Pith. sign in

REVIEW 2 major objections 5 minor 84 references

Near Data Acceleration with Concurrent Host Access

T0 review · 2 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read By letting host CPUs and near-data accelerators interleave fine-grained requests to the same DDR4 ranks, Chopim captures up to 97% of otherwise idle rank bandwidth while holding host performance steady, and speeds up SVRG logistic…

desk verdict Chopim's fine-grain host/NDA sharing mechanisms are original and well evaluated, but the headline results rest on an unvalidated replicated-FSM synchronization assumption that needs explicit handling of refresh and error cases. read the letter →

arxiv 1908.06362 v2 pith:RPQA5ZMS submitted 2019-08-18 cs.AR

Benjamin Y. Cho , Yongkee Kwon , Sangkug Lym , Mattan Erez This is my paper · ORCID
classification cs.AR
keywords near-dataprocessingprocessing-in-memoryDDR4bankpartitioningread/writeturnarounddatalayoutSVRGconcurrentaccess
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that near-data accelerators (NDAs) inside main memory need not be confined to their own ranks or time slices. It presents Chopim, a combined software/hardware proposal that lets the host CPU and NDAs interleave requests to the same DDR4 banks and ranks at fine granularity, capturing up to 97% of rank bandwidth the host is not using while host instructions-per-cycle stays effectively unchanged. Four mechanisms carry the result: bank partitioning that isolates host-only banks from shared banks, throttling of NDA writes to avoid read/write-turnaround penalties, a data layout that aligns NDA operands across ranks using coarse-grain allocation and OS page coloring, and replicated memory-controller state machines that keep the host and NDA controllers synchronized without explicit signaling. In an SVRG logistic-regression case study, concurrent host-NDA execution trains 2x faster than host-only execution and 1.6x faster than serialized host-NDA execution, with no data copies.

What carries the argument

The central mechanism is a bank-partitioning address remap combined with replicated memory-controller FSMs. In the remap, the most-significant physical address bits and bank bits are swapped when an address would otherwise fall into a bank reserved for shared data, so host-only addresses are always steered to host-only banks and aliasing is avoided; this works with huge pages and modern XOR-based interleaving. The replicated FSMs, synchronized by the DDR interface clock, let the host track NDA memory transactions from a deterministic access pattern, enabling non-packetized DDR4 operation without NDA-to-host signaling. Supporting these are a data layout that aligns operands to the same 'system row' across ranks using coarse-grain allocation and OS page coloring, and two write-throttling policies (next-rank prediction over a dedicated pin, and stochastic issue) that curb the read/write turnaround overhead caused by interleaved NDA writes.

What would settle it

Implement the non-packetized DDR4 mode in a cycle-accurate simulator or prototype and inject a single refresh or timing perturbation during an NDA operation: if the host-side and NDA-side FSMs diverge, transactions are lost or corrupted, refuting the synchronization claim. Alternatively, measure actual NDA bandwidth utilization on a DDR4 module with the dedicated pin wired and compare it with the simulated 97% of idle-rank-bandwidth figure; a large gap would falsify the fine-grain interleaving results.

Watch

Extended reading notes

Core claim

The discovery is that fine-grain temporal interleaving of host and NDA memory requests to the same physical DRAM devices is not only feasible but decisively better than dedicating ranks to each side or coarse time-multiplexing. Chopim achieves this by remapping addresses so that host-only memory never lands in banks reserved for shared data, by issuing NDA commands only into ranks that are momentarily idle while throttling NDA writes (via next-rank prediction or stochastic issue), and by placing all operands of an NDA instruction in the same rank through system-row-aligned allocation and physical-frame coloring. On the non-packetized DDR4 interface, the system replicates the memory-controller finite-state machine on both sides of the channel and synchronizes them with the DDR clock, so the host can predict NDA memory behavior from the (deterministic) NDA access pattern alone, without any NDA-to-host signaling. The paper reports up to 97% of unutilized rank bandwidth used by NDAs with host IPC preserved, and 2x/1.6x speedups for SVRG logistic regression versus host-only and non-concurrent baselines.

Load-bearing premise

The load-bearing premise is that the replicated memory-controller finite-state machines on the host and NDA sides stay perfectly synchronized using the DDR interface clock and deterministic NDA access patterns, with no explicit NDA-to-host signaling; if DRAM refresh, error handling, or timing uncertainty makes the states diverge, the non-packetized DDR4 mode fails, and the results also assume a dedicated pin for next-rank write prediction that is not implemented on real systems.

Editorial extensions

If this is right

  • NDAs can be integrated into commodity DDR4 DIMMs without dedicating ranks, so the full memory capacity stays available to the host even during acceleration.
  • Host and NDAs can collaborate on the same data concurrently without copying, enabling algorithm redesigns such as delayed-update SVRG that overlap phases and cut training time.
  • Write-intensive NDA operations no longer force a severe host slowdown when combined with the throttling mechanisms.
  • Scaling the system with more ranks improves NDA throughput more than rank partitioning does, because idle-rank opportunities grow with rank count.
  • The bank-partitioning and layout techniques are compatible with huge pages and advanced address interleaving, removing a practical obstacle to near-data processing adoption.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The replicated-FSM synchronization idea could generalize to other memory-side engines that have deterministic access patterns, potentially enabling concurrent host access with in-memory processing on future DRAM generations.
  • The OS/compiler interface implied by Chopim—colored allocation and system-row alignment—suggests that memory allocation policies could be exposed to programming models so more applications obtain NDA locality automatically.
  • Because the SVRG result depends on staleness, similar delayed-update variants of other iterative algorithms could benefit from concurrent host-NDA access, but their convergence guarantees would need separate analysis.
  • If the dedicated pin for next-rank prediction is unacceptable, stochastic write throttling provides a fallback but with a performance trade-off; an in-band signaling scheme could eliminate the pin while retaining robustness.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper presents Chopim, a hardware/software co-design that enables concurrent access to main memory by a host CPU and near-data accelerators (NDAs) sharing the same DRAM ranks. The main contributions are: bank partitioning to prevent host accesses from reducing NDA row-buffer locality; mechanisms to throttle NDA writes (next-rank prediction and stochastic issue) to avoid read/write turnaround penalties; a data layout that keeps NDA operands locally aligned while preserving the host's sophisticated address interleaving; and replicated memory-controller FSMs that allow the host and NDA-side controllers to coordinate in a non-packetized DDR4 interface without NDA-to-host signaling. The evaluation, based on gem5+Ramulator simulations with SPEC2006/2017 host workloads and NDA microbenchmarks, reports up to 97% of unutilized rank bandwidth utilization for NDAs with preserved host IPC, and a 2x speedup (host-only baseline) or 1.6x speedup (non-concurrent baseline) for an SVRG logistic-regression case study.

