Search NASA⌕ Search

SEARCH · Search NASA

Results for “High efficiency acceleration”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,765 records · Page 98

Muon Collider Progress: Accelerators

A muon collider would be a powerful tool for exploring the energy-frontier with leptons, and would complement the studies now under way at the LHC. Such a device would offer several important benefits. Muons, like electrons, are point particles so the full center-of-mass energy is available for particle production. Moreover, on account of their higher mass, muons give rise to very little synchrotron radiation and produce very little beamstrahlung. The first feature permits the use of a circular collider that can make efficient use of the expensive rf system and whose footprint is compatible with an existing laboratory site. The second feature leads to a relatively narrow energy spread at the collision point. Designing an accelerator complex for a muon collider is a challenging task. Firstly, the muons are produced as a tertiary beam, so a high-power proton beam and a target that can withstand it are needed to provide the required luminosity of ~1 × 10{sup 34} cm{sup –2}s{sup –1}. Secondly, the beam is initially produced with a large 6D phase space, which necessitates a scheme for reducing the muon beam emittance (“cooling”). Finally, the muon has a short lifetime so all beam manipulations must be done very rapidly. The Muon Accelerator Program, led by Fermilab and including a number of U.S. national laboratories and universities, has undertaken design and R&D activities aimed toward the eventual construction of a muon collider. Design features of such a facility and the supporting R&D program are described.

43 PARTICLE ACCELERATORS↗

Muon Acceleration: Neutrino Factory and Beyond

We summarize the current state of a concept for muon acceleration aimed at a future Neutrino Factory and extendable to Higgs Factory. The main thrust of these studies was to reduce the overall cost while maintaining performance by exploring the interplay between the complexity of the cooling systems and the acceptance of the accelerator complex. To ensure adequate survival for the short-lived muons, acceleration must occur at high average gradient. The need for large transverse and longitudinal acceptances drives the design of the acceleration system to an initially low RF frequency, e.g., 325 MHz, which is then increased to 650 MHz as the transverse size shrinks with increasing energy. High-gradient normal conducting RF cavities at these frequencies require extremely high peak-power RF sources. Hence superconducting RF (SRF) cavities are chosen. We consider an SRF-efficient design based on a multi-pass (4.5) ?dogbone? RLA, extendable to multi-pass FFAG-like arcs.

Bogacz, Alex↗

Crab Waist Collision at DAFNE

DAFNE is an accelerator complex consisting of a double ring lepton collider working at the c.m. energy of the {Phi}-resonance (1.02 GeV) and an injection system. In its original configuration the collider consisted of two independent rings, each {approx}97 m long, sharing two 10 m long interaction regions (IR1 and IR2) where the KLOE and FINUDA or DEAR detectors were respectively installed. A full energy injection system, including an S-band linac, 180 m long transfer lines and an accumulator/damping ring, provides fast and high efficiency electron positron injection also in topping-up mode during collisions. Recently the DAFNE collider has been upgraded in order to implement a new collision scheme based on large Piwinski angle and cancellation of the synchro-betatron resonances by means of electromagnetic sextupoles (Crab-Waist compensation). The novel approach has proved to be effective in improving beam-beam interaction and collider luminosity.

43 PARTICLE ACCELERATORS↗

High-Energy Pulsar Models: Developments and New Questions

The past few years have seen a major advance in observational knowledge of high-energy (HE) pulsars. The Fermi Large Area Telescope (LAT) and AGILE have increased the number of known gamma-ray pulsars by an order of magnitude, its members being divided roughly equally among millisecond pulsars (MSPs), young radio-loud pulsars, and young radio-quiet pulsars. Many new and diverse emission characteristics are being measured, while radio and X-ray follow-up observations increase the pulsar detection rate and enrich our multiwavelength picture of these extreme sources. The wealth of new data has provided impetus for further development and improvement of existing theoretical pulsar models. Geometric light curve (LC) modelling has uncovered three broad classes into which HE pulsars fall: those where the radio profile leads, is aligned with, or lags the gamma-ray profile. For example, the original MSP and original black widow system are members of the second class, requiring co-located emission regions and thereby breaking with traditional notions of radio emission origin. These models imply narrow accelerator gaps in the outer magnetosphere, indicating copious pair production even in MSP magnetospheres that were previously thought to be pair-starved. The increased quality and variety of the LCs necessitate construction of ever more sophisticated models. We will review progress in global magnetosphere solutions which specify a finite conductivity on field lines above the stellar surface, filling the gap between the standard vacuum and force-free (FF; plasma-filled) models. The possibility of deriving phase-resolved spectra for the brightest pulsars, coupled with the fact that the HE pulsar population is sizable enough to allow sampling of various pulsar geometries, will enable much more stringent testing of future radiation models. Reproduction of the observed phase-resolved behavior of this disparate group will be one of the next frontiers in pulsar science, impacting on our understanding of particle acceleration, emission, and magnetosphere geometry. One may now also study evolutionary trends of the measured or inferred quantities, and probe pulsar visibility and population properties such as radiation beam sizes of different pulsar classes, as well as the distribution of spin-down power, gamma-ray luminosity, conversion efficiency, spectral index, and cutoff energy across the population. Lastly, the recent detection of very-high-energy (VHE) pulsations from the Crab pulsar generated quite a few ideas to explain this emission, leading to an extension of standard models and possibly even a bridge between the physics of pulsars and pulsar wind nebulae (PWNe).

