Muhammet Mustafa Ozdal

Muhammet Mustafa Ozdal

Auto-generated profile, based on publicly available information

Muhammet Mustafa Ozdal

Meta Platforms, Inc., Menlo Park, USA

Research Focus & Metrics

Academic Career Length

22Years (2005-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
19 May 2017 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Closed access
AAAndrey Ayupov
Andrey Ayupov
SYSerif Yesil
Serif Yesil
MOMuhammet Mustafa Ozdal
Muhammet Mustafa Ozdal
TKTaemin Kim
Taemin Kim
SBSteven Burns
Steven Burns
OOOzcan Ozturk
Ozcan Ozturk
Graph applications have been gaining importance in the last decade due to emerging big data analytics problems such as Web graphs, social networks, and biological networks. For these applications, traditional CPU and GPU architectures suffer in terms of performance and power consumption due to irregular communications, random memory accesses, and load balancing problems. It has been shown that specialized hardware accelerators can achieve much better power and energy efficiency compared to the general purpose CPUs and GPUs. In this paper, we present a template-based methodology specifically targeted for hardware accelerator design of big-data graph applications. Important architectural features that are key for energy efficient execution are implemented in a common template. The proposed template-based methodology is used to design hardware accelerators for different graph applications with little effort. Compared to an application-specific high-level synthesis methodology, we show that the proposed methodology can generate hardware accelerators with up to 18× better energy efficiency and requires less design effort.

5 November 2012
Verified
Conference Paper
Open access
PDF file
MOMuhammet Mustafa Ozdal
Muhammet Mustafa Ozdal
RHRenato Fernandes Hentschke
Renato Fernandes Hentschke

20 June 2025
Verified
Conference Paper
Open access
WCWeiwei Chu
Weiwei Chu
XXXinfeng Xie
Xinfeng Xie
JYJiecao Yu
Jiecao Yu
JWJie Wang
Jie Wang
APAmar Phanishayee
Amar Phanishayee
JPJong-Soo Park
Jong-Soo Park

22 February 2005
Verified
Conference Paper
Closed access
MOM.M. Ozdal
M.M. Ozdal
MWM.D.F. Wong
M.D.F. Wong
As the clock frequencies used in industrial applications increase, the timing requirements on routing problems become tighter, and current routing tools can not successfully handle these constraints any more. We focus on the high-performance single-layer bus routing problem, where the objective is to match the lengths of all nets belonging to each bus. An effective approach to solve this problem is to allocate extra routing resources around short nets during routing; and use those resources for length extension afterwards. We first propose a provably optimal algorithm for routing nets with min-area max-length constraints. Then, we extend this algorithm to the case where minimum constraints are given as exact length bounds. We also prove that this algorithm is optimal within a constant factor. Both algorithms proposed are also shown to be scalable for large circuits, since the respective time complexities are O(A) and O(A log A), where A is the area of the intermediate region between chips.

20 April 2023 in Future Generation Computer Systems
Verified
Journal Article
Closed access
SGSelcuk Gulcan
Selcuk Gulcan
MOMuhammet Mustafa Ozdal
Muhammet Mustafa Ozdal
CACevdet Aykanat
Cevdet Aykanat

19 March 2019 in IEEE Transactions on Parallel and Distributed Systems
Verified
Journal Article
Closed access
MOMuhammet Mustafa Ozdal
Muhammet Mustafa Ozdal
Memory access is known to be the main bottleneck for shared-memory parallel graph applications especially for large and irregular graphs. Propagation blocking (PB) idea was proposed recently to improve the parallel performance of PageRank and sparse matrix and vector multiplication operations. The idea is based on separating parallel computation into two phases, binning and accumulation, such that random memory accesses are replaced with contiguous accesses. In this paper, we propose an algorithm that allows execution of these two phases concurrently. We propose several improvements to increase parallel throughput, reduce memory overhead, and improve work efficiency. Our experimental results show that our proposed algorithms improve shared-memory parallel throughput by a factor of up to 2× compared to the original PB algorithms. We also show that the memory overhead can be reduced significantly (from 170 percent down to less than 5 percent) without significant degradation of performance. Finally, we demonstrate that our concurrent execution model allows asynchronous parallel execution, leading to significant work efficiency in addition to throughput improvements.