StFX Scholar (St. Francis Xavier University)
Not a member yet
55051 research outputs found
Sort by
Letter from Donald MacGregor to his brother, September 18, 1945
Donald updates his brother on his and his family's health, and contemplates whether emigration to Nova Scotia would be feasible, economically
Desertification trends following the cessation of anthropogenic carbon dioxide (CO2)
Desertification remains a global issue, particularly in arid and semi-arid regions, where changes in atmospheric carbon dioxide (CO2) influence plant-water dynamics. This study examines the impact of CO2 stabilization and decline on desertification risk using climate model data from the Zero Emissions Commitment Model Intercomparison Project. Data from multiple Earth system models were analyzed under two scenarios: the 1pctCO2 experiment, which simulates a gradual CO2 increase followed by stabilization, and businessas-usual scenarios with continued emissions. Results indicate that while CO2 fertilization initially increases plant water use efficiency, a subsequent decline in CO2 concentration increases water stress, particularly in dryland regions. The study further reveals that differences in various variables between the two scenarios may affect ecosystem stability. These findings suggest that desertification risk may become a more prevalent issue even after emissions cease, highlighting the need for adaptive land management strategies in a post-emissions world
Decoding the style of Jack DeJohnette : an analysis of his approach to the drums
Jack DeJohnette has contributed to the development of jazz drumming as well as the evolution of this musical idiom itself. His open and melodic approach to time, ear-catching solo vocabulary, and versatility with a vast variety of contemporary jazz styles make him a significant figurehead for our era of jazz drumming. This thesis will analyze the development of his style through analytical transcriptions and musical concepts to cover several aspects of his playing and highlight the context and effect of said qualities. This thesis will also explore the open and interactive style of playing that DeJohnette employs. It will trace his influences and philosophies, through the idiosyncrasies of his playing, and discusses how his piano playing possibly influences his approach to drumming. DeJohnette’s career from its earliest days to the present has included collaborations with several renowned musicians such as: Miles Davis, Charles Lloyd, Jackie McLean, Bill Evans, Keith Jarrett, John Abercrombie, Kenny Wheeler, Dave Holland, Gary Peacock, and Pat Metheny. Key recordings will be highlighted in a select discography to outline specific points in his career as well as his creativity and foothold in the jazz tradition
An adaptive QoE-driven DVFS: A comprehensive approach to enhancing user experience and optimizing power across concurrent application workloads
Dynamic Voltage and Frequency Scaling (DVFS) is one of the most common techniques used for manipulating the frequency of a processor according to the needs of the application workload that is being run on the said system by the user. Common techniques often try to address this challenge by matching the frequency of the processor to the demands of the system, often in terms of utilization seen when the applications are run. The main goal here is to save power and, in turn, reduce heat generation while optimizing the frequency with which the processor runs in order to accommodate the workload. These techniques often make sure that the performance of the application workload is not being compromised in the search for saving power, reducing temperature, or the additional operating frequency of the processor. One familiar term that is used while addressing the performance of the applications is QoE (Quality of Experience). Recent practices, including the most advanced frameworks like GearDVFS and zTT, often focus only on frames per second (FPS) as QoE quantifying metrics in their experiments. While FPS is a strong indicator of user experience for certain applications like video playback or gaming, it does not adequately reflect or generalize the user experience requirements across the wider range of interactive application workloads. For example, to a user communicating using Skype or Zoom, maintaining the lowest achievable latency and constant jitter is more important to ensure his Quality of Experience (QoE) than FPS. Or, for a person wanting to book the flight tickets on his mobile phone, we need to prioritize quick page loading times and the least time-to-first-byte (TTFB) rather than the visual quality of the page with FPS. In the above two of many real-time scenarios, we see that FPS does not play any critical role in defining any rich or valuable QoE for the users. For a user who is running and switching multiple Google Tabs, page loading time and latency are quite essential QoE quantifiers compared to others. As discussed, in scenarios where multiple applications are simultaneously run and all of them have a completely different set of QoE requirements and metrics, this complexity requires a more QoE-driven DVFS that should dynamically prioritize and strike a perfect balance between the different QoE requirements of all the applications that are being run on the system. One more scenario where both GearDVFS and zTT fall short is when dealing with multiple concurrently running applications that each have distinct and complex QoE needs, and both of them do not keep the user’s interests in mind while coming up with a governor that meets this criterion. For instance, zTT relies on a lightweight Deep Q-network structure and a specified target of 30 frames per second, to optimize the QoE of a single, active foreground application. However, it is not at all adaptable enough to handle the demands of even more than one application at once.