gamma rays:↗

Optimization of Isentropic Compression Loads on Current-Adder Pulsed Power Accelerator Architectures

The Thor pulsed power generator is being developed at Sandia National Laboratories. The design consists of up to 288 decoupled and transit time isolated capacitor-switch units, called “bricks”, that can be individually triggered to achieve a high degree of pulse tailoring for magnetically-driven isentropic compression experiments (ICE). The connecting transmission lines are impedance matched to the bricks, allowing the capacitor energy to be efficiently delivered to an ICE strip-line load with peak pressures of over 100 GPa. Thor will drive experiments to explore equation of state, material strength, and phase transition properties of a wide variety of materials. We present an optimization process for producing tailored current pulses, a requirement for many material studies, on the Thor generator. This technique, which is unique to the novel “current-adder” architecture used by Thor, entirely avoids the iterative use of complex circuit models to converge to the desired electrical pulse. We describe the optimization procedure for the Thor design and show results for various materials of interest. Also, we discuss the extension of these concepts to the megajoule-class Neptune machine design. Given this design, we are able to design shockless ramp-driven experiments in the 1 TPa range of material pressure.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Computing the Properties of Matter with Leadership Computing Resources (Closeout Report for DE-SC0018121)

In order to add more capabilities to Halide, we have designed a new framework called Tiramisu and integrated this framework into Halide. Since Tiramisu enables Halide to target heterogeneous architectures, our development efforts have been refocused on Tiramisu. Most high-performance computer systems today are complex and increasingly heterogeneous; they may have CPUs, GPUs and FPGAs. Achieving best performance requires taking full advantage of all these different architectures. To address this issue, we have designed Tiramisu, an optimization framework that enables Halide (and other DSLs) to target heterogeneous architectures. Tiramisu is an optimization framework that takes as input a high level, architecture-independent representation of code and a set of scheduling and data mapping commands that guide code transformation. The input can either be generated by a domain-specific language (DSL) compiler such as Halide or directly written by a programmer. Tiramisu then applies the user-specified code and data-layout transformations and generates an architecture-specific, low-level intermediate representation (IR) that takes advantage of modern architectural features such as multicore parallelism, non-uniform memory (NUMA) hierarchies, clusters, and accelerators like GPUs and FPGAs. We integrated Tiramisu within Halide and implemented a representative set of benchmarks to evaluate this integration. Tiramisu is now open source and is available for public use (http://tiramisu-compiler.org/). A paper about Tiramisu was published, it shows that Tiramisu extends Halide with many new capabilities and that Tiramisu can generate efficient code for multicores, GPUs, FPGAs and distributed heterogeneous systems. The performance of code generated by the Tiramisu backends matches or exceeds hand optimized reference implementations. For example, the multicore backend matches the highly optimized Intel MKL library on many kernels and shows speedups reaching 4x over the original Halide. In addition to making Tiramisu more robust, we have used Tiramisu to implement a set of representative tensor operation for constructing baryon building blocks required for multi baryon contractions in LQCD. In order to implement this code, we needed to generalize Tiramisu in two ways: first we needed to support indirect array accesses, and second, we needed to add support for complex numbers to Tiramisu. The code generated by Tiramisu is 6x faster than the reference code. Our efforts towards an MPI based multi-node version of tiramisu have matured and the resulting code scales well on multiple nodes (tests up to 512 KNL nodes have been undertaken).

