + |