Pith. sign in

REVIEW 3 major objections 6 minor 48 references

The Fallacy of Independent Ceilings: Characterizing Coupled Load-Branch Stall Interaction

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Isolated perfect-branch and perfect-cache speedups do not compose multiplicatively: the joint speedup exceeds the product on 70% of simulated irregular kernels, so independent-ceiling analysis can understate headroom.

desk verdict JSS's product baseline is the wrong null, so the headline coupling numbers don't prove symbiosis; the paper's conditional-gain recommendation is still worth adopting. read the letter →

arxiv 2608.11380 v1 pith:CPG5ZHWW submitted 2026-08-11 cs.AR cs.PF

Matthew Constant , Resit Sendag This is my paper · ORCID
classification cs.ARcs.PF
keywords branchmispredictioncachemissesirregularworkloadsjointspeedupsynergysymbioticstalllatencyperformanceevaluationout-of-orderprocessors
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

Branch mispredictions and data-cache misses are usually evaluated as separate bottlenecks, and the product of their isolated speedups is treated as the joint ceiling. This paper argues that in irregular hot loops—where a hard-to-predict branch depends on a long-latency cache-missing load—that product is the wrong baseline. It introduces joint speedup synergy (JSS), the measured joint speedup divided by the product of the isolated speedups, and reports that across 53 simulated kernels 70% have JSS above 1, 40% exceed the product by more than 6%, and the highest-pressure kernels reach JSS values from 1.23 to 3.29. The practical point is that evaluations reporting only isolated perfect-branch and perfect-cache speedups can understate the headroom that jointly addressing both bottlenecks would unlock.

What carries the argument

The machinery is a paired evaluation protocol built on four simulation modes: baseline, perfect branch direction, perfect cache (all demand accesses hit at L1 latency), and both perfect together. The load-bearing metric is joint speedup synergy (JSS), the ratio of the measured joint speedup to the product of the two isolated speedups, which operationalizes the independent-ceilings assumption. The supporting mechanism is symbiotic stall latency (SSL): in a hot loop containing an irregular load and a data-dependent branch, the two penalties interact through reorder-buffer occupancy, squash rate, and commit starvation, so each isolated fix leaves the other exposed. A cheap screen, symbiotic stall opportunity ($\mathrm{SSO} = \mathrm{MPKI}_{\mathrm{BP}}\mathrm{MPKI}_{\mathrm{cache}} / (\mathrm{MPKI}_{\mathrm{BP}} + \mathrm{MPKI}_{\mathrm{cache}})$), uses committed branch-misprediction and demand-miss rates per kilo-instruction to select workloads that merit the expensive perfect-mode validation. High-JSS kernels are then grouped into four software patterns—neighbor access, hash lookup, linked-structure traversal, and data-dependent modification—which locate where the coupling usually arises.

What would settle it

On a synthetic loop whose branch outcome and load address are engineered to be independent (the branch resolves from a register, and the load misses in a separate dependency chain), run the four configurations and compute JSS. If JSS stays above 1, the metric is measuring baseline arithmetic rather than interaction; if JSS collapses to 1 when the load is software-prefetched or the branch is replaced by an always-taken outcome, the coupling interpretation is confirmed.

Watch

Extended reading notes

Core claim

The central claim is that removing one stall source in a coupled loop shifts, rather than removes, the bottleneck. With perfect branch direction, squashes drop but surviving correct-path work fills the reorder buffer behind long-latency loads; with a perfect cache, the pipeline reaches the mispredicting branch more often per unit time and squash pressure rises. Only when both are perfected can the loop body drain quickly and commit fully, so the measured joint speedup exceeds the product of the isolated speedups. The paper reports $JSS = \mathrm{Speedup}_{\mathrm{perfect}} / (\mathrm{Gain}_{\mathrm{pBP}} \cdot \mathrm{Gain}_{\mathrm{pCache}})$ above 1 on 37 of 53 kernels (70%) and above 1.06 on 21 of 53 (40%), interpreting these surpluses as evidence that independent-ceiling analysis understates attainable gain in high-pressure irregular workloads. It explicitly scopes the claim: near-unity JSS workloads compose well enough for conventional analysis, and high JSS is an upper-bound diagnostic, not a prediction for a deployable mechanism.

Load-bearing premise

The argument rests on taking the product of the two isolated speedups as the baseline for independent effects; if the correct independence baseline is higher—because independent penalties add in cycles rather than multiplying speedups—then part of the reported synergy is not coupling at all.