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Investigation of a metallic photonic crystal high power microwave mode converter

It is demonstrated that an L band metallic photonic crystal TEM-TE{sub 11} mode converter is suitable for narrow band high power microwave application. The proposed mode converter is realized by partially filling metallic photonic crystals along azimuthal direction in a coaxial transmission line for phase-shifting. A three rows structure is designed and simulated by commercial software CST Microwave Studio. Simulation results show that its conversion efficiency is 99% at the center frequency 1.58 GHz. Over the frequency range of 1.56-1.625 GHz, the conversion efficiency exceeds 90 %, with a corresponding bandwidth of 4.1 %. This mode converter has a gigawatt level power handling capability which is suitable for narrow band high power microwave application. Using magnetically insulated transmission line oscillator(MILO) as a high power microwave source, particle-in-cell simulation is carried out to test the performance of the mode converter. The expected TE{sub 11} mode microwave output is obtained and the MILO works well. Mode conversion performance of the converter is tested by far-field measurement method. And the experimental result confirms the validity of our design. Then, high power microwave experiment is carried out on a Marx-driven Blumlein water line pulsed power accelerator. Microwave frequency, radiated pattern and power are measured in the far-field region and the results agree well with simulation results. The experiment also reveals that no microwave breakdown or pulse shortening took place in the experimental setup.

36 MATERIALS SCIENCE↗

DIFFUSE HARD X-RAY EMISSION IN STARBURST GALAXIES AS SYNCHROTRON FROM VERY HIGH ENERGY ELECTRONS

The origin of the diffuse hard X-ray (2-10 keV) emission from starburst galaxies is a long-standing problem. We suggest that synchrotron emission of 10-100 TeV electrons and positrons (e {sup {+-}}) can contribute to this emission, because starbursts have strong magnetic fields. We consider three sources of e {sup {+-}} at these energies: (1) primary electrons directly accelerated by supernova remnants, (2) pionic secondary e {sup {+-}} created by inelastic collisions between cosmic ray (CR) protons and gas nuclei in the dense interstellar medium of starbursts, and (3) pair e {sup {+-}} produced between the interactions between 10 and 100 TeV {gamma}-rays and the intense far-infrared (FIR) radiation fields of starbursts. We create one-zone steady-state models of the CR population in the Galactic center (R {<=} 112 pc), NGC 253, M82, and Arp 220's nuclei, assuming a power-law injection spectrum for electrons and protons. We consider different injection spectral slopes, magnetic field strengths, CR acceleration efficiencies, and diffusive escape times, and include advective escape, radiative cooling processes, and secondary and pair e {sup {+-}}. We compare these models to extant radio and GeV and TeV {gamma}-ray data for these starbursts, and calculate the diffuse synchrotron X-ray and inverse Compton (IC) luminosities of these starbursts in the models which satisfy multiwavelength constraints. If the primary electron spectrum extends to {approx}PeV energies and has a proton/electron injection ratio similar to the Galactic value, we find that synchrotron emission contributes 2%-20% of their unresolved, diffuse hard X-ray emission. However, there is great uncertainty in this conclusion because of the limited information on the CR electron spectrum at these high energies. IC emission is likewise a minority of the unresolved X-ray emission in these starbursts, from 0.1% in the Galactic center to 10% in Arp 220's nuclei, with the main uncertainty being the starbursts' magnetic field. We also model generic starbursts, including submillimeter galaxies, in the context of the FIR-X-ray relation, finding that anywhere between 0% and 16% of the total hard X-ray emission is synchrotron for different parameters, and up to 2% in the densest starbursts assuming an E {sup -2.2} injection spectrum and a diffusive escape time of 10 Myr (E/3 GeV){sup -1/2} (h/100 pc). Neutrino observations by IceCube and TeV {gamma}-ray data from HESS, VERITAS, and CTA can further constrain the synchrotron X-ray emission of starbursts. Our models do not constrain the possibility of hard, second components of primary e {sup {+-}} from sources like pulsars in starbursts, which could enhance the synchrotron X-ray emission further.