Significance. If the mechanisms work as described, Chopim addresses real and largely unexplored challenges: fine-grain temporal interleaving of host and NDA requests to shared ranks, a bank-partitioning scheme compatible with huge pages and modern address hashing, and a data layout that enables colocated host/NDA collaboration without data copies. The evaluation is reasonably thorough, covering read- and write-intensive extremes, multiple host workload mixes, a scalable rank configuration, and a realistic machine-learning application. The main weakness is that the non-packetized DDR4 mode, which underpins all of the headline results, relies on an unvalidated replicated-FSM synchronization assumption, and the next-rank prediction mechanism depends on a dedicated sideband pin that is assumed but not modeled with physical delays.

major comments (2)
  1. [III-D, VII] The replicated-FSM design is load-bearing for the entire DDR4 evaluation, yet its correctness is only argued informally. Section III-D describes the FSMs staying synchronized via the DDR interface clock and deterministic NDA access patterns, and the text states that 'all our experiments rely on this.' However, deterministic NDA access patterns specify which commands an NDA will eventually issue, not the cycle-level timing of when those commands are issued. Host priority decisions, NDA write-buffer drain thresholds, and especially JEDEC refresh scheduling are controller policies whose outcomes must be identically predicted by both FSMs. The simulator does not model refresh, ECC scrubbing, or error handling, and no re-synchronization or divergence-recovery mechanism is presented. Without a formal argument or a simulated fault-injection/sensitivity study showing that the replicated state cannot diverge under these events, the correctness of concurrent host/NDA access in non-packetized DDR4 is not established, and the claimed 97% bandwidth utilization and SVRG speedups inherit this uncertainty.
  2. [III-B, VII] The next-rank prediction mechanism used for the rest of the paper depends on a dedicated sideband pin for the inhibit signal. Section III-B states, 'For now, we assume that this information is communicated over a dedicated pin and plan to develop other signaling mechanisms that can piggyback on existing host DRAM commands at a later time.' The simulation does not model the pin's propagation delay, the decision logic latency, or the command scheduling overhead of sending the inhibit signal ahead of the host transaction. Since Section VII states that next-rank prediction is used for the rest of the paper, and Figures 12 and 15 report results based on it, the reported host-IPC and SVRG numbers could change materially once this signaling path is modeled with realistic delays. The authors should either implement a delay model for the pin and predictor or provide a sensitivity analysis over the signal latency.
minor comments (5)
  1. [III-A] In the paragraph on data layout across ranks, 'we reply on OS page coloring' should read 'we rely on OS page coloring.'
  2. [VII, Takeaway 6] Takeaway 6 says collaborative SVRG is 'sped up by 50%,' while the abstract and Section IV say 2x over host-only execution, which is a 100% improvement. Please make the reported percentage and the speedup factors consistent.
  3. [III-C] The bank-partitioning scheme's real-system feasibility depends on the OS being able to split the physical address space and enforce page coloring as described. The evaluation appears to assume this OS behavior; a short discussion or simulation of allocator-induced fragmentation or of the cost of clearing memory before repartitioning would strengthen the claim of compatibility with huge pages.
  4. [III-A, V] The paper acknowledges that the proposed per-chip word layout relies on in-DRAM ECC with limited coverage and defers server-grade ECC to future work. Since the target is main memory in a server system, this limitation should at least be stated more prominently in the contributions or in the concluding discussion.
  5. [VI] The results for the nine mixes are presented without any indication of run-to-run variability or statistical significance. Since the same mixes are used across all evaluations, a brief statement about the sensitivity of the conclusions to the specific SimPoint choices would be useful.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: Chopim's central claims are architectural mechanisms evaluated against independent baselines, with tuning confined to the SVRG case study.

full rationale

This is an engineering systems paper, not a derivation chain, and I find no step in which a claimed prediction or result reduces by construction to its own inputs. The bank-partitioning, write-throttling, replicated-FSM, and data-layout mechanisms are implemented in a gem5/Ramulator simulation and evaluated against concrete baselines: rank partitioning, idealized unutilized-rank-bandwidth, and host-only execution. The 'up to 97% of unutilized bandwidth' claim is a measured ratio against an explicitly idealized model, not a value produced by fitting that model. The SVRG case study uses 'best-tuned' learning rates and epoch sizes, but these hyperparameters tune the algorithm comparison; they are not used to fit or derive the architecture-level bandwidth or IPC results. The paper's self-citations, including the bank-partitioning prior work [36] with an overlapping author, are contextual and contrastive ('existing bank partitioning mechanisms [36], [52], [57] are incompatible...'), not load-bearing evidence for Chopim's correctness. The main caveats are assumptions rather than circularities: Section III-D assumes replicated host- and NDA-side memory-controller FSMs remain synchronized using the DDR interface clock and deterministic NDA access patterns without NDA-to-host signaling, and Section III-B assumes a dedicated pin for next-rank prediction. These are unvalidated implementation assumptions and correctness risks, but they are not reductions of the paper's conclusions to its premises. The paper is self-contained against external benchmarks and simulators, so no equation-level or self-citation-based circularity is present.

Assumptions & free parameters 5 free parameters · 7 assumptions · 2 invented entities

The central claims rest on simulator fidelity, OS page-coloring support, linear address mappings, and the replicated-FSM assumption. The paper does not ship code or hardware, so these are assumptions rather than verified artifacts.

free parameters (5)
  • Number of banks reserved for NDAs = 1 bank per rank in evaluation (2-bank example in Section III-C)
    Bank partitioning strength trades host parallelism against NDA interference; the paper selects one bank per rank without a sensitivity analysis.
  • Stochastic write issue probability = 1/4 and 1/16
    Configurable coin weight; the paper evaluates two fixed values and does not provide a selection rule.
  • Number of OS page colors = 8 colors, 4 GiB shared region each
    Coloring granularity is a design choice; it determines how many operands can share alignment.
  • System-row allocation granularity = 2 MiB for a 1 TiB DDR4 system
    Chosen to match one DRAM row per bank; affects memory fragmentation and OS allocator support.
  • SVRG hyperparameters = lambda=1e-3, momentum=0.9, learning rate 'best-tuned', epochs N, N/2, N/4
    The convergence speedup depends on hyperparameters selected per configuration; the exact best-tuned values are not reported.
assumptions (7)
  • domain assumption Ramulator and gem5 model DRAM timing, scheduling, and CPU behavior faithfully enough for the reported speedups.
    The central results are produced by simulation; first-order accuracy depends on simulator fidelity.
  • domain assumption The Intel Skylake address mapping is representative of modern interleaving, and all address mappings of interest are linear.
    Section III-A uses this mapping for the layout and coloring argument; nonlinear or hashed future mappings may break the alignment guarantee.
  • domain assumption OS page coloring and coarse-grain memory allocation can be implemented with acceptable fragmentation.
    Section III-A states fragmentation is negligible but provides no quantitative evaluation.
  • ad hoc to paper Replicated FSMs remain synchronized using the DDR interface clock and deterministic NDA access patterns, with no NDA-to-host signaling.
    Section III-D; this is the load-bearing premise for the non-packetized DDR4 mode.
  • domain assumption In-DRAM ECC with limited coverage is sufficient for NDA local accesses.
    Section III-A explicitly defers stronger ECC; if DRAM faults are a concern, the single-chip word layout is not fully protected.
  • domain assumption Coherence can be handled by explicit copies and memory fences because shared data is mostly read-only.
    Sections II and IV rely on cache bypassing and fences; a general coherence protocol would be needed for frequently written shared data.
  • domain assumption A rank cannot serve host and NDA requests simultaneously.
    Section I defines the problem this way; if simultaneous access were possible, the entire contention framework would change.
invented entities (2)
  • Replicated host and NDA memory controller FSMs
    purpose: Coordinate bank and timing state in non-packetized DDR4 without NDA-to-host signaling.
    Described in Section III-D and used in all experiments, but only simulated; no independent hardware validation exists.
  • Dedicated sideband pin for next-rank prediction inhibit signal
    purpose: Carry the host-to-NDA stall signal for write throttling.
    Section III-B assumes the pin exists and plans to develop other signaling later; no implementation is demonstrated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Near Data Acceleration with Concurrent Host Access." pith.science (2026). https://pith.science/paper/RPQA5ZMS

@misc{pith2026190806362,
  author       = {Pith},
  title        = {Pith review of: Near Data Acceleration with Concurrent Host Access},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RPQA5ZMS}},
  note         = {Machine review of arXiv:1908.06362}
}
read the original abstract

Near-data accelerators (NDAs) that are integrated with main memory have the potential for significant power and performance benefits. Fully realizing these benefits requires the large available memory capacity to be shared between the host and the NDAs in a way that permits both regular memory access by some applications and accelerating others with an NDA, avoids copying data, enables collaborative processing, and simultaneously offers high performance for both host and NDA. We identify and solve new challenges in this context: mitigating row-locality interference from host to NDAs, reducing read/write-turnaround overhead caused by fine-grain interleaving of host and NDA requests, architecting a memory layout that supports the locality required for NDAs and sophisticated address interleaving for host performance, and supporting both packetized and traditional memory interfaces. We demonstrate our approach in a simulated system that consists of a multi-core CPU and NDA-enabled DDR4 memory modules. We show that our mechanisms enable effective and efficient concurrent access using a set of microbenchmarks, and then demonstrate the potential of the system for the important stochastic variance-reduced gradient (SVRG) algorithm.

Figures

Figures reproduced from arXiv: 1908.06362 by the authors.

Figure 1
Figure 1. Exemplary NDA architecture. sulting from NDA writes that mainly interfere with the host. We provide two solutions in this context. First, we develop a new bank-partitioning scheme that limits interference to just those memory regions that are shared by the host and NDAs, thus enabling colocating host-only tasks with tasks that use the NDAs. This new scheme is the first that is compatible with huge pages and also wit… view at source ↗
Figure 2
Figure 2. Rank idle-time breakdown vs. idleness granularity. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Example data layout across ranks for concurrent access [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Baseline and proposed host-side address mapping. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Global MC state tracking when the host (left) and [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Collaboration between host and NDAs in SVRG. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Overview of NDA architecture. identify such cases and more intelligently allocate memory and regions to minimize copies. For this paper, we do not implement such a compiler. Instead, programs are written to directly interact with a runtime system that is implemented wi…
Figure 8
Figure 8. Figure 8: Average gradient example code. This code corresponds [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: PE architecture and execution flow of AXPY. [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]
Figure 10
Figure 10. Figure 10: Impact of coarse-grain NDA operations. (X-axis: the [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]
Figure 13
Figure 13. Figure 13: Impact of NDA operations and operand size. [PITH_FULL_IMAGE:figures/full_fig_p011_13.png]
Figure 14
Figure 14. Figure 14: Scalability Chopim vs. rank partitioning. [PITH_FULL_IMAGE:figures/full_fig_p011_14.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

84 extracted references · 78 canonical work pages

  1. [1]