Editorial extensions

If this is right

  • Near-unity JSS or low SSO supports conventional isolated branch and cache evaluation; high SSO with high JSS indicates that joint or conditional measurements are required.
  • Branch predictor studies should report gains under both the baseline memory system and a strong-cache configuration, because cache misses can mask a predictor's real value.
  • Prefetcher and cache studies should report gains under both baseline branch prediction and strong branch prediction, because misprediction squashes can mask memory-side value.
  • The four software patterns give a concrete screen for benchmark selection, and full-application results should be sliced by phase or checkpoint rather than aggregate-only, since aggregate metrics hide short coupled intervals.

Reading between the lines

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

  • The multiplicative independence baseline is the weakest step: under an additive cycle-accounting model, truly independent penalties would already predict a joint speedup above the product, so some of the reported JSS surplus may be arithmetic rather than interaction.
  • The conditional-gain ordering suggests a practical design heuristic: reducing the cheaper bottleneck first raises the marginal value of fixing the other, so the paper's conditional columns double as a priority guide.
  • SSO could plausibly serve as a dynamic trigger for enabling coupled mechanisms on the fly, a use the paper leaves for future work.
  • The four-pattern taxonomy hints at compiler-directed prefetching or load pinning for the specific load feeding the mispredicting branch; nothing in the paper tests that intervention.
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

3 major / 6 minor

Summary. This manuscript argues that treating perfect-branch and perfect-cache speedups as independent ceilings is fallacious for irregular workloads. It defines Joint Speedup Synergy (JSS) as the measured joint perfect-mode speedup divided by the product of the two isolated speedups, and reports that on 53 simulated kernels 70% have JSS > 1 and 40% exceed the independence product by more than 6%. It introduces Symbiotic Stall Opportunity (SSO) as an MPKI-based screening metric, classifies high-JSS loops into four software patterns, and presents gem5 measurements of ROB occupancy, squash rate, and commit starvation to support the claim that removing one penalty exposes the other. The paper concludes that evaluations should report conditional branch-after-cache and cache-after-branch gains alongside isolated and joint speedups.

Significance. If the JSS definition and its null model were correct, the paper would provide a practical caution against multiplicative composition of isolated ceilings and a useful taxonomy of load–branch coupling patterns. The study is reproducible in design (deterministic gem5 SE mode, clear perfect-mode definitions, open benchmark suites), and the pipeline-level observations in Figures 6 and 7 are informative. However, the central quantitative result—that JSS > 1 on 70% of kernels demonstrates coupling—is not supported, because the product-of-isolated-speedups baseline is mathematically the wrong independence null under the additive CPI model the paper itself cites. The load-bearing claim therefore fails, and the reported percentages become an artifact of metric definition rather than a demonstrated architectural phenomenon.