79 ASTRONOMY AND ASTROPHYSICS↗

Demonstration of low-level biogenic fuel content using quench curve and direct liquid scintillation counting (LSC) methods

In this work, an investigation into the direct method for liquid scintillation counting (LSC) quantification of biogenic fuel content in fuel blendstock is presented. This method development intentionally used a colored matrix (~99 wt%), fossil diesel fraction, and low-level biogenic fractions (~1 wt%) to determine the applicability of the LSC technique in colored blend mixtures. LSC procedures for quench correction, including the use of an instrument internal standard (Quantulus SQP values), color quench curves, and an internal 14 C standard spike on samples containing low-level amounts of biogenic gasoline, jet, and diesel mixed with fossil fuel, were compared. Samples were analyzed on both Perkin-Elmer Quantulus and Tri-Carb instruments, and the 14 C contents were compared with those determined by accelerator mass spectrometry (AMS). Results from the Tri-Carb instrument showed that percent differences of <10 % compared to AMS-reported values are achievable at 5-hour count times despite colored samples. A comparison between the impact of chemical and color quenching in the fuel samples showed that color quenching in highly colored samples reduces efficiency significantly more than chemical quenching, indicating that chemical quench curves are not appropriate for highly colored biofuel/fossil-fuel blended samples. Results indicate that both the direct method with internal spike quench correction and the use of a color-quench curve provide accurate results for 1 % biogenic fuel blends. Additionally, we have explored the use of Fourier transform infrared (FTIR) spectroscopy in measuring the biogenic content of the colored blended fuel samples. Preliminary results show the presence of biogenic carbon-derived fingerprints that are absent in fossil-derived sample. However, further work is needed to develop an FTIR-based quantitative method.

09 BIOMASS FUELS↗

Complexity in the Photofunctionalization of Single-Wall Carbon Nanotubes with Hypochlorite

The reaction of aqueous suspensions of single-wall carbon nanotubes (SWCNTs) with UV-excited sodium hypochlorite has previously been reported to be an efficient route for doping nanotubes with oxygen atoms. Here, we have investigated how this reaction system is affected by pH level, dissolved O 2 content, and radical scavengers and traps. Products were characterized with near-IR fluorescence, Raman, and XPS spectroscopy. The reaction is greatly accelerated by removal of dissolved O 2 and strongly suppressed by TEMPO, a radical trap. Alcohols added as radical scavengers alter the reaction efficiency and the product peak emission wavelengths. Photofunctionalization with 300 nm irradiation is substantially less efficient at pH levels low enough to protonate the OCl – ion to HOCl. We deduce that in mildly treated high pH samples, the main product is sp 2 hybridized O-doped adducts formed by reaction of SWCNTs with atomic oxygen in its 3 P (ground) level. By contrast, treatment under low pH conditions leads to sp 3 hybridized SWCNT adducts formed by the addition of secondary radicals from reactions of • OH and • Cl. There is also evidence for additional photoreactions of product species under stronger irradiation. Researchers using photoexcited hypochlorite for SWCNT functionalization should be alert to the range of products and the sensitivity to reaction conditions in this system.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Biexciton and trion energy transfer from CdSe/CdS giant nanocrystals to Si substrates

Observation of energy transfer (ET) from multiexcitonic (MX) complexes in nanocrystal quantum dots (NQDs) has been severely restricted due to efficient nonradiative Auger recombination leading to very low MX emission quantum yields. Here we employed “giant” CdSe/CdS NQDs with suppressed Auger recombination to study ET of biexcitons (BX) and charged excitons (trions) into Si substrate. Photoluminescence (PL) measurements of (sub)monolayers of gNQDs controllably assembled on various interacting surfaces and augmented by single gNQD's imaging reveal appearance of BX spectral signatures and progressive acceleration of PL lifetimes of all excitonic species on Si substrates. From statistical analysis of a large number of PL lifetime traces, representative exciton, trion and BX ET efficiencies are measured as ~75%, 55% and 45% respectively. Detailed analysis of the MX's radiative rates demonstrate the crucial role of the radiative (waveguide) ET in maintaining high overall transfer efficiency despite the prevalent Auger recombination. Furthermore, our observations point towards practical utilization of MX-bearing nanocrystals in future optoelectronics architectures.