To address this challenge, our approach proposes an advanced DVFS mechanism that incorporates and upholds an extended set of QoE metrics tailored to different application types. By dynamically profiling and weighting these metrics, our DVFS model aims to maintain optimal performance across diverse and concurrent application types. Furthermore, our model uses techniques to dynamically take into account the user-specific requirements in terms of the priority of each application and each of the QoE metrics in an application while trying to maximize and enhance their personal experience, instead of the one-size-fits-all approach. Power efficiency is also a very critical aspect of our DVFS design, which is often undermined by agnostic approaches like GearDVFS and schedutil , which push the CPUs to maintain arbitrary utilization levels (e.g., 80%) or by aggressive scaling strategies in OnDemand governors. These methods waste energy by either over-allocating resources or reacting excessively to transient demands. Our model overcomes this by dynamically adjusting CPU and GPU frequencies based on application and user-specific QoE metrics, weights, and real-time priorities. This ensures a fine-grained balance between power consumption and performance, avoiding unnecessary energy expenditure while enhancing user satisfaction across diverse and concurrent workloads. This thesis and research presents an adaptive DVFS mechanism that departs from all the static policy-based traditional optimization techniques and introduces a novel adaptive and user-trainable mechanism where a user will be able to train, store, and deploy multiple configurations of DVFS solutions based on different scenarios and user- and context-specific requirements in terms of power and performance. Each configuration that the user trains and stores can be deployed to specific scenarios that target a certain performance along with power thresholds. This ensures that the users will be able to pull and deploy a DVFS strategy that is quite relevant to the current scenario and achieves their performance requirement along with power conservation for that particular context. This approach empowers the user to call the terms on the power and performance tradeoff according to his needs while tackling the scenario. This makes every user empowered and gives way for them to customize a DVFS solution of their own based on their use case scenarios and requirements. This also enables all the users to manage their energy consumption and application performances proactively and represents a significant departure from all the traditional one-size-fits-all models and paves a way for more personalized, intelligent, and context-aware DVFS solutions. The user will be able to define their own objectives for QoE, like the target for each metric, the weight for each metric among the metrics, and the priority of the application among concurrently running active applications. Based on these user-customized inputs, each user will be able to train, store, and deploy a multitude of DVFS models, all of which are tailored to distinct power and performance expectations from the user. These models can be very easily retrieved and deployed with very minimal user intervention. By empowering each user to dictate how energy and performance tradeoffs are balanced based on each of their contexts, the framework represents a significant departure from traditional static policy-based frameworks and innovates a new paradigm for intelligent and personalized DVFS management. This level of granularity and efficiency cannot be attained with static and traditional DVFS policies and thus highlights the transformative potential of power-performance modeling with user personalization.
Furthermore, the proposed governor was able to achieve significant power conservation as compared to the standard Linux baseline governors and has demonstrated superior power and performance efficiency when compared to the state-of-the-art zTT governor. The proposed governor also has shown robust adaptability by successfully retraining and deploying specific models based on changing user-defined QoE targets and power thresholds and thereby showcasing its ability in enabling dynamic optimization across diverse usage contexts and empowering users to meet specific performance-power tradeoffs with minimal intervention
Letter from John Menzies to John McGregor, March 7, 1811
John Menzies is providing an account of financial matters related to the estate of the recipient's brother, Captain Roberts. He details a balance of £579.8.1 ¾ due to the estate, with £540.17.3 ½ accounted for by agents in London after deductions. John McGregor is set to receive 1/5 of the £579.8.1 3/4
Age and petrogenesis of Li pegmatites in southern Newfoundland
Canada’s efforts to develop low-carbon technologies has increased the demand for critical
minerals, including lithium. Southern Newfoundland is a region displaying favourable
geological conditions to host lithium pegmatites, particularly in the aureoles of voluminous,
geochemically evolved plutonic rocks. The present research is focussed on the Killick
pegmatite field, a swarm of lithium-cesium-tantalum (LCT) pegmatites discovered in 2021 in
southern Newfoundland. Fieldwork focussed on mapping and sampling multiple
spodumene-bearing pegmatite dikes. The dykes intruded metasiltstone, meta-quartz
arenite, and metavolcanic rocks of the Dolman Cove formation. Due to the homogeneity of
the >10 dykes in the Killick swarm, three dykes form the focus of petrographic and mineral
chemistry data collection (Kraken, East and Hockey Stick dykes). The dykes are of variable
thickness with irregular widths ranging from <1 up to <5m. The textures in the Kraken dyke
are mostly equigranular and dykes show no internal zoning, but the East and Hockey Stick
dykes have irregular zoning patterns including layered aplite zones in the contact with the
host rock and coarser-grained cores with spodumene up to 15cm in length. Most plagioclase
grains are albite, which likely formed due to the alteration of K-feldspar. Muscovite is
present as both primary and secondary muscovite. Spodumene is acicular in shape, pale
green, up to 15 cm in length, and it formed during the late magmatic stage. In some cases,
the spodumene is altered to clay minerals. The garnet grains are Mn-rich spessartine and is
either magmatic or xenocrystic. The tourmalines are acicular, schorl in composition, and
mainly present in the contact zones of the pegmatites with the host rocks. The columbite –
tantalite minerals are mostly manganocolumbite to manganotantalite that are rich in HFSE.