major comments (3)
  1. [Section IV.C.5, Eqs. (2)–(6)] The independence baseline is not the product of isolated speedups. Under the additive CPI model cited in references [20–23], if CPI_base = CPI_exec + CPI_branch + CPI_cache with the two penalties independent, then S_b = CPI_base/(CPI_exec+CPI_cache), S_c = CPI_base/(CPI_exec+CPI_branch), S_joint = CPI_base/CPI_exec, and JSS = S_joint/(S_b·S_c) = CPI_base/CPI_exec > 1. For example, equal branch and cache penalties of 25% each yield JSS = 1.125 with zero coupling. The paper never states or justifies an assumption that branch and cache penalties occupy disjoint multiplicative time fractions; indeed its cited interval models are additive. Consequently, JSS > 1 on 37 of 53 kernels (70%) is exactly what an uncoupled additive model predicts when both penalty classes are present, and it does not demonstrate that isolated ceilings understate joint headroom.
  2. [Section V.A, Table III columns 4–5] The conditional-gain argument is also expected under the additive null. If perfect-bp is already applied, the additional speedup from perfect-cache is CPI_base/CPI_exec, which is always greater than the isolated cache gain CPI_base/(CPI_exec+CPI_branch); similarly for the other order. The fact that a conditional gain exceeds its isolated counterpart (e.g., tc: 3.27 vs. 1.85, 3.12 vs. 1.76) is therefore not evidence of mutual amplification; it is a mathematical consequence of the additive independent-penalty model. The text presents these numbers as evidence of coupling without comparing them to the additive null.
  3. [Section IV.D, Figure 4] The thresholds used to declare 'measurable coupling' (JSS > 1) and 'actionable headroom' (JSS > 1.06) are applied without a null-model calibration. The proper test of coupling is whether JSS exceeds the additive expectation CPI_base/CPI_exec, which depends on the size of the execution component; a fixed 6% threshold cannot distinguish coupling from baseline arithmetic. The paper's own counterexamples bisort and health, which exhibit JSS ≤ 1 while clearing the SSO screen, show that the metric can move in both directions, and the paper provides no statistical model for how much JSS would be expected under independence for each kernel. The headline percentages are therefore not interpretable as a measure of how often independent-ceiling analysis fails.
minor comments (6)
  1. [Abstract] The phrase 'interactionsymbiotic stall latency' should read 'interaction symbiotic stall latency'.
  2. [Section III.B, Figure 3] The text says 'IPC rises only to 2.35×baseline' but this is a speedup value, not an IPC value; please use consistent terminology throughout (IPC vs. speedup).
  3. [Section IV.C.4, Eq. (1)] SSO is the harmonic mean of the two MPKI values divided by two; stating this explicitly would help readers interpret the threshold SSO > 5.
  4. [Figure 4] The kernel labels in the figure are too small to read in print; please enlarge them or use a two-panel layout with a separate legend.
  5. [Section V.A.4, Table VIII] The table caption says 'Comparison Sort' but the code excerpt is the partitioning loop of IntroSort; rename to 'IntroSort partition' for accuracy.
  6. [Section IV.D] The claim 'kernels with SSO > 20 show JSS from 1.23 to 3.29' should specify whether this is among screened kernels only; bisort and health are said to exceed the SSO threshold yet have JSS ≤ 1, so the statement needs qualification.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: JSS and SSO are transparently defined measured quantities, and the multiplicative independence baseline is an explicit assumption rather than a fitted input.

full rationale

The paper's central quantities are measured, not derived. JSS is defined (Eq. 6) as the ratio of the observed joint perfect-mode speedup to the product of the two isolated perfect-mode speedups, and SSO is defined (Eq. 1) directly from MPKI values. The statement that JSS > 1 indicates the product baseline understates joint headroom is an explicit interpretive label attached to the metric, not a hidden derivation, and the paper supports it with independently measured conditional gains, ROB occupancy, squash rates, and commit activity. The multiplicative 'Expected speedup' (Eq. 2) is stated as an Amdahl-style independence assumption with citation [25]; whether that baseline is the correct null under an additive CPI model is a modeling-validity question, not a circularity, and the paper's conditional-gain and pipeline measurements would remain meaningful even if the baseline were replaced. The SSO > 5 and JSS > 1.06 thresholds are analytic choices made on the same dataset rather than fitted parameters, and the paper explicitly presents bisort and health as counterexamples, so no fitted input is renamed as a prediction. No load-bearing self-citation appears: reference [4], which shares an author, is background motivation for Node Tracker and does not carry the JSS/SSO framework. No step in the claimed derivation chain reduces to its own inputs, so the appropriate finding is no significant circularity.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The central claim rests on a chosen independence baseline (the product of speedups) that is not the standard additive composition used in CPI-stack analysis, on arbitrary thresholds, and on an idealization of perfect cache. No physical entity is invented; SSL is a label for measured pipeline behavior.

free parameters (4)
  • SSO screening threshold = 5
    Chosen as the cutoff for detailed analysis in Section IV.D; no principled derivation is given and the value appears selected after inspecting the SSO distribution.
  • JSS actionable threshold = 1.06
    Defines the 'exceed the independence product by more than 6%' bucket in Section IV.D and the abstract; presented as conservative but arbitrary.
  • SSO greater than 20 bucket = 20
    Used in the abstract and Section IV.D to report the headline JSS range of 1.23 to 3.29; no independent justification for this cutoff.
  • Baseline prefetcher selection = per-benchmark best of Stride, SPP, BOP, IMP
    The baseline uses the best prefetcher per benchmark (Section IV.B), a per-workload choice that changes isolated speedups and therefore all JSS values.
assumptions (4)
  • domain assumption The product of isolated perfect-branch and perfect-cache speedups is the correct independence baseline for jointly perfecting both (Equation 2).
    Invoked in Section II.C and Section IV.C. Under additive cycle accounting, the model cited from references [20-23], independent penalties yield a joint speedup larger than the product, so JSS greater than 1 can occur with zero interaction.
  • domain assumption Perfect-cache is approximated by making all demand accesses hit the L1D timing path with zero refill latency.
    Stated in Section IV.C.2. This removes long-latency misses entirely and is an upper-bound idealization, not a realistic hierarchy; it directly affects absolute JSS values.
  • domain assumption SE-mode deterministic simulation of single-threaded kernels with 100M-instruction regions is representative of the studied workloads.
    Section IV.B. This excludes OS, multicore, and shared-cache effects that could alter branch-cache coupling.
  • ad hoc to paper The functional form of SSO as a harmonic product of MPKI values is a valid screening proxy.
    Equation (1). No derivation is given; it is a heuristic chosen so that SSO is high only when both branch and cache miss rates are high.
invented entities (1)
  • Symbiotic Stall Latency (SSL) independent evidence
    purpose: Names the proposed mechanism by which cache-miss and branch-misprediction penalties interact through ROB occupancy, squash rate, and commit bandwidth.
    The paper provides direct pipeline measurements (ROB occupancy, squash events, iteration commit ratio) that would support or falsify the mechanism, though the JSS metric used to quantify it is confounded by the multiplicative baseline.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Fallacy of Independent Ceilings: Characterizing Coupled Load-Branch Stall Interaction." pith.science (2026). https://pith.science/paper/CPG5ZHWW

