Nastaran Hajinazar

Nastaran Hajinazar

Auto-generated profile, based on publicly available information

Nastaran Hajinazar

Intel, Hillsboro, USA

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
17 December 2019 in Proceedings of the ACM on Measurement and Analysis of Computing Systems
Verified
Journal Article
Open access
SGSaugata Ghose
Saugata Ghose
TLTianshi Li
Tianshi Li
NHNastaran Hajinazar
Nastaran Hajinazar
DCDamla Senol Cali
Damla Senol Cali
OMOnur Mutlu
Onur Mutlu
It has become increasingly difficult to understand the complex interactions between modern applications and main memory, composed of Dynamic Random Access Memory (DRAM) chips. Manufacturers are now selling and proposing many different types of DRAM, with each DRAM type catering to different needs (e.g., high throughput, low power, high memory density). At the same time, memory access patterns of prevalent and emerging applications are rapidly diverging, as these applications manipulate larger data sets in very different ways. As a result, the combined DRAM-workload behavior is often difficult to intuitively determine today, which can hinder memory optimizations in both hardware and software. In this work, we identify important families of workloads, as well as prevalent types of DRAM chips, and rigorously analyze the combined DRAM-workload behavior. To this end, we perform a comprehensive experimental study of the interaction between nine different DRAM types and 115 modern applications and multiprogrammed workloads. We draw 12 key observations from our characterization, enabled in part by our development of new metrics that take into account contention between memory requests due to hardware design. Notably, we find that (1) newer DRAM technologies such as DDR4 and HMC often do not outperform older technologies such as DDR3, due to higher access latencies and, also in the case of HMC, poor exploitation of locality; (2) there is no single memory type that can effectively cater to all of the components of a heterogeneous system (e.g., GDDR5 significantly outperforms other memories for multimedia acceleration, while HMC significantly outperforms other memories for network acceleration); and (3) there is still a strong need to lower DRAM latency, but unfortunately the current design trend of commodity DRAM is toward higher latencies to obtain other benefits. We hope that the trends we identify can drive optimizations in both hardware and software design. To aid further study, we open-source our extensively-modified simulator, as well as a benchmark suite containing our applications.

28 October 2023
Verified
Conference Paper
Open access
KKKonstantinos Kanellopoulos
Konstantinos Kanellopoulos
RBRahul Bera
Rahul Bera
KSKosta Stojiljkovic
Kosta Stojiljkovic
FBF. Nisa Bostanci
F. Nisa Bostanci
CFCan Firtina
Can Firtina
OMOnur Mutlu
Onur Mutlu
Conventional virtual memory (VM) frameworks enable a virtual address to flexibly map to any physical address. This flexibility necessitates large data structures to store virtual-to-physical mappings, which leads to high address translation latency and large translation-induced interference in the memory hierarchy, especially in data-intensive workloads. On the other hand, restricting the address mapping so that a virtual address can only map to a specific set of physical addresses can significantly reduce address translation overheads by making use of compact and efficient translation structures. However, restricting the address mapping flexibility across the entire main memory severely limits data sharing across different processes and increases data accesses to the swap space of the storage device even in the presence of free memory. We propose Utopia, a new hybrid virtual-to-physical address mapping scheme that allows both flexible and restrictive hash-based address mapping schemes to harmoniously co-exist in the system. The key idea of Utopia is to manage physical memory using two types of physical memory segments: restrictive segments and flexible segments. A restrictive segment uses a restrictive, hash-based address mapping scheme that maps virtual addresses to only a specific set of physical addresses and enables faster address translation using compact translation structures. A flexible segment employs the conventional fully-flexible address mapping scheme. By mapping data to a restrictive segment, Utopia enables faster address translation with lower translation-induced interference. At the same time, Utopia retains the ability to use the flexible address mapping to (i) support conventional VM features such as data sharing and (ii) avoid storing data in the swap space of the storage device when program data does not fit inside a restrictive segment. Our evaluation using 11 diverse data-intensive workloads shows that Utopia improves performance by 24% in a single-core system over the baseline conventional four-level radix-tree page table design, whereas the best prior state-of-the-art contiguity-aware translation scheme improves performance by 13%. Utopia provides 95% of the performance benefits of an ideal address translation scheme where every translation request hits in the first-level TLB. All of Utopia’s benefits come at a modest cost of 0.64% area overhead and 0.72% power overhead compared to a modern high-end CPU. The source code of Utopia is freely available at https://github.com/CMU-SAFARI/Utopia.

17 February 2021 in bioRxiv
Verified
Preprint
Open access
PDF file
JKJeremie S. Kim
Jeremie S. Kim
CFCan Firtina
Can Firtina
MCMeryem Banu Cavlak
Meryem Banu Cavlak
DCDamla Senol Cali
Damla Senol Cali
NHNastaran Hajinazar
Nastaran Hajinazar
OMOnur Mutlu
Onur Mutlu
As genome sequencing tools and techniques improve, researchers are able to incrementally assemble more accurate reference genomes, which enable sensitivity in read mapping and downstream analysis such as variant calling. A more sensitive downstream analysis is critical for a better understanding of the genome donor (e.g., health characteristics). Therefore, read sets from sequenced samples should ideally be mapped to the latest available reference genome that represents the most relevant population. Unfortunately, the increasingly large amount of available genomic data makes it prohibitively expensive to fully re-map each read set to its respective reference genome every time the reference is updated. There are several tools that attempt to accelerate the process of updating a read data set from one reference to another (i.e., remapping) by 1) identifying regions that appear similarly between two references and 2) updating the mapping location of reads that map to any of the identified regions in the old reference to the corresponding similar region in the new reference. The main drawback of existing approaches is that if a read maps to a region in the old reference that does not appear with a reasonable degree of similarity in the new reference, the read cannot be remapped. We find that, as a result of this drawback, a significant portion of annotations (i.e., coding regions in a genome) are lost when using state-of-the-art remapping tools. To address this major limitation in existing tools, we propose AirLift, a fast and comprehensive technique for remapping alignments from one genome to another. Compared to the state-of-the-art method for remapping reads (i.e., full mapping), AirLift reduces 1) the number of reads (out of the entire read set) that need to be fully mapped to the new reference by up to 99.99% and 2) the overall execution time to remap read sets between two reference genome versions by 6.7×, 6.6×, and 2.8× for large (human), medium (C. elegans), and small (yeast) reference genomes, respectively. We validate our remapping results with GATK and find that AirLift provides similar accuracy in identifying ground truth SNP and INDEL variants as the baseline of fully mapping a read set. Code Availability AirLift source code and readme describing how to reproduce our results are available at https://github.com/CMU-SAFARI/AirLift.

17 October 2025
Verified
Conference Paper
Open access
KZKaiyang Zhao
Kaiyang Zhao
YCYuang Chen
Yuang Chen
XXXenia Xu
Xenia Xu
DSDan Schatzberg
Dan Schatzberg
NHNastaran Hajinaza
Nastaran Hajinaza
DSDimitrios Skarlatos
Dimitrios Skarlatos