Apatite, monazite, and beryl are also present in minor proportions. The high Li and Cs
content in the dykes and the presence of spodumene, spessartine, and columbite-tantalite
group minerals indicate that the pegmatites are highly fractionated. Taken together, these
new data provide a better understanding of the magmatic-hydrothermal processes that
resulted in the emplacement of LCT pegmatites in southern Newfoundland. The pegmatites
may be related by a remelting or anatexis of nearby, voluminous, geochemically evolved
two-mica granites (Peter Snout and Rose Blanche plutons) or originated by the remelting of
the metasedimentary host rock
Building mathematical partnerships: an immigrant father’s perceptions of his child's elementary mathematics learning experiences
There is widespread agreement among education researchers that establishing partnerships among families, educators, and schools benefits students significantly. However, many families face barriers, such as work schedules or differing communication styles, that prevent full engagement in these partnerships. This is particularly true for immigrant families, especially immigrant fathers, who have historically been excluded from educational partnerships. Research indicates that many parents perceive a stark contrast between the ways they learned mathematics in school and how their children are taught, often contributing to feelings of disconnection between home and school. This study aimed to explore immigrant fathers’ experiences with educational partnerships from a mathematics perspective, specifically examining their perceptions of their child’s elementary mathematics experiences. Two theoretical frameworks guided this investigation: Pirie and Kieren’s (1994a) Theory for the Dynamical Growth of Mathematical Understanding (P-K Theory) and Moll et al.'s (1992) concept of Funds of Knowledge (FoK). These frameworks provided a lens through which to collect, view, and analyze the data. The study involved two interviews with a newcomer father living in an urban setting in Nova Scotia, Canada
Exploring The Immigration Experience And Learning Process: A Case Study Of Arab Women In Rural Nova Scotia
In this vastly different part of the world from where they came from, Arab Muslim women have experienced a unique journey marked by challenges, achievements, and growth. Conducted in the heart of rural Nova Scotia, Canada, this research seeks to investigate unexplored areas, examining how migration and resettlement in a rural town impacted the learning processes of Arab immigrant women, particularly in terms of transformative learning. This is a qualitative study based on in-depth interviews with five Arab women who arrived in Canada with their families as privately sponsored refugees. The research first examines the impact of the migration journey on Arab women and the challenges they faced upon arriving in Canada such as feelings of isolation, absence of extended family, language barriers, and adjustment to an entirely new cultural identity as Arabs in a Western society. Additionally, the study reveals the strategies they used to overcome racial and religious discrimination. The research also explores whether their integration process included transformative learning. Finally, the study examines the significance of informal learning in the women’s lives and how it facilitated access to formal education credentials. The findings demonstrate that over the years, immigrant Arab women underwent numerous changes and adjustments in their skills and ways of thinking without feeling coerced. They were motivated by the recognition that these changes, although gradual and challenging, were essential for their personal and professional development and for the well-being of their families
Na gaisgich òga = The young heroes : a youth Gaelic mentorship program in Nova Scotia
This thesis examines Na Gaisgich Òga/The Young Heroes, a youth Gaelic mentorship program in Nova Scotia. Launched in 2013, the program integrates five key elements important to successful language revitalization: community-based initiatives, immersion-based learning, mentorship, intergenerational learning, and cultural learning. Through a questionnaire of past participants, the study assesses the program’s impact on language acquisition, cultural engagement, and community involvement. Findings highlight the importance of mentorship and community connections in language retention, while also identifying areas for improvement, such as expanding post-program engagement and providing continued support for learners. By exploring other international revitalization initiatives and examining contrasts and comparisons to Na Gaisgich Òga, the thesis offers valuable insights into the program’s strengths and areas for future development. Recommendations include strengthening mentorship relationships, increasing immersion opportunities, and fostering a stronger sense of community. This research contributes to ensuring the long-term sustainability of Gaelic in Nova Scotia
Letter from Malcolm McGregor to Malcolm McGregor, August 23, 1831
In this letter to his friend of the same name, Malcolm McGregor mentions his plans to travel to North America in the the coming months. He mentions the health and welfare of various family members and friends, including John McGregor, D McGregor, John Stewart, Duncan McGregor. The low price of livestock is discussed