@misc{pith2026260811380,
  author       = {Pith},
  title        = {Pith review of: The Fallacy of Independent Ceilings: Characterizing Coupled Load-Branch Stall Interaction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CPG5ZHWW}},
  note         = {Machine review of arXiv:2608.11380}
}
read the original abstract

Branch mispredictions and data-cache misses are usually evaluated as separate bottlenecks: studies report perfect-branch or perfect-cache speedups as isolated upper bounds and often treat their product as the joint ceiling. In irregular workloads, however, hard-to-predict branches and cache-missing loads often occur in the same hot loops. Removing one penalty can expose the other: faster memory reaches mispredicted branches sooner, while better branch prediction leaves more long-latency loads in the out-of-order window. We call this interaction symbiotic stall latency (SSL). This paper quantifies when isolated ceilings fail using joint speedup synergy (JSS), the observed joint perfect-branch/perfect-cache speedup divided by the product of the isolated speedups. Values above one mean independent-ceiling analysis understates attainable gain. Across 53 simulated workloads, 70% show measurable coupling (JSS > 1), though many are near unity, especially in lower-pressure cases. With a conservative threshold, 40% exceed the independence product by more than 6%, and kernels with SSO > 20 show JSS from 1.23 to 3.29. We introduce symbiotic stall opportunity (SSO), a lightweight MPKI-based screen for workloads that merit full joint simulation. We map high-SSO workloads to four recurring software patterns: neighbor access, hash lookup, linked-structure traversal, and data-dependent modification. We connect SSL to reorder-buffer occupancy, squash rate, and commit starvation under isolated perfect modes. The resulting methodology is simple: use SSO to screen, JSS to validate, and report conditional branch-after-cache and cache-after-branch gains when evaluating branch predictors, prefetchers, caches, or coupled branch/memory mechanisms. Our contribution is a measurement framework showing when isolated perfect modes are adequate and when they understate joint performance headroom.

Figures

Figures reproduced from arXiv: 2608.11380 by the authors.

