cs.ARMar 8, 2026
SaveMitigating the Memory Bottleneck with Machine Learning-Driven and Data-Aware Microarchitectural Techniques
Abstract
Modern applications process massive data volumes that overwhelm the storage and retrieval capabilities of memory systems, making memory the primary performance and energy-efficiency bottleneck of computing systems. Although many microarchitectural techniques attempt to hide or tolerate long memory access latency, rapidly growing data footprints continue to outpace technology scaling, requiring more effective solutions. This dissertation shows that modern processors observe large amounts of application and system data during execution, yet many microarchitectural mechanisms make decisions largely independent of this information. Through four case studies, we demonstrate that such data-agnostic design leads to substantial missed opportunities for improving performance and energy efficiency. To address this limitation, this dissertation advocates shifting microarchitecture design from data-agnostic to data-informed. We propose mechanisms that (1) learn policies from observed execution behavior (data-driven design) and (2) exploit semantic characteristics of application data (data-aware design). We apply lightweight machine learning techniques and previously underexplored data characteristics across four processor components: a reinforcement learning-based hardware data prefetcher that learns memory access patterns online; a perceptron predictor that identifies memory requests likely to access off-chip memory; a reinforcement learning mechanism that coordinates data prefetching and off-chip prediction; and a mechanism that exploits repeatability in memory addresses and loaded values to eliminate predictable load instructions. Our extensive evaluation shows that the proposed techniques significantly improve performance and energy efficiency compared to prior state-of-the-art approaches.
Figures & tables
Figure 1.1 : Performance comparison of three contemporary prefetchers - stride [ 302 ] , SPP [ 465 ] , and Bingo [ 120 ] - across diverse applications. No prefetcher provides the highest performance across all applications.
Figure 1.2 : Fraction of dynamic loads that fetch the same data value from the same load address throughout the entire execution of an application (on the primary y-axis), and the performance improvement potential by ideally value-predicting and eliminating the execution of such loads (on the secondary y-axis).
Figure 3.1 : Interaction between an agent and the environment in a reinforcement learning system.
Figure 3.2 : Overview of a single-layer perceptron model. Each blue circle denotes an input and the green circle denotes the output of the perceptron.
| Workload type | # traces | |
| Single-core | AI/ML applications (AIML) | 81 |
| Graph Mining Suite (GMS) | 43 | |
| Google datacenter workloads (Google) | 359 | |
| Total | 483 | |
Table 4.1: DPC4 workload traces used for evaluation.
Figure 5.1 : Comparison of (a) coverage, overprediction, and (b) performance of two recently-proposed prefetchers, SPP [ 465 ] and Bingo [ 120 ] , and our new proposal, Pythia.
Figure 5.2 : Formulating the prefetcher as an RL-agent.
| Feature | Control-flow | Data-flow | ||
|---|---|---|---|---|
| Info. | History | Info. | History | |
| Last 3-PCs | PC | last 3 | ✖ | ✖ |
| Last 4-deltas | ✖ | ✖ | Cacheline delta | last 4 |
| PC+Delta | PC | current | Cacheline delta | current |
| Last 4-PCs+Page no. | PC | last 4 | Page no. | current |
Table 5.1: Example program features.
Figure 5.3 : Overview of Pythia.
Algorithm 1 Pythia’s reinforcement learning based prefetching algorithm
Figure 5.4 : (a) The QVStore is comprised of multiple vaults. (b) Each vault is comprised of multiple planes. (c) Index generation from feature value.
Figure 5.5 : Pipelined organization of QVStore search operation. The illustration depicts three program features, each having three planes.
| Features | PC+Delta , Sequence of last-4 deltas |
| Prefetch Action List | {-6,-3,-1,0,1,3,4,5,10,11,12,16,22,23,30,32} |
| Reward Level Values | = , = , = , = , = , = , = |
| Hyperparameters | , , |
Table 5.2: Basic Pythia configuration derived from our automated design-space exploration.
| Control-flow Components | Data-flow Components |
| 1. PC of load request 2. PC-path (XOR-ed last-3 PCs) 3. PC XOR-ed branch-PC 4. None | 1. Load cacheline address 2. Page number 3. Page offset 4. Load address delta 5. Sequence of last-4 offsets 6. Sequence of last-4 deltas 7. Offset XOR-ed with delta 8. None |
Table 5.3: List of program control-flow and data-flow components used to derive the list of features for exploration.
| Structure | Description | Size |
|---|---|---|
| QVStore | • # vaults = • # planes/vault = • # entries/plane = feature dimension ( ) action dimