| year | cpus |
storage | tools | applications |
| 1954 | First computer--ORACLE
(Oak Ridge Automated Computer And Logical Engine)
photo | | | |
| 1985 | First distributed memory parallel computer delivered --
Intel iPSC/1, 32 cpus, Serial # 1
Photo 7568-85 7217-85 7219-85 |
(1983)CDIAC originated; won Data Center awardLANL's CFS storage system | | |
| 1986 | First shared memory parallel computer delivered --
Sequent Balance 8000, 8 cpus
Photo 8799-97 |
| message-passing multiprocessor simulator
Denelcor HEP multiprocessor simulator | |
| 1987 | | |
benchmarking of Ametek S14, Intel iPSC, and
nCube hypercubes | |
| 1988 | Distributed memory --
Intel iPSC/2, 64 cpus, Serial # 1
Photo 7027-91 |
| Hypercube simulation on a LAN
benchmarking of Intel iPSC/2 hypercube
| |
| 1989 | | |
PICL developed
ParaGraph developed
remote host facility for Intel hypercubes |
|
| 1990 | Distributed memory --
Intel iPSC/860, 128 cpus, Serial # 1
Photo 8056-90 8054-90 |
Storage Tek Tape Robot 1.1 TBARM archive established | |
World's first publicly available 3D deterministic radiation
transport code (TORT) released by ORNL
Gordon Bell award (price/performance)
1st prize IBM Supercomputing competition
(Scientific Excellence in Supercomputing)
Cray GigaFLOP award for superconductivity calculations |
| 1991 |
Shared memory --
Kendall Square KSR1-32, Serial # 1
Photo 01344-92 |
|
PVM
benchmarking of Intel iPSC/860 and nCube 6400 hypercubes |
CHAMMP |
1992
CCS added |
Kendall Square upgraded to KSR1-64 gif
Intel Touchstone Sigma with 66 GP nodes (aka Paragon XP/S 5)
Serial # 1
Intel Paragon XP/S 35 with 512 GP nodes
gif |
|
PVM application wins SC92 Heterogeneous Computing Challenge
("Most Heterogeneous")
hypercube clock synchronization algorithm |
Spectral transform method parallelized |
| 1993 |
XP/S 35 message co-processors enabled |
Storage Tek upgraded to 4.8 TB |
XPVM wins
Heterogeneous computing challenge
(most elegant interface)instrumenting message-passing Fortran | |
1994
CCII formed |
KSR has 100% uptime
Paragon XP/S 35 passes acceptance test
XP/S 35 upgraded to 32MB per node
XP/S 5 message co-processors enabled
IBM SP-2 delivered 05945-96
Intel MP XP/S 14 with 96 MP nodes arrives |
108 GB Max Strat RAID disk
CREO Optical 1 TB/tapeIBM 3495 Tape Robot 100 TB
NSL-UniTree softwareARM archive moved to UniTree |
PICL 2.0 released
DOLIB and
DONIO
PVM3 wins R&D 100
ParaGraph performance analysis package adopted and marketed by Intel SSD.
EES verifies interprocessor communication latency of Intel Paragon MP
at target figure of 30 microseconds
EES research on the effect on performance of job
assignment to paragon processors used to design and
evaluate scheduling algorithms.
|
CHAMMP run of PCCM2 on 512 nodes of XP/S 35 simulating 250 days
Spectral algorithms for shallow water code
Order N scaling for LSMS up to 512 atoms
Solved microscopic theory of the
vortex
state in superconductors
Magnetic moment inhomogeneities in
disordered
NiCu alloys
CETEP ported to paragons to study Ge adatom on Si
GCT 1.0 run on 512 processors (using 1.3 million grid points)
PICS Molecular dynamics world record
Car crash modeling
gif
TORT converted from CTSS to UNICOS
|
1995
gif |
MP Paragon XP/S 150
gif
Photo 00425-95 00424-95
HiPPI nodes on Paragons
P6 (Pentium Pro) testingXP/S 150 passes acceptance test |
NASA DAAC at ORNL | Cumulvs introduced
DONIO improves Large Scale Seismic Application I/O by factor of 25
Performance testing of Giganet's ATM/OC-12 interface on Paragon
MPI-1 specifications
HPF and C++ cross-complilers installed |
PCCM2
T170 run continuously to one year
Nanotechnology
CMD run on over 10,000,000 atoms on XP/S 35
LSMS run up to 1024 atoms on XP/S 150
|
| 1996 |
R&D 100 for GigaNet OC-12C board
jpg,
jpg
TORC (Tennessee-Oak Ridge Cluster) using Linux/NT, Miranet/ATM
SGI Onyx Reality Engine 2 for VizLab
SGI Onyx Infinite Reality (4 processor SMP) for VizLab
XP/S 5 upgraded to XP/S 10 |
IBM 3495 tape library replaced with two IBM 3494's
Over 3 TB in UniTree |
Secure PVM
PVM with IP multicast
Beta testing the Chen-1000 server
Secure Online Facilities
gif
HFIR project
Electronic Notebooks
Electron Microscope Project
Metacomputing project wins Gold medal for Concurrency at SC96
Steering team wins Silver medal for Innovation at SC96
PSTSWM shallow water code becomes part of ParkBench
|
BES Award for Outstanding Sci.
Accomp. in M&C (
GMR)
XP/S 35 & XP/S 150 linked for runs of LSMS
04412-96
Native ATM over ESnet between ORNL and SNL Paragons
CalTech/JPL world record for synthetic forces (10,000)
linked runs of LSMS over ATM on 512 nodes each of ORNL/SNL Paragons
Linked runs of LSMS on ORNL Paragon and PSC Cray t3D
Multi-tasking speeds up TORT by 4.5 on 8 processor Cray Y-MP
2-d decomposition of CCM3
new cartesian algorithm parallelized for climate
0424-96
coupled models of climate systems in heterogeneous environment
|
| 1997 |
IBM SP2 upgraded to 24 nodes
Stone SouperComputer announced
Photo 127422-97
KSR1-64 donated to The Computer Museum at Moffett Field |
HPSS replaces UniTree (5 million files, 8 TB moved)
HPSS wins R&D 100 award
|
MPI-2 specificationsAVS/Eigen interface used for LSMS
CoVise installed for VizLab DB2 installed |
linked run of CTH on ORNL/SNL Paragons
MCNP (LANL's premiere stochastic radiation transport code)
ported to XP/S 150 and run on 1024 nodes
linked run of CTH over ATM using PVM on 2048 nodes (1024 each)
of ORNL/SNL paragons
CalTech/JPG sets another world record for synthetic forces (50,000)
|
| 1998 |
Stone SouperComputer up to 126 nodes
127387-97
SRC procurement announced
Benchmarking on T0 (T-zero),
C-plant, and
Tera |
HPSS/ARM up to 3 million files, 12 TB |
HarnessAtlas |
Protein structure prediction
Large scale genome analysis
Advanced models for CCM3
M4GC adds CLM to improve spin dynamics
regional climate science |
| FUTURE | hardware?
opto-electronic neurocomputing
Quantum communications | storage? | tools? |
applications? |