    The design space of data-parallel memory systems

    Jung Ho Ahn, Mattan Erez, and William J Dally. The design space of data-parallel memory systems. In SC06: Proceedings of the 2006 ACM/IEEE conference on Supercomputing , 2006

    2006

  2. [2]

    A scalable processing-in-memory accelerator for parallel graph processing

    Junwhan Ahn, Sungpack Hong, Sungjoo Yoo, Onur Mutlu, and Kiyoung Choi. A scalable processing-in-memory accelerator for parallel graph processing. In 2015 ACM/IEEE 42nd Annual International Symposium on Computer Architecture (ISCA) , pages 105–117, June 2015

    2015

  3. [3]

    Pim-enabled instructions: a low-overhead, locality-aware processing- in-memory architecture

    Junwhan Ahn, Sungjoo Yoo, Onur Mutlu, and Kiyoung Choi. Pim-enabled instructions: a low-overhead, locality-aware processing- in-memory architecture. In Computer Architecture (ISCA), 2015 ACM/IEEE 42nd Annual International Symposium on , pages 336–348. IEEE, 2015

    2015

  4. [4]

    Berkin Akin, Franz Franchetti, and James C. Hoe. Data reorganization in memory using 3d-stacked dram. In Proceedings of the 42Nd Annual International Symposium on Computer Architecture , ISCA ’15, pages 131–143, New York, NY , USA, 2015. ACM

    2015

  5. [5]

    Hamlet architecture for parallel data reorganization in memory

    Berkin Akin, Franz Franchetti, and James C Hoe. Hamlet architecture for parallel data reorganization in memory. IEEE Micro, 36(1):14–23, Jan 2016

    work page 2016

  6. [6]

    Netdimm: Low-latency near- memory network interface architecture

    Mohammad Alian and Nam Sung Kim. Netdimm: Low-latency near- memory network interface architecture. In Proceedings of the 52nd An- nual IEEE/ACM International Symposium on Microarchitecture , pages 699–711. ACM, 2019

    work page 2019

  7. [7]

    Application-transparent near-memory processing architecture with memory channel network,

    Mohammad Alian, Seung Won Min, Hadi Asgharimoghaddam, Ashutosh Dhar, Dong Wang, Adam Roewer, Thomas McPadden, Oliver OHalloran, Deming Chen, Jinjun Xiong, Daehoon Kim, Wen-mei Hwu, and Nam Sung Kim. Application-transparent near-memory processing architecture with memory channel network,. In The 51st Annual IEEE/ACM International Symposium on Microarchitec...

    work page 2018

  8. [8]

    Chameleon: Versatile and practical near-dram accel- eration architecture for large memory systems

    Hadi Asghari-Moghaddam, Young Hoon Son, Jung Ho Ahn, and Nam Sung Kim. Chameleon: Versatile and practical near-dram accel- eration architecture for large memory systems. In Microarchitecture (MICRO), 2016 49th Annual IEEE/ACM International Symposium on , pages 1–13. IEEE, 2016

    work page 2016

Show all 84 references
  1. [9]

    A neural probabilistic language model

    Yoshua Bengio, R ´ejean Ducharme, Pascal Vincent, and Christian Jauvin. A neural probabilistic language model. Journal of machine learning research, 3(Feb):1137–1155, 2003

    2003

  2. [10]

    The gem5 simulator

    Nathan Binkert, Bradford Beckmann, Gabriel Black, Steven K Rein- hardt, Ali Saidi, Arkaprava Basu, Joel Hestness, Derek R Hower, Tushar Krishna, Somayeh Sardashti, et al. The gem5 simulator. ACM SIGARCH Computer Architecture News, 39(2):1–7, 2011

    2011

  3. [11]

    Google workloads for consumer devices: Mitigating data movement bottlenecks

    Amirali Boroumand, Saugata Ghose, Youngsok Kim, Rachata Ausavarungnirun, Eric Shiu, Rahul Thakur, Daehyun Kim, Aki Kuusela, Allan Knies, Parthasarathy Ranganathan, and Onur Mutlu. Google workloads for consumer devices: Mitigating data movement bottlenecks. In Proceedings of th...

    2018

  4. [12]

    Malladi, Hongzhong Zheng, and Onur Mutlu

    Amirali Boroumand, Saugata Ghose, Minesh Patel, Hasan Hassan, Bran- don Lucia, Rachata Ausavarungnirun, Kevin Hsieh, Nastaran Hajinazar, Krishna T. Malladi, Hongzhong Zheng, and Onur Mutlu. Conda: Effi- cient cache coherence support for near-data accelerators. In Proceedings of...

    2019

  5. [13]

    Lazypim: An efficient cache coherence mechanism for processing-in-memory

    Amirali Boroumand, Saugata Ghose, Minesh Patel, Hasan Hassan, Brandon Lucia, Kevin Hsieh, Krishna T Malladi, Hongzhong Zheng, and Onur Mutlu. Lazypim: An efficient cache coherence mechanism for processing-in-memory. IEEE Computer Architecture Letters , 16(1):46– 50, 2016

    2016

  6. [14]

    Regan: A pipelined reram-based accelerator for generative adversarial networks

    Fan Chen, Linghao Song, and Yiran Chen. Regan: A pipelined reram-based accelerator for generative adversarial networks. In Design Automation Conference (ASP-DAC), 2018 23rd Asia and South Pacific , pages 178–183. IEEE, 2018

    2018

  7. [15]

    Cacti-3dd: Architecture-level modeling for 3d die-stacked dram main memory

    Ke Chen, Sheng Li, Naveen Muralimanohar, Jung Ho Ahn, Jay B Brock- man, and Norman P Jouppi. Cacti-3dd: Architecture-level modeling for 3d die-stacked dram main memory. In Design, Automation & Test in Europe Conference & Exhibition (DATE), 2012 , pages 33–38. IEEE, 2012

    2012

  8. [16]

    Prime: A novel processing-in-memory architecture for neural network computation in reram-based main mem- ory

    Ping Chi, Shuangchen Li, Cong Xu, Tao Zhang, Jishen Zhao, Yongpan Liu, Yu Wang, and Yuan Xie. Prime: A novel processing-in-memory architecture for neural network computation in reram-based main mem- ory. In Proceedings of the 43rd International Symposium on Computer Architectu...