Figure 1
Figure 1. Coupled load–branch stalls in a hot loop. Load L1 (irregular access, cache MPKI) pairs with branch B1 (data-dependent, branch MPKI). Top path: better direction prediction lowers branch MPKI but raises ROB occupancy because fewer instructions are squashed while loads still miss. Bottom path: better caching/prefetching lowers cache MPKI and speeds retirement, but can increase squash events per cycle as the core reache… view at source ↗
Figure 2
Figure 2. Pipeline snapshots of SSL on the mst hash loop. Three iterations of the branch–load pair are shown; time scales differ across panels. The isolated perfect modes still leave either surviving load latency or wrong-path churn, whereas joint perfect yields the short, commit-ready loop body that JSS quantifies in Section IV. perfect-bp and perfect-cache speedups. The quan￾titative metrics in [PITH_FULL_IMAGE:figures/ful… view at source ↗
Figure 3
Figure 3. mst hash-loop behavior under four simulation configurations. Bars compare baseline, perfect-bp, perfect-cache, and perfect. (a) ROB occupancy rises with perfect-bp because fewer squashes leave long-latency loads in flight. (b) Squashes per 1K cycles jump under perfect-cache when memory is fast but B1 remains mispredicted. (c) IPC shows isolated modes far below perfect. (d) Iteration commit ratio: fraction of observe… view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Population-wide SSO screening and JSS measurement for all simulated kernels. Inexpensive MPKI-based screening (Equation (1)) is followed by perfect-mode JSS validation (Equation (6)). Observation: high SSO is a useful screen for JSS > 1, but it must be validated with p…
Figure 5
Figure 5. Figure 5: Joint speedup synergy (JSS) by benchmark and algorithmic family. Each bar is JSS = Speedupobserved/Expectedspeedup (Equation (6)); color encodes family (Table IV). Family membership alone does not determine JSS: NNA/HTL concentrate high surplus, while several LDST kern…
Figure 6
Figure 6. Figure 6: reports three pipeline signatures on screened kernels. Under perfect-bp, fewer squashes leave more correct-path instructions in flight; if they depend on irregular loads, the ROB fills with useful but blocked work (geomean occupied entries ∼122→157). Under perfect-cach…
Figure 7
Figure 7. Figure 7: Hot-loop iteration throughput (five representative kernels). perfect-bp recovers speculative iterations but can lengthen drain time on memory-bound loops; perfect-cache shortens snapshots but not when squashes dominate. Joint perfect approaches the short, fully committ…
Figure 8
Figure 8. Figure 8: ROB scaling under isolated perfect modes (256/512/1024 entries). perfect-bp benefits when fewer squashes expose load latency; perfect-cache is insensitive because wrong-path work dominates. Under joint perfect, IPC is likewise flat (not shown). depth, dominate; joint p…
Figure 9
Figure 9. Figure 9: Speedup across simulation modes (screened kernels, normalized to base). Bars show perfect-bp, perfect-cache, and joint perfect speedup. On coupled-stall kernels (tc, mst, RandAcc), the joint bar far exceeds either isolated mode; near-independent kernels (health, tsp) a…
Figure 10
Figure 10. Figure 10: SPEC CPU2017 int: (a) SSO screening, (b) IPC under different modes, and (c) mcf checkpoint speedups. Only 605.mcf_s is high-SSO/high-JSS in aggregate. Per-checkpoint speedups show why aggregation can hide SSL: CPT9 is cache-dominated (JSS ≈ 1.05), whereas CPT8 and CPT…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 42 canonical work pages

  1. [1]

    J. L. Hennessy and D. A. Patterson,Computer archi- tecture: a quantitative approach. Elsevier, 2011

    work page 2011

  2. [2]

    A survey of techniques for dynamic branch prediction,

    S. Mittal, “A survey of techniques for dynamic branch prediction,”Concurrency and Computation: Practice and Experience, vol. 31, no. 1, e4666, 2019

    work page 2019

  3. [3]

    Evaluation of hardware data prefetchers on server processors,

    M. Bakhshalipour, S. Tabaeiaghdaei, P. Lotfi-Kamran, and H. Sarbazi-Azad, “Evaluation of hardware data prefetchers on server processors,”ACM Computing Sur- veys (CSUR), vol. 52, no. 3, pp. 1–29, 2019

    work page 2019

  4. [4]

    Fast key-value lookups with node tracker,

    M. Cavus, M. Shatnawi, R. Sendag, and A. K. Uht, “Fast key-value lookups with node tracker,”ACM Transactions on Architecture and Code Optimization, vol. 18, no. 3, 34:1–34:26, 2021.DOI: 10.1145/3452099

    work page doi:10.1145/3452099 2021

  5. [5]

    The 3rd data prefetching championship,

    “The 3rd data prefetching championship,” 2019. [On- line]. Available: https://dpc3.compas.cs.stonybrook.edu/ [6]4th data prefetching championship (dpc4), Co-located with HPCA 2026, Sydney, Australia, 2026. [Online]. Available: https://sites.google.com/view/dpc4- 2026/ home

    work page 2019

  6. [7]

    [Online]

    cbp,Championship branch prediction (cbp-5), 2016. [Online]. Available: https://www.jilp.org/cbp2016/ [8]6th championship branch prediction (cbp2025), Co- located with ISCA 2025, Tokyo, Japan, 2025. [Online]. Available: https : / / ericrotenberg . wordpress . ncsu . edu / cbp2025/

    work page 2016

  7. [9]

    Tage-sc-l branch predictors again,

    A. Seznec, “Tage-sc-l branch predictors again,” 2016

    work page 2016

  8. [10]

    A new case for the tage branch predic- tor,

    A. Seznec, “A new case for the tage branch predic- tor,” inProceedings of the 44th Annual IEEE/ACM International Symposium on Microarchitecture, 2011, pp. 117–127

    work page 2011

Show all 48 references
  1. [11]

