A collection of SciNet images. Click on the titles to see the images full-size, or the images to find out what story they came from.
The SciNet supercomputer is used to investigate how antimicrobial agents function in molecular detail, a key step toward rationally designing new antibiotic drugs that work in similar ways.
Flow over a galaxy moving upwards through a galaxy cluster [Credit: Jonathan Dursi, CITA/SciNet, University of Toronto]
SciNet Analysist Scott Northup works with users on issues of Parallel I/O [Credit: HPCS2010, SciNet, University of Toronto]
The team at their booth at University College
A partial view of the SciNet datacentres, one of the most energy efficient in Canada. [Credit: SciNet, University of Toronto]
High performance computing is being used to study how a drug candidate, scyllo-inositol, may combat neurodegeneration in Alzheimer’s patients. SciNet’s computing power is crucial to the investigation of small-molecule therapeutics for the treatment of Alzheimer’s and related neurodegenerative diseases.
The nearby Virgo galaxy cluster, showing inferred magnetic field orientations. [Credit: Jonathan Dursi, CITA/SciNet, University of Toronto]
University of Toronto to Acquire Canada’s Most Powerful Supercomputer from IBM
Gas circling the Milky Way’s supermassive black hole is roiled by magnetic turbulence, and is spared the fate of being accreted into its centre. [Credit: Ue-Li Pen, University of Toronto]
Infiniband cabling [Credit: SciNet, University of Toronto]
[Credit: SciNet, University of Toronto]
Cooling pipes at the SciNet datacentre [credit: SciNet]
A still from a highly chemically complex simulation of biofuel combustion, the largest ever performed. This simulation will help make biofuel combustion in aeronautics reliable, clean, and economical. [Credit: S. Dworkin, MIE, University of Toronto]
SciNet CTO Chris Loken with SOSCIP’s BG/Q Supercomputers
Internal cables (Credit: SciNet, University of Toronto)
ESO telescope image (left) courtesy of ESO; Alpha Centauri B data from Nature
Observations, Modelled Observations, and Simulations of galaxies moving through a magnetic field [Credit: Jonathan Dursi, SciNet/CITA, University of Toronto]
CTO Software Danny Gruner at the SciNet Data Centre
A proton-proton collision event in the CMS experiment producing two high-energy photons (red towers). This is CERN we would expect to see from the decay of a Higgs boson but it is also consistent with background Standard Model physics processes. © CERN 2012
Networking cables at the SciNet data centre [credit: SciNet]
Chris Loken, CTO SciNet, leads a tour of HPC professionals through the energy-efficient SciNet Datacentre in 2010
SOSCIP’s BG/Q Supercomputers near final installation at SciNet’s data centre
Gravity waves emitted after the black hole merger. See also a movie of the wave generation on Dr. Pfeiffer’s webpage at http://www.cita.utoronto.ca/~pfeiffer/images/0093g_800x600.mpg . [Credit: Harald Pfeiffer, CITA, University of Toronto]
The ACT telescope [Credit: The ACT Project]
FIG. 3. (Color online) Time frequency spectra (left) for scheme I for two pump-probe time delays tdel = 5.0 fs (a) and tdel = 5.2 fs (b) and free time evolution (no probe pulse) of electron velocity (right) following excitation by the pump pulse corresponding to scheme I. From “Monitoring attosecond dynamics of coherent electron-nuclear wave packets by molecular high-order-harmonic generation”. Timm Bredtmann, Szczepan Chelkowski, and Andre ́ D. Bandrauk, Phys Rev A, 84 021401
However, in high dimension, the neighborhood becomes exponentially large, and one requires an exponential number of training examples to cover it. To cover and discriminate among N regions in input space, one would need O(N) examples with a local learning algorithm, but N can grow to the power fo the space dimension.
Cable arms [Credit: SciNet, University of Toronto]
Back of compute racks showing cabling and fans. [Credit: SciNet]
SOSCIP BG/Qs operating at SciNet
The spring-like elasticity of skin, lungs, blood vessels, and uterine tissue is imparted by the protein elastin. Large-scale computer simulations of elastin are providing the molecular insight necessary for the design of elastin-based biomaterials. [Credit: Pomes Group, Sick Kids Hospital/University of Toronto]
Some of the students here are cabling up what will be the “Goliath” cluster, a cluster of 3 old Pentium-4 desktops with 100Mb ethernet.