    2016

  9. [17]

    Next big thing: Ddr4 3ds

    JS Choi. Next big thing: Ddr4 3ds

  10. [18]

    Openmp: An industry-standard api for shared-memory programming

    Leonardo Dagum and Ramesh Menon. Openmp: An industry-standard api for shared-memory programming. Computing in Science & Engi- neering, 5(1):46–55, 1998

    1998

  11. [19]

    Large scale distributed deep networks

    Jeffrey Dean, Greg Corrado, Rajat Monga, Kai Chen, Matthieu Devin, Mark Mao, Andrew Senior, Paul Tucker, Ke Yang, Quoc V Le, et al. Large scale distributed deep networks. In Advances in neural informa- tion processing systems , pages 1223–1231, 2012

    2012

  12. [20]

    A white paper on the benefits of chipkill-correct ecc for pc server main memory

    Timothy J Dell. A white paper on the benefits of chipkill-correct ecc for pc server main memory. IBM Microelectronics Division , 11:1–23, 1997

    1997

  13. [21]

    The true processing in memory accelerator

    Fabrice Devaux. The true processing in memory accelerator. In 2019 IEEE Hot Chips 31 Symposium (HCS) , pages 1–24. IEEE, 2019

    2019

  14. [22]

    The mondrian data engine

    Mario Drumond, Alexandros Daglis, Nooshin Mirzadeh, Dmitrii Ustiu- gov, Javier Picorel, Babak Falsafi, Boris Grot, and Dionisios Pnev- matikatos. The mondrian data engine. In Proceedings of the 44th Annual International Symposium on Computer Architecture , pages 639–

  15. [23]

    Nda: Near-dram acceleration architecture leveraging commodity dram devices and standard memory modules

    Amin Farmahini-Farahani, Jung Ho Ahn, Katherine Morrow, and Nam Sung Kim. Nda: Near-dram acceleration architecture leveraging commodity dram devices and standard memory modules. In High Per- formance Computer Architecture (HPCA), 2015 IEEE 21st International Symposium on, page...

    2015

  16. [24]

    Practical near-data processing for in-memory analytics frameworks

    Mingyu Gao, Grant Ayers, and Christos Kozyrakis. Practical near-data processing for in-memory analytics frameworks. In Parallel Architecture and Compilation (PACT), 2015 International Conference on, pages 113–

    2015

  17. [25]

    Tetris: Scalable and efficient neural network acceleration with 3d memory

    Mingyu Gao, Jing Pu, Xuan Yang, Mark Horowitz, and Christos Kozyrakis. Tetris: Scalable and efficient neural network acceleration with 3d memory. In Proceedings of the Twenty-Second International Conference on Architectural Support for Programming Languages and Operating System...

    2017

  18. [26]

    Processing in memory: The terasys massively parallel pim array

    Maya Gokhale, Bill Holmes, and Ken Iobst. Processing in memory: The terasys massively parallel pim array. Computer, 28(4):23–31, 1995

    1995

  19. [27]

    Understanding the Linux virtual memory manager

    Mel Gorman. Understanding the Linux virtual memory manager . Prentice Hall Upper Saddle River, 2004

    2004

  20. [28]

    3d- stacked memory-side acceleration: Accelerator and system design

    Qi Guo, Nikolaos Alachiotis, Berkin Akin, Fazle Sadi, Guanglin Xu, Tze Meng Low, Larry Pileggi, James C Hoe, and Franz Franchetti. 3d- stacked memory-side acceleration: Accelerator and system design. In In the Workshop on Near-Data Processing (WoNDP)(Held in conjunction with M...

    2014

  21. [29]

    Performance implications of nocs on 3d-stacked memories: Insights from the hybrid memory cube

    Ramyad Hadidi, Bahar Asgari, Jeffrey Young, Burhan Ahmad Mudassar, Kartikay Garg, Tushar Krishna, and Hyesoon Kim. Performance implications of nocs on 3d-stacked memories: Insights from the hybrid memory cube. In 2018 IEEE International Symposium on Performance Analysis of Sys...

    2018

  22. [30]

    Simpoint 3.0: Faster and more flexible program phase analysis

    Greg Hamerly, Erez Perelman, Jeremy Lau, and Brad Calder. Simpoint 3.0: Faster and more flexible program phase analysis. Journal of Instruction Level Parallelism, 7(4):1–28, 2005

    2005

  23. [31]

    Spec cpu2006 benchmark descriptions

    John L Henning. Spec cpu2006 benchmark descriptions. ACM SIGARCH Computer Architecture News , 34(4):1–17, 2006

    2006

  24. [32]

    Accelerating linked-list traversal through near-data processing

    Byungchul Hong, Gwangsun Kim, Jung Ho Ahn, Yongkee Kwon, Hongsik Kim, and John Kim. Accelerating linked-list traversal through near-data processing. In Parallel Architecture and Compilation Tech- niques (PACT), 2016 International Conference on , 2016

    2016

  25. [33]

    Hsieh, E

    K. Hsieh, E. Ebrahim, G. Kim, N. Chatterjee, M. O’Connor, N. Vijayku- mar, O. Mutlu, and S. W. Keckler. Transparent offloading and mapping (tom): Enabling programmer-transparent near-data processing in gpu systems. In 2016 ACM/IEEE 43rd Annual International Symposium on Compute...

    2016

  26. [34]

    Accelerating pointer chasing in 3d-stacked memory: Challenges, mechanisms, evaluation

    Kevin Hsieh, Samira Khan, Nandita Vijaykumar, Kevin K Chang, Ami- rali Boroumand, Saugata Ghose, and Onur Mutlu. Accelerating pointer chasing in 3d-stacked memory: Challenges, mechanisms, evaluation. In 2016 IEEE 34th International Conference on Computer Design (ICCD) , pages ...