36 MATERIALS SCIENCE↗

Charge state breeding experiences and plans at TRIUMF

At the Isotope Separation and ACceleration (ISAC) facility at TRIUMF, an electron cyclotron resonance ion source (ECRIS) has been set up for the charge state breeding of radioactive ions. In order to reduce background from stable ions generated in the ECRIS, several measures, including changing materials for the plasma chamber and the surrounding components, have been implemented. Further reduction has been achieved by using the post-accelerator chain as a mass filter. Since the implementation of those measures in 2013, physics experiments with accelerated radioactive isotopes of Rb, Sr, K, and Mg have been performed. In most cases, a charge breeding efficiency of several percent has been achieved. With the planned expansion of the isotope production capabilities at TRIUMF within the Advanced Rare IsotopE Laboratory project, two new target stations, one using photo-fission induced by a high-power electron beam at 50 MeV and the other one using 480 MeV protons as at ISAC, will be put into operation within the next 5 yr. Additionally, a new electron beam ion source (EBIS) based charge state breeding system will be installed. Background from such a source is expected to be much lower. The drawback is that for the efficient operation of such a system, pulsed beam operation is required, which makes the installation of an additional ion buncher in front of the EBIS necessary.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Insights into collaborative separation process of photogenerated charges and superior performance of solar cells

ZnO nanowires/Cu{sub 4}Bi{sub 4}S{sub 9} (ZnO/CBS) and ZnO nanowires/CBS-graphene nanoplates (ZnO/CBS-GNs), as well as two types of solar cells were prepared. The photovoltaic responses of CBS-GNs and ZnO/CBS-GNs can be improved with incorporation of GNs. The transient surface photovoltage (TPV) can provide detailed information on the separation and transport of photogenerated carriers. The multichannel separation process from the TPVs indicates that the macro-photoelectric signals can be attributed to the photogenerated charges separated at the interface of CBS/GNs, rather than CBS/ZnO. The multi-interfacial recombination is the major carrier loss, and the hole selective p-V{sub 2}O{sub 5} can efficiently accelerate the charge extraction to the external circuit. The ZnO/CBS-GNs cell exhibits the superior performance, and the highest efficiency is 10.9%. With the adequate interfaces of CBS/GNs, GNs conductive network, energy level matching, etc., the excitons can easily diffuse to the interface of CBS/GNs, and the separated electrons and holes can be collected quickly, inducing the high photoelectric properties. Here, a facile strategy for solid state solar cells with superior performance presents a potential application.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Development of a bi-solvent liquid scintillator with slow light emission

One of the most promising approaches for the next generation of neutrino experiments is the realization of large hybrid Cherenkov/scintillation detectors made possible by recent innovations in photodetection technology and liquid scintillator chemistry. The development of a potentially suitable future detector liquid with particularly slow light emission is discussed in the present publication. This cocktail is compared with respect to its fundamental characteristics (scintillation efficiency, transparency, and time profile of light emission) with liquid scintillators currently used in large-scale neutrino detectors. In addition, the optimization of the admixture of wavelength shifters for a scintillator with particularly high light emission is presented. Furthermore, the pulse-shape discrimination capabilities of the novel medium was studied using a pulsed particle accelerator driven neutron source. Beyond that, purification methods based on column chromatography and fractional vacuum distillation for the co-solvent DIN (Diisopropylnaphthalene) are discussed.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Development of an accurate EPID-based output measurement and dosimetric verification tool for electron beam therapy