    Multiperspective perceptron predictor,

    D. A. Jim ´enez, “Multiperspective perceptron predictor,” 2016

    2016

  2. [12]

    Best-offset hardware prefetching,

    P. Michaud, “Best-offset hardware prefetching,” in2016 IEEE International Symposium on High Performance Computer Architecture (HPCA), IEEE, 2016, pp. 469– 480

    2016

  3. [13]

    Path confidence based lookahead prefetching,

    J. Kim, S. H. Pugsley, P. V . Gratz, A. N. Reddy, C. Wilkerson, and Z. Chishti, “Path confidence based lookahead prefetching,” in2016 49th Annual IEEE/ACM International Symposium on Microarchitec- ture (MICRO), IEEE, 2016, pp. 1–12

    2016

  4. [14]

    Linearizing irregular mem- ory accesses for improved correlated prefetching,

    A. Jain and C. Lin, “Linearizing irregular mem- ory accesses for improved correlated prefetching,” in Proceedings of the 46th Annual IEEE/ACM Interna- tional Symposium on Microarchitecture, ser. MICRO- 46, Davis, California: Association for Computing Ma- chinery, 2013, pp. 247...

    2013

  5. [15]

    Efficiently prefetching complex address patterns,

    M. Shevgoor, S. Koladiya, R. Balasubramonian, C. Wilkerson, S. H. Pugsley, and Z. Chishti, “Efficiently prefetching complex address patterns,” in2015 48th Annual IEEE/ACM International Symposium on Mi- croarchitecture (MICRO), 2015, pp. 141–152.DOI: 10. 1145/2830772.2830793

    2015

  6. [16]

    Imp: Indirect memory prefetcher,

    X. Yu, C. J. Hughes, N. Satish, and S. Devadas, “Imp: Indirect memory prefetcher,” inProceedings of the 48th International Symposium on Microarchitecture, 2015, pp. 178–190

    2015

  7. [17]

    Address- branch correlation: A novel locality for long-latency hard-to-predict branches,

    H. Gao, Y . Ma, M. Dimitrov, and H. Zhou, “Address- branch correlation: A novel locality for long-latency hard-to-predict branches,” inProceedings of the 14th IEEE International Symposium on High Performance Computer Architecture (HPCA), 2008, pp. 134–145

    2008

  8. [18]

    EX- ACT: Explicit dynamic-branch prediction with active updates,

    M. Al-Otoom, E. Forbes, and E. Rotenberg, “EX- ACT: Explicit dynamic-branch prediction with active updates,” inProceedings of the 2010 ACM International Conference on Computing Frontiers (CF), 2010, pp. 1– 10

    2010

  9. [19]

    Control-flow decoupling,

    R. Sheikh, J. Tuck, and E. Rotenberg, “Control-flow decoupling,” in2012 45th Annual IEEE/ACM Interna- tional Symposium on Microarchitecture, IEEE, 2012, pp. 329–340

    2012

  10. [20]

    A performance counter architecture for computing accurate CPI components,

    S. Eyerman, L. Eeckhout, T. Karkhanis, and J. E. Smith, “A performance counter architecture for computing accurate CPI components,” inProceedings of the 12th International Conference on Architectural Support for Programming Languages and Operating Systems (AS- PLOS), 2006, pp. 175–184

    2006

  11. [21]

    Charac- terizing the branch misprediction penalty,

    S. Eyerman, J. E. Smith, and L. Eeckhout, “Charac- terizing the branch misprediction penalty,” inIEEE International Symposium on Performance Analysis of Systems and Software (ISPASS), 2006, pp. 48–58.DOI: 10.1109/ISPASS.2006.1620789

    2006 arXiv

  12. [22]

    A first-order su- perscalar processor model,

    T. S. Karkhanis and J. E. Smith, “A first-order su- perscalar processor model,” inProceedings of the 31st Annual International Symposium on Computer Archi- tecture (ISCA), 2004, pp. 338–349.DOI: 10.1109/ISCA. 2004.1310786

    2004 arXiv

  13. [23]

    A top-down method for performance anal- ysis and counters architecture,

    A. Yasin, “A top-down method for performance anal- ysis and counters architecture,” inIEEE International Symposium on Performance Analysis of Systems and Software (ISPASS), 2014, pp. 35–44.DOI: 10 . 1109 / ISPASS.2014.6844459

    2014

  14. [24]

    Branch prediction is not a solved problem: Measurements, opportunities, and future directions,

    C.-K. Lin and S. Tarsa, “Branch prediction is not a solved problem: Measurements, opportunities, and future directions,” inIEEE International Symposium on Workload Characterization (IISWC), 2019, pp. 228– 238.DOI: 10.1109/IISWC47752.2019.9042108

    2019

  15. [25]

    Amdahl’s law in the multicore era,

    M. D. Hill and M. R. Marty, “Amdahl’s law in the multicore era,”Computer, vol. 41, no. 7, pp. 33–38, 2008

    2008

  16. [26]

    The gap benchmark suite,

    S. Beamer, K. Asanovi, and D. Patterson, “The gap benchmark suite,” arXiv preprint, 2015. arXiv: 1508 . 03619

    2015

  17. [27]

    The gem5 simulator,

    N. Binkert et al., “The gem5 simulator,”ACM SIGARCH computer architecture news, vol. 39, no. 2, pp. 1–7, 2011

    2011

  18. [28]

    Classifying memory access patterns for prefetching,

    G. Ayers, H. Litz, C. Kozyrakis, and P. Ranganathan, “Classifying memory access patterns for prefetching,” inProceedings of the Twenty-Fifth International Con- ference on Architectural Support for Programming Lan- guages and Operating Systems, 2020, pp. 513–526

    2020

  19. [29]

    Dynamic branch prediction with perceptrons,

    D. A. Jimnez and C. Lin, “Dynamic branch prediction with perceptrons,” inProceedings HPCA Seventh In- ternational Symposium on High-Performance Computer Architecture, IEEE, 2001, pp. 197–206

    2001

  20. [30]

    Evaluation of hardware- based stride and sequential prefetching in shared- memory multiprocessors,

    F. Dahlgren and P. Stenstrom, “Evaluation of hardware- based stride and sequential prefetching in shared- memory multiprocessors,”IEEE Transactions on Paral- lel and Distributed Systems, vol. 7, no. 4, pp. 385–398, 1996

    1996

  21. [31]

    A prefetching technique for irregular accesses to linked data structures,

    M. Karlsson, F. Dahlgren, and P. Stenstrom, “A prefetching technique for irregular accesses to linked data structures,” inProceedings Sixth International Symposium on High-Performance Computer Archi- tecture. HPCA-6 (Cat. No. PR00550), IEEE, 2000, pp. 206–217

    2000

  22. [32]

    Dependence based prefetching for linked data structures,

    A. Roth, A. Moshovos, and G. S. Sohi, “Dependence based prefetching for linked data structures,” inPro- ceedings of the eighth international conference on Ar- chitectural support for programming languages and operating systems, 1998, pp. 115–126

    1998

  23. [33]

    Se- mantic locality and context-based prefetching using reinforcement learning,

    L. Peled, S. Mannor, U. Weiser, and Y . Etsion, “Se- mantic locality and context-based prefetching using reinforcement learning,” in2015 ACM/IEEE 42nd An- nual International Symposium on Computer Architec- ture (ISCA), IEEE, 2015, pp. 285–297

    2015

  24. [34]

    Understanding memory access patterns for prefetching,

    P. Braun and H. Litz, “Understanding memory access patterns for prefetching,” inInternational Workshop on AI-assisted Design for Architecture (AIDArc), held in conjunction with ISCA, 2019

    2019

  25. [35]

    A study of slipstream processors,

    Z. Purser, K. Sundaramoorthy, and E. Rotenberg, “A study of slipstream processors,” inProceedings of the 33rd annual ACM/IEEE international symposium on Microarchitecture, 2000, pp. 269–280

    2000

  26. [36]

    Simulation of computer archi- tectures: Simulators, benchmarks, methodologies, and recommendations,

    J. J. Yi and D. J. Lilja, “Simulation of computer archi- tectures: Simulators, benchmarks, methodologies, and recommendations,”IEEE Transactions on computers, vol. 55, no. 3, pp. 268–280, 2006

    2006

  27. [37]

    A case for MLP-aware cache replacement,

    M. K. Qureshi, D. N. Lynch, O. Mutlu, and Y . N. Patt, “A case for MLP-aware cache replacement,” inPro- ceedings of the 33rd Annual International Symposium on Computer Architecture (ISCA), 2006, pp. 167–177

    2006

  28. [38]

