John Paul Shen

This paper explores Speculative Precomputation, a technique that uses idle thread context in a multithreaded architecture to improve performance of single-threaded applications. It attacks program stalls from data cache misses by pre-computing future memory accesses in available thread contexts, and prefetching these data. This technique is evaluated by simulating the performance of a research processor based on the Itanium™ ISA supporting Simultaneous Multithreading. Two primary forms of Speculative Precomputation are evaluated. If only the non-speculative thread spawns speculative threads, performance gains of up to 30% are achieved when assuming ideal hardware. However, this speedup drops considerably with more realistic hardware assumptions. Permitting speculative threads to directly spawn additional speculative threads reduces the overhead associated with spawning threads and enables significantly more aggressive speculation, overcoming this limitation. Even with realistic costs for spawning threads, speedups as high as 169% are achieved, with an average speedup of 76%.

EGFR-activating mutations are observed in approximately 15% to 20% of patients with non-small cell lung cancer. Tyrosine kinase inhibitors have provided an illustrative example of the successes in targeting oncogene addiction in cancer and the role of tumor-specific adaptations conferring therapeutic resistance. The compound osimertinib is a third-generation tyrosine kinase inhibitor, which was granted full FDA approval in March 2017 based on targeting EGFR T790M resistance. The compound has received additional FDA approval as first-line therapy with improvement in progression-free survival by suppressing the activating mutation and preventing the rise of the dominant resistance clone. Drug development has been breathtaking in this space with other third-generation compounds at various stages of development: rociletinib (CO-1686), olmutinib (HM61713), nazartinib (EGF816), naquotinib (ASP8273), mavelertinib (PF-0647775), and AC0010. However, therapeutic resistance after the administration of third-generation inhibitors is complex and not fully understood, with significant intertumoral and intratumoral heterogeneity. Repeat tissue and plasma analyses on therapy have revealed insights into multiple mechanisms of resistance, including novel second site EGFR mutations, activated bypass pathways such as MET amplification, HER2 amplification, RAS mutations, BRAF mutations, PIK3CA mutations, and novel fusion events. Strategies to understand and predict patterns of mutagenesis are still in their infancy; however, technologies to understand synthetically lethal dependencies and track cancer evolution through therapy are being explored. The expansion of combinatorial therapies is a direction forward targeting minimal residual disease and bypass pathways early based on projected resistance.

The worldwide burden of sickle cell disease is enormous, with over 200,000 infants born with the disease each year in Africa alone. Induction of fetal hemoglobin is a validated strategy to improve symptoms and complications of this disease. The development of targeted therapies has been limited by the absence of discrete druggable targets. We developed a unique bead-based strategy for the identification of inducers of fetal hemoglobin transcripts in primary human erythroid cells. A small-molecule screen of bioactive compounds identified remarkable class-associated activity among histone deacetylase (HDAC) inhibitors. Using a chemical genetic strategy combining focused libraries of biased chemical probes and reverse genetics by RNA interference, we have identified HDAC1 and HDAC2 as molecular targets mediating fetal hemoglobin induction. Our findings suggest the potential of isoform-selective inhibitors of HDAC1 and HDAC2 for the treatment of sickle cell disease.