Purpose: To develop an efficient and robust tool for output measurement and absolute dose verification of electron beam therapy by using a high spatial-resolution and high frame-rate amorphous silicon flat panel electronic portal imaging device (EPID). Methods: The dosimetric characteristics of the EPID, including saturation, linearity, and ghosting effect, were first investigated on a Varian Clinac 21EX accelerator. The response kernels of the individual pixels of the EPID to all available electron energies (6, 9, 12, 16, and 20 MeV) were calculated by using Monte Carlo (MC) simulations, which formed the basis to deconvolve an EPID raw images to the incident electron fluence map. The two-dimensional (2D) dose distribution at reference depths in water was obtained by using the constructed fluence map with a MC simulated pencil beam kernel with consideration of the geometric and structural information of the EPID. Output factor measurements were carried out with the EPID at a nominal source–surface distance of 100 cm for 2 × 2, 3 × 3, 6 × 6, 10 × 10, and 15 × 15 cm{sup 2} fields for all available electron energies, and the results were compared with that measured in a solid water phantom using film and a Farmer-type ion chamber. The dose distributions at a reference depth specific to each energy and the flatness and symmetry of the 10 × 10 cm{sup 2} electron beam were also measured using EPID, and the results were compared with ion chamber array and water scan measurements. Finally, three patient cases with various field sizes and irregular cutout shapes were also investigated. Results: EPID-measured dose changed linearly with the monitor units and showed little ghosting effect for dose rate up to 600 MU/min. The flatness and symmetry measured with the EPID were found to be consistent with ion chamber array and water scan measurements. The EPID-measured output factors for standard square fields of 2 × 2, 3 × 3, 6 × 6, 10 × 10, 15 × 15 cm{sup 2} agreed with film and ion chamber measurements. The average discrepancy between EPID and ion chamber/film measurements was 0.81% ± 0.60% (SD) and 1.34% ± 0.75%, respectively. For the three clinical cases, the difference in output between the EPID- and ion chamber array measured values was found to be 1.13% ± 0.11%, 0.54% ± 0.10%, and 0.74% ± 0.11%, respectively. Furthermore, the γ-index analysis showed an excellent agreement between the EPID- and ion chamber array measured dose distributions: 100% of the pixels passed the criteria of 3%/3 mm. When the γ-index was set to be 2%/2 mm, the pass rate was found to be 99.0% ± 0.07%, 98.2% ± 0.14%, and 100% for the three cases. Conclusions: The EPID dosimetry system developed in this work provides an accurate and reliable tool for routine output measurement and dosimetric verification of electron beam therapy. Coupled with its portability and ease of use, the proposed system promises to replace the current film-based approach for fast and reliable assessment of small and irregular electron field dosimetry.

60 APPLIED LIFE SCIENCES↗

Space-Charge Effects in Ionization Beam Profile Monitors

Ionization profile monitors (IPMs) are widely used in accelerators for non-destructive and fast diagnostics of high energy particle beams. At high beam intensities, the space-charge forces make the measured IPM profiles significantly different from those of the beams. We analyze dynamics of the secondaries in IPMs and develop an effective algorithm to reconstruct the beam sizes from the measured IPM profiles. Efficiency of the developed theory is illustrated in application to the Fermilab 8 GeV proton Booster IPMs.

Shiltsev, Vladimir D.↗

Parity Violating Deep Inelastic Electron Scattering from the Deuteron at 6 GeV

An experiment that measured the parity violating (PV) asymmetry A d in e- 2 H deep inelastic scattering (DIS) at Q 2 ~ 1.10 and 1.90 (GeV/c) 2 and x B ~ 0.3 was completed in experimental Hall A at the Thomas Jefferson National Accelerator Facility. The asymmetry can be used to extract the neutral weak coupling combination (2C 2u -C 2d ), providing a factor of five to six improvement over the current world data. To achieve this precision, asymmetries of the 10 -4 level needed to be measured at event rates up to 500 kHz with high electron detection efficiency and high pion background rejection capability. A specialized scaler-based counting data acquisition system (DAQ) with hardware-based particle identification was successfully implemented. The statistical quality of the asymmetry measurement agreed with the Gaussian distribution to over five orders of magnitudes and the experimental goal of 3-4% statistical uncertainty was achieved. The design and performance of the new DAQ system is presented with the preliminary asymmetry results given in the end.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

LDMS-GPU: Lightweight Distributed Metric Service (LDMS) for NVIDIA GPGPUs

GPUs are now a fundamental accelerator for many high-performance computing applications. They are viewed by many as a technology facilitator for the surge in fields like machine learning and Convolutional Neural Networks. To deliver the best performance on a GPU, we need to create monitoring tools to ensure that we optimize the code to get the most performance and efficiency out of a GPU. Since NVIDIA GPUs are currently the most commonly implemented in HPC applications and systems, NVIDIA tools are the solution for performance monitoring. The Light-Weight Distributed Metric System (LDMS) at Sandia is an infrastructure widely adopted for large-scale systems and application monitoring. Sandia has developed CPU application monitoring capability within LDMS. Therefore, we chose to develop a GPU monitoring capability within the same framework. In this report, we discuss the current limitations in the NVIDIA monitoring tools, how we overcame such limitations, and present an overview of the tool we built to monitor GPU performance in LDMS and its capabilities. Also, we discuss our current validation results. Most of the performance counter results are the same in both vendor tools and our tool when using LDMS to collect these results. Furthermore, our tool provides these statistics during the entire runtime of the tool as a time series and not just aggregate statistics at the end of the application run. This allows the user to see the progress of the behavior of the applications during their lifetime.

97 MATHEMATICS AND COMPUTING↗