    2016

  27. [35]

    Eigen: C++ template library for linear algebra, 2013

    B Jacob, G Guennebaud, et al. Eigen: C++ template library for linear algebra, 2013

    2013

  28. [36]

    Balancing dram locality and parallelism in shared memory cmp systems

    Min Kyu Jeong, Doe Hyun Yoon, Dam Sunwoo, Mike Sullivan, Ikhwan Lee, and Mattan Erez. Balancing dram locality and parallelism in shared memory cmp systems. In High Performance Computer Architecture (HPCA), 2012 IEEE 18th International Symposium on , pages 1–12. IEEE, 2012

    2012

  29. [37]

    Accelerating stochastic gradient descent using predictive variance reduction

    Rie Johnson and Tong Zhang. Accelerating stochastic gradient descent using predictive variance reduction. In Advances in neural information processing systems, pages 315–323, 2013

    2013

  30. [38]

    Cacti-io: Cacti with off-chip power-area-timing models

    Norman P Jouppi, Andrew B Kahng, Naveen Muralimanohar, and Vaishnav Srinivas. Cacti-io: Cacti with off-chip power-area-timing models. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, 23(7):1254–1267, 2015

    2015

  31. [39]

    Flexram: Toward an advanced intelligent memory system

    Yi Kang, Wei Huang, Seung-Moon Yoo, Diana Keen, Zhenzhou Ge, Vinh Lam, Pratap Pattnaik, and Josep Torrellas. Flexram: Toward an advanced intelligent memory system. In Proceedings 1999 IEEE International Conference on Computer Design: VLSI in Computers and Processors (Cat. No. ...

    1999

  32. [40]

    Neurocube: A programmable digital neuromor- phic architecture with high-density 3d memory

    Duckhwan Kim, Jaeha Kung, Sek Chai, Sudhakar Yalamanchili, and Saibal Mukhopadhyay. Neurocube: A programmable digital neuromor- phic architecture with high-density 3d memory. In Computer Architec- ture (ISCA), 2016 ACM/IEEE 43rd Annual International Symposium on , pages 380–39...

    2016

  33. [41]

    Toward standardized near-data processing with unrestricted data place- ment for gpus

    Gwangsun Kim, Niladrish Chatterjee, Mike O’Connor, and Kevin Hsieh. Toward standardized near-data processing with unrestricted data place- ment for gpus. In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis , page 24. ...

    2017

  34. [42]

    Ramulator: A fast and extensible dram simulator

    Yoongu Kim, Weikun Yang, and Onur Mutlu. Ramulator: A fast and extensible dram simulator. IEEE Computer architecture letters , 15(1):45–49, 2016

    2016

  35. [43]

    Execube-a new architecture for scaleable mpps

    Peter M Kogge. Execube-a new architecture for scaleable mpps. In 1994 International Conference on Parallel Processing Vol. 1, volume 1, pages 77–84. IEEE, 1994

    1994

  36. [44]

    Processing in memory: Chips to petaflops

    Peter M Kogge, Jay B Brockman, Thomas Sterling, and Guang Gao. Processing in memory: Chips to petaflops. In Workshop on Mixing Logic and DRAM: Chips that Compute and Remember at ISCA , volume 97. Citeseer, 1997

    1997

  37. [45]

    Tensordimm: A practical near-memory processing architecture for embeddings and ten- sor operations in deep learning

    Youngeun Kwon, Yunjae Lee, and Minsoo Rhu. Tensordimm: A practical near-memory processing architecture for embeddings and ten- sor operations in deep learning. In Proceedings of the 52nd Annual IEEE/ACM International Symposium on Microarchitecture , pages 740–

  38. [46]

    Coordinated and efficient huge page management with ingens

    Youngjin Kwon, Hangchen Yu, Simon Peter, Christopher J Rossbach, and Emmett Witchel. Coordinated and efficient huge page management with ingens. In OSDI, volume 16, pages 705–721, 2016

    2016

  39. [47]

    Slow learners are fast

    John Langford, Alexander Smola, and Martin Zinkevich. Slow learners are fast. arXiv preprint arXiv:0911.0491 , 2009

    2009 arXiv

  40. [48]

    Bssync: Processing near memory for machine learning workloads with bounded staleness consistency models

    Joo Hwan Lee, Jaewoong Sim, and Hyesoon Kim. Bssync: Processing near memory for machine learning workloads with bounded staleness consistency models. In 2015 International Conference on Parallel Architecture and Compilation (PACT), pages 241–252, Oct 2015

    2015

  41. [49]

    Drisa: A dram-based reconfigurable in-situ accelerator

    Shuangchen Li, Dimin Niu, Krishna T Malladi, Hongzhong Zheng, Bob Brennan, and Yuan Xie. Drisa: A dram-based reconfigurable in-situ accelerator. In Proceedings of the 50th Annual IEEE/ACM International Symposium on Microarchitecture, pages 288–301. ACM, 2017

    2017

  42. [50]

    Pinatubo: A processing-in-memory architecture for bulk bitwise operations in emerging non-volatile memories

    Shuangchen Li, Cong Xu, Qiaosha Zou, Jishen Zhao, Yu Lu, and Yuan Xie. Pinatubo: A processing-in-memory architecture for bulk bitwise operations in emerging non-volatile memories. In Design Automation Conference (DAC), 2016 53nd ACM/EDAC/IEEE , pages 1–6. IEEE, 2016

    2016

  43. [51]

    Processing-in-memory for energy-efficient neural network train- ing: A heterogeneous approach

    Jiawen Liu, Hengyu Zhao, Matheus A Ogleari, Dong Li, and Jishen Zhao. Processing-in-memory for energy-efficient neural network train- ing: A heterogeneous approach. In 2018 51st Annual IEEE/ACM International Symposium on Microarchitecture (MICRO) , pages 655–

    2018

  44. [52]

    A software memory partition approach for eliminating bank-level interference in multicore systems

    Lei Liu, Zehan Cui, Mingjie Xing, Yungang Bao, Mingyu Chen, and Chengyong Wu. A software memory partition approach for eliminating bank-level interference in multicore systems. In Proceedings of the 21st international conference on Parallel architectures and compilation techni...

