No abstract

Generalized Sparse Matrix-Matrix Multiplication (SpGEMM) is a ubiquitous task in various engineering and scientific applications. However, inner product based SpGEMM introduces redundant input fetches for mismatched nonzero operands, while outer product based approach [1] suffers from poor output locality due to numerous partial product matrices. Inefficiency in the reuse of either inputs or outputs data leads to extensive and expensive DRAM access.To address this problem, this paper proposes an efficient sparse matrix multiplication accelerator architecture, SpArch, which jointly optimizes the data locality for both input and output matrices. We first design a highly parallelized streamingbased merger to pipeline the multiply and merge stage of partial matrices so that partial matrices are merged on chip immediately after produced. We then propose a condensed matrix representation that reduces the number of partial matrices by three orders of magnitude and thus reduces DRAM access by 5.4×. We further develop a Huffman tree scheduler to improve the scalability of the merger for larger sparse matrices, which reduces the DRAM access by another 1.8×. We also resolve the increased input matrix read induced by the new representation using a row prefetcher with near-optimal buffer replacement policy, further reducing the DRAM access by 1.5×. Evaluated on 20 benchmarks, SpArch reduces the total DRAM access by 2.8× over previous state-of-the-art. On average, SpArch achieves 4×, 19×, 18×, 17×, 1285× speedup and 6×, 164×, 435×, 307×, 62× energy savings over OuterSPACE, MKL, cuSPARSE, CUSP, and ARM Armadillo, respectively.

The role of COREX shaft furnace plays is similar with that of lump zone in blast furnace, but there are still some differences existed between them, especially the burden distribution. However, the research on burden distribution in COREX shaft furnace is quite few. Therefore, a three dimensional model is established in present work based on Discrete Element Method (DEM), after validated by industrial experiments, the model is used to investigate the burden profiles and distribution along radius in the upper part of COREX shaft furnace, the studied parameters include the distributor angle, rotating speed and length, and the stock line. The results show that the distributor angle affects the burden profile most, stock line is next, while rotating speed and length of distributor are least. The small particles segregate more intensively than large and medium particles. Additionally, it is better to reduce the distributor length and stock line in order to obtain a stable burden distribution along radius.

Currently, the reliable prognostic biomarkers for WHO grade II diffuse astrocytomas (DA) are still limited. We investigated the relations between the level of 5-Hydroxymethylcytosine (5hmC), an oxidated production of 5-methylcytosine (5mC) by the ten eleven translocated (TET) enzymes, and clinicopathological features of glioma patients. With an identified anti-5hmC antibody, we performed immunohistochemistry in 287 glioma cases. We detected that 5hmC variably reduced in most gliomas and 5hmC reduction was closely associated with higher pathological grades and shortened survival of glioma patients. In multivariate analysis, 5hmC had no independent prognostic value in the entire patient cohort. However, multivariate analysis within subtypes of gliomas revealed that 5hmC was still a prognostic marker confined to DA. In addition, we detected that IDH1 mutation by DNA sequencing was associated with favorable survival within DA. Lastly, we detected that the combination of 5hmC/KI67 was a useful prognostic marker for restratification of DA.

No abstract