Skip directly to content

Coming dissertations at Uppsala university

  • Rethinking Dynamic Instruction Scheduling and Retirement for Efficient Microarchitectures Author: Mehdi Alipour Link: Publication date: 2020-02-27 10:37

    Out-of-order execution is one of the main micro-architectural techniques used to improve the performance of both single- and multi-threaded processors. The application of such a processor varies from mobile devices to server computers. This technique achieves higher performance by finding independent instructions and hiding execution latency and uses the cycles which otherwise would be wasted or caused a CPU stall. To accomplish this, it uses scheduling resources including the ROB, IQ, LSQ and physical registers, to store and prioritize instructions.

    The pipeline of an out-of-order processor has three macro-stages: the front-end, the scheduler, and the back-end. The front-end fetches instructions, places them in the out-of-order resources, and analyzes them to prepare for their execution. The scheduler identifies which instructions are ready for execution and prioritizes them for scheduling. The back-end updates the processor state with the results of the oldest completed instructions, deallocates the resources and commits the instructions in the program order to maintain correct execution.

    Since out-of-order execution needs to be able to choose any available instructions for execution, its scheduling resources must have complex circuits for identifying and prioritizing instructions, which makes them very expansive, therefore, limited. Due to their cost, the scheduling resources are constrained in size. This limited size leads to two stall points respectively at the front-end and the back-end of the pipeline. The front-end can stall due to fully allocated resources and therefore no more new instructions can be placed in the scheduler. The back-end can stall due to the unfinished execution of an instruction at the head of the ROB which prevents other resources from being deallocated, preventing new instructions from being inserted into the pipeline.

    To address these two stalls, this thesis focuses on reducing the time instructions occupy the scheduling resources. Our front-end technique tackles IQ pressure while our back-end approach considers the rest of the resources. To reduce front-end stalls we reduce the pressure on the IQ for both storing (depth) and issuing (width) instructions by bypassing them to cheaper storage structures. To reduce back-end stalls, we explore how we can retire instructions earlier, and out-of-order, to reduce the pressure on the out-of-order resource.

  • Prognostic signficance of tumor cell markers in diffuse large B-cell lymphoma with special emphasis on lymphoma localization Author: Maysaa Abdulla Link: Publication date: 2020-02-27 09:02

    Diffuse large B-cell lymphoma (DLBCL) is the most common type of high-grade B-cell lymphoma with different clinical, morphological, immunophenotypical, and molecular features. DLBCL is curable in 60-70% of patients when treated with standard immunochemotherapy R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone).

    The main aim of this thesis is to identify prognostic factors in DLBCL by studying tumor markers (paper I and II), site of disease (paper III) and tumor microenvironment markers in primary DLBCL of the CNS (PCNSL) (paper IV) in order to better identify different risk groups of DLBCL patients.

    In papers I-III, we studied DLBCL patients treated homogeneously with R-CHOP. The negative prognostic impact of double protein expression of MYC and BCL2 so called “double-expressor lymphoma” (DEL) was a common finding in the three papers. In paper I, we detected DEL in 27% of patients, distributed with no significant difference between the germinal center derived B-cell subgroup (GCB) in 52% of cases and the non-GCB subgroup in 37% of cases. There was no significant difference in survival between GCB and non-GCB patients. The diagnosis in most of the patients with DEL was made on core needle biopsy in this paper. This finding was more thoroughly investigated in paper III with attention paid to the site of biopsy. In paper II, we evaluated the concordance of cell of origin (COO) assignment between gene expression profile (GEP) and immunohistochemistry (IHC) to identify the best predictor of survival in a DLBCL cohort including patients from Sweden and Denmark. The overall concordance between the two methods was 83%. We found that ABC/non-GCB subtype identified by both GEP and IHC is associated with the worst outcome. This finding indicates the importance of precise risk stratification in the era of precision medicine. DEL was more common in ABC patients categorized by GEP. In paper III, we identified abdominal lymph node involvement by radiological examination in 63% of DLBCL patients with an inferior survival, adverse clinical characteristics and significantly more frequent DEL. These findings may indicate a distinct biological behavior in patients with abdominal nodal disease. In paper IV, we demonstrated a significant association between IDO1 and PD-L1 in PCNSL patients. This finding indicates the crucial immunosuppressive role of these two molecules. In addition, in PCNSL low frequencies of MYC and BCL2 translocations and high frequency of BCL6 translocation was observed and DEL was detected in 49% of cases. Contrary to our results in systemic DLBCL in papers I-III, there was no significant prognostic impact of DEL in PCNSL.

  • Alternative back contacts for CZTS thin film solar cells Author: Sven Englund Link: Publication date: 2020-02-26 13:50

    In this thesis, alternative back contacts for Cu2ZnSnS4 (CZTS) thin film solar cells were investigated. Back contacts for two different configurations were studied, namely traditional single-junction cells with opaque back contacts and transparent back contacts for possible use in either tandem or bifacial solar cell configuration.

    CZTS is processed under chemically challenging conditions, such as high temperature and high chalcogen partial pressure. This places great demands on the back contact. Mo is the standard choice as back contact, but reacts with chalcogens to form MoS(e)2 while the CZTS decomposes, mainly into detrimental secondary phases. Thin MoS(e)2 is assumed to be beneficial for the electrical contact, but excessive thickness is detrimental to solar cell performance. The back contact acts as diffusion medium for Na during annealing when soda-lime glass is used as substrate. Na influences both defect passivation and doping in CZTS and increases the efficiency of the solar cells. The ability of the back contact to facilitate Na diffusion is an important property that must be monitored.

    Titanium nitride (TiN) as an interlayer between the opaque molybdenum (Mo) and CZTS as well as complete replacement of Mo with TiN back contacts were investigated. TiN was found to be chemically stable in typical anneal conditions. Formation of MoS(e)2 was observed only in areas where the TiN interlayers did not fully cover the Mo, following from the surface roughness of Mo and insufficient step-coverage of the sputter-deposition of TiN. Thick TiN interlayers (200 nm) were found to increase the diffusion of Na to the absorber layer from the glass substrate. For precursors annealed in sulfur atmosphere, improved device efficiency was observed for increased TiN thickness.

    Transparent back contacts can be used in either tandem configurations where two or more absorber materials are used to more efficiently use different parts of the solar spectra, or in bifacial solar cells to allow light to reach the absorber layer from two sides and thus increase the photocurrent. Thus far only a few studies have investigated transparent back contact materials in CZTS solar cell devices. Antimony-doped tin oxide (ATO) was studied as a transparent back contact for CZTS. Annealing of bare ATO resulted in complete reaction with S to form Sn–S compounds. When annealed below the CZTS, ATO was found to be stable at low temperature (<550 °C), and in some aspects even improved its properties. ATO back contacts resulted in significantly increased formation of Sn–S secondary phases on the CZTS absorber surface compared to the Mo reference. Sn–S secondary compounds on the absorber surface made it challenging to obtain good device performance. Adhesion and device behavior could be improved by pre-addition of NaF on the precursor prior to annealing.