    2012

  45. [53]

    Get out of the valley: power- efficient address mapping for gpus

    Yuxi Liu, Xia Zhao, Magnus Jahre, Zhenlin Wang, Xiaolin Wang, Yingwei Luo, and Lieven Eeckhout. Get out of the valley: power- efficient address mapping for gpus. In 2018 ACM/IEEE 45th Annual International Symposium on Computer Architecture (ISCA) , pages 166–

    2018

  46. [54]

    Concurrent data structures for near-memory computing

    Zhiyu Liu, Irina Calciu, Maurice Herlihy, and Onur Mutlu. Concurrent data structures for near-memory computing. In Proceedings of the 29th ACM Symposium on Parallelism in Algorithms and Architectures , pages 235–245. ACM, 2017

    2017

  47. [55]

    Reram-based processing-in-memory architecture for recurrent neural network accel- eration

    Yun Long, Taesik Na, and Saibal Mukhopadhyay. Reram-based processing-in-memory architecture for recurrent neural network accel- eration. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, 26(12), 2018

    2018

  48. [56]

    The ibm z13 memory subsystem for big data

    Patrick J Meaney, Lawrence D Curley, Glenn D Gilda, Mark R Hodges, Daniel J Buerkle, Robert D Siegl, and Roger K Dong. The ibm z13 memory subsystem for big data. IBM Journal of Research and Development, 59(4/5):4–1, 2015

    2015

  49. [57]

    Software- hardware cooperative dram bank partitioning for chip multiprocessors

    Wei Mi, Xiaobing Feng, Jingling Xue, and Yaocang Jia. Software- hardware cooperative dram bank partitioning for chip multiprocessors. In Proceedings the IFIP International Conference on Network and Parallel Computing, 2010

    2010

  50. [58]

    Cacti 6.0: A tool to model large caches

    Naveen Muralimanohar, Rajeev Balasubramonian, and Norman P Jouppi. Cacti 6.0: A tool to model large caches. HP laboratories , pages 22–31, 2009

    2009

  51. [59]

    Graphpim: Enabling instruction-level pim offloading in graph computing frameworks

    Lifeng Nai, Ramyad Hadidi, Jaewoong Sim, Hyojong Kim, Pranith Kumar, and Hyesoon Kim. Graphpim: Enabling instruction-level pim offloading in graph computing frameworks. In High Performance Computer Architecture (HPCA), 2017 IEEE International Symposium on, pages 457–468. IEEE, 2017

    2017

  52. [60]

    Active memory cube: A processing-in- memory architecture for exascale systems

    Ravi Nair, Samuel F Antao, Carlo Bertolli, Pradip Bose, Jose R Brunheroto, Tong Chen, C-Y Cher, Carlos HA Costa, Jun Doi, Con- stantinos Evangelinos, et al. Active memory cube: A processing-in- memory architecture for exascale systems. IBM Journal of Research and Development, ...

    2015

  53. [61]

    Active pages: A computation model for intelligent memory

    Mark Oskin, Frederic T Chong, and Timothy Sherwood. Active pages: A computation model for intelligent memory. In Proceedings. 25th Annual International Symposium on Computer Architecture , pages 192– 203, 1998

    1998

  54. [62]

    Wait of a decade: Did spec cpu 2017 broaden the performance horizon? In 2018 IEEE International Symposium on High Performance Computer Architecture (HPCA), pages 271–282

    Reena Panda, Shuang Song, Joseph Dean, and Lizy K John. Wait of a decade: Did spec cpu 2017 broaden the performance horizon? In 2018 IEEE International Symposium on High Performance Computer Architecture (HPCA), pages 271–282. IEEE, 2018

    2017

  55. [63]

    A case for intelligent ram

    David Patterson, Thomas Anderson, Neal Cardwell, Richard Fromm, Kimberly Keeton, Christoforos Kozyrakis, Randi Thomas, and Katherine Yelick. A case for intelligent ram. IEEE micro, 17(2):34–44, 1997

    1997

  56. [64]

    Scheduling techniques for gpu architectures with processing-in-memory capabilities

    Ashutosh Pattnaik, Xulong Tang, Adwait Jog, Onur Kayiran, Asit K Mishra, Mahmut T Kandemir, Onur Mutlu, and Chita R Das. Scheduling techniques for gpu architectures with processing-in-memory capabilities. In Proceedings of the 2016 International Conference on Parallel Archi- t...

    2016

  57. [65]

    Hybrid memory cube (hmc)

    J Thomas Pawlowski. Hybrid memory cube (hmc). In 2011 IEEE Hot Chips 23 Symposium (HCS) , pages 1–24. IEEE, 2011

    2011

  58. [66]

    Reverse engineering intel dram addressing and exploitation

    Peter Pessl, Daniel Gruss, Cl ´ementine Maurice, Michael Schwarz, and Stefan Mangard. Reverse engineering intel dram addressing and exploitation. arXiv preprint arXiv:1511.08756 , 2015

    2015 arXiv

  59. [67]

    Drama: Exploiting dram addressing for cross-cpu attacks

    Peter Pessl, Daniel Gruss, Cl ´ementine Maurice, Michael Schwarz, and Stefan Mangard. Drama: Exploiting dram addressing for cross-cpu attacks. In USENIX Security Symposium , pages 565–581, 2016

    2016

  60. [68]

    Nu-minebench 2.0

    Jayaprakash Pisharath, Ying Liu, Wei-keng Liao, Alok Choudhary, Gokhan Memik, and Janaki Parhi. Nu-minebench 2.0. Technical report, Technical report, Northwestern University, 2005