    Using simpoint for accurate and efficient simulation,

    E. Perelman, G. Hamerly, M. Van Biesbrouck, T. Sher- wood, and B. Calder, “Using simpoint for accurate and efficient simulation,”ACM SIGMETRICS Performance Evaluation Review, vol. 31, no. 1, pp. 318–319, 2003

    2003

  29. [39]

    Smarts: Accelerating microarchitecture simula- tion via rigorous statistical sampling,

    R. E. Wunderlich, T. F. Wenisch, B. Falsafi, and J. C. Hoe, “Smarts: Accelerating microarchitecture simula- tion via rigorous statistical sampling,”SIGARCH Com- put. Archit. News, vol. 31, no. 2, pp. 84–97, May 2003,ISSN: 0163-5964.DOI: 10.1145/871656.859629 [Online]. Availabl...

    2003

  30. [41]

    Nda: Preventing speculative execution attacks at their source,

    O. Weisse, I. Neal, K. Loughlin, T. F. Wenisch, and B. Kasikci, “Nda: Preventing speculative execution attacks at their source,” inProceedings of the 52nd Annual IEEE/ACM International Symposium on Microarchitec- ture, 2019, pp. 572–586

    2019

  31. [42]

    Hardware accelera- tion of database operations,

    J. Casper and K. Olukotun, “Hardware accelera- tion of database operations,” inProceedings of the 2014 ACM/SIGDA international symposium on Field- programmable gate arrays, 2014, pp. 151–160

    2014

  32. [43]

    M. C. Carlisle,Olden: Parallelizing Programs with Dynamic Data Structures on Distributed-Memory Ma- chines. USA: Princeton University, 1996

    1996

  33. [44]

    Brief announcement: The problem based benchmark suite,

    J. Shun et al., “Brief announcement: The problem based benchmark suite,” inProceedings of the twenty-fourth annual ACM symposium on Parallelism in algorithms and architectures, 2012, pp. 68–70

    2012

  34. [45]

    Asynchronous memory access chaining,

    O. Kocberber, B. Falsafi, and B. Grot, “Asynchronous memory access chaining,”Proceedings of the VLDB Endowment, vol. 9, no. 4, pp. 252–263, 2015

    2015

  35. [46]

    Crono: A benchmark suite for multithreaded graph algorithms executing on futuristic multicores,

    M. Ahmad, F. Hijaz, Q. Shi, and O. Khan, “Crono: A benchmark suite for multithreaded graph algorithms executing on futuristic multicores,” in2015 IEEE In- ternational Symposium on Workload Characterization, IEEE, 2015, pp. 44–55

    2015

  36. [47]

    Introducing the graph 500,

    J. A. Ang, B. W. Barrett, K. B. Wheeler, and R. C. Murphy, “Introducing the graph 500,” 2010. [48]The Boost Graph Library: User Guide and Reference Manual. USA: Addison-Wesley Longman Publishing Co., Inc., 2002,ISBN: 0201729148

    2010

  37. [49]

    The pagerank citation ranking: Bringing order to the web.,

    L. Page, S. Brin, R. Motwani, and T. Winograd, “The pagerank citation ranking: Bringing order to the web.,” Stanford InfoLab, Technical Report 1999-66, Nov. 1999, Previous number = SIDL-WP-1999-0120. [Online]. Available: http: // ilpubs. stanford. edu: 8090/ 422/

    1999

  38. [50]

    Spec cpu2017: Next-generation compute benchmark,

    J. Bucek, K.-D. Lange, and J. v. Kistowski, “Spec cpu2017: Next-generation compute benchmark,” in Companion of the 2018 ACM/SPEC International Con- ference on Performance Engineering, ser. ICPE ’18, Berlin, Germany: Association for Computing Machin- ery, 2018, pp. 41–42,ISBN: ...

    2018

  39. [51]

    Runa- head execution: An alternative to very large instruction windows for out-of-order processors,

    O. Mutlu, J. Stark, C. Wilkerson, and Y . N. Patt, “Runa- head execution: An alternative to very large instruction windows for out-of-order processors,” inThe Ninth In- ternational Symposium on High-Performance Computer Architecture, 2003. HPCA-9 2003. Proceedings, IEEE, 2003,...

    2003

  40. [52]

    From single core to multi-core: Preparing for a new expo- nential,

    J. Parkhurst, J. Darringer, and B. Grundmann, “From single core to multi-core: Preparing for a new expo- nential,” inProceedings of the 2006 IEEE/ACM inter- national conference on Computer-aided design, 2006, pp. 67–72

    2006

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.