    2005

  61. [69]

    Hog- wild: A lock-free approach to parallelizing stochastic gradient descent

    Benjamin Recht, Christopher Re, Stephen Wright, and Feng Niu. Hog- wild: A lock-free approach to parallelizing stochastic gradient descent. In Advances in neural information processing systems , pages 693–701, 2011

    2011

  62. [70]

    Dally, Ujval J

    Scott Rixner, William J. Dally, Ujval J. Kapasi, Peter Mattson, and John D. Owens. Memory access scheduling. In Proceedings of the 27th Annual International Symposium on Computer Architecture , ISCA ’00, pages 128–138, New York, NY , USA, 2000. ACM

    2000

  63. [71]

    Fast bulk bitwise and and or in dram

    Vivek Seshadri, Kevin Hsieh, Amirali Boroum, Donghyuk Lee, Michael A Kozuch, Onur Mutlu, Phillip B Gibbons, and Todd C Mowry. Fast bulk bitwise and and or in dram. IEEE Computer Architecture Letters, 14(2):127–131, 2015

    2015

  64. [72]

    Ambit: In-memory accelerator for bulk bitwise operations using commodity dram technology

    Vivek Seshadri, Donghyuk Lee, Thomas Mullins, Hasan Hassan, Amirali Boroumand, Jeremie Kim, Michael A Kozuch, Onur Mutlu, Phillip B Gibbons, and Todd C Mowry. Ambit: In-memory accelerator for bulk bitwise operations using commodity dram technology. In Proceedings of the 50th A...

    2017

  65. [73]

    Isaac: A convolutional neural network accelerator with in-situ analog arithmetic in crossbars

    Ali Shafiee, Anirban Nag, Naveen Muralimanohar, Rajeev Balasubra- monian, John Paul Strachan, Miao Hu, R Stanley Williams, and Vivek Srikumar. Isaac: A convolutional neural network accelerator with in-situ analog arithmetic in crossbars. In Proceedings of the 43rd International...

    2016

  66. [74]

    Pipelayer: A pipelined reram-based accelerator for deep learning

    Linghao Song, Xuehai Qian, Hai Li, and Yiran Chen. Pipelayer: A pipelined reram-based accelerator for deep learning. In High Per- formance Computer Architecture (HPCA), 2017 IEEE International Symposium on, pages 541–552. IEEE, 2017

    2017

  67. [75]

    Graphr: Accelerating graph processing using reram

    Linghao Song, Youwei Zhuo, Xuehai Qian, Hai Li, and Yiran Chen. Graphr: Accelerating graph processing using reram. In High Per- formance Computer Architecture (HPCA), 2018 IEEE International Symposium on, pages 531–543. IEEE, 2018

    2018

  68. [76]

    High bandwidth memory (hbm) dram

    JEDEC Standard. High bandwidth memory (hbm) dram. JESD235, 2013

    2013

  69. [77]

    A logic-in-memory computer

    Harold S Stone. A logic-in-memory computer. IEEE Transactions on Computers, C-19(1):73–78, Jan 1970

    1970

  70. [78]

    Hunter, and Lizy K

    Jeffrey Stuecheli, Dimitris Kaseridis, David Daly, Hillery C. Hunter, and Lizy K. John. The virtual write queue: Coordinating dram and last- level cache policies. In Proceedings of the 37th Annual International Symposium on Computer Architecture , ISCA ’10, pages 72–82. ACM, 2010

    2010

  71. [79]

    Energy-efficient sql query exploiting rram-based process-in-memory structure

    Yuliang Sun, Yu Wang, and Huazhong Yang. Energy-efficient sql query exploiting rram-based process-in-memory structure. In Non-Volatile Memory Systems and Applications Symposium (NVMSA), 2017 IEEE 6th, pages 1–6. IEEE, 2017

    2017

  72. [80]

    Data access optimization in a processing- in-memory system

    Zehra Sura, Arpith Jacob, Tong Chen, Bryan Rosenburg, Olivier Sal- lenave, Carlo Bertolli, Samuel Antao, Jose Brunheroto, Yoonho Park, Kevin O’Brien, et al. Data access optimization in a processing- in-memory system. In Proceedings of the 12th ACM International Conference on C...

    2015

  73. [81]

    Palloc: Dram bank-aware memory allocator for performance isolation on multicore platforms

    Heechul Yun, Renato Mancuso, Zheng-Pei Wu, and Rodolfo Pellizzoni. Palloc: Dram bank-aware memory allocator for performance isolation on multicore platforms. In Real-Time and Embedded Technology and Applications Symposium (RTAS), 2014 IEEE 20th , pages 155–166. IEEE, 2014

    2014

  74. [82]

    Top-pim: throughput- oriented programmable processing in memory

    Dongping Zhang, Nuwan Jayasena, Alexander Lyashevsky, Joseph L Greathouse, Lifan Xu, and Michael Ignatowski. Top-pim: throughput- oriented programmable processing in memory. In Proceedings of the 23rd international symposium on High-performance parallel and distributed computi...

    2014

  75. [83]

    Graphp: Reducing communication for pim-based graph processing with efficient data partition

    Mingxing Zhang, Youwei Zhuo, Chao Wang, Mingyu Gao, Yongwei Wu, Kang Chen, Christos Kozyrakis, and Xuehai Qian. Graphp: Reducing communication for pim-based graph processing with efficient data partition. In High Performance Computer Architecture (HPCA), 2018 IEEE International...

    2018

  76. [84]

    A permutation-based page interleaving scheme to reduce row-buffer conflicts and exploit data locality

    Zhao Zhang, Zhichun Zhu, and Xiaodong Zhang. A permutation-based page interleaving scheme to reduce row-buffer conflicts and exploit data locality. In Proceedings 33rd Annual IEEE/ACM International Symposium on Microarchitecture. MICRO-33 2000, pages 32–41. IEEE, 2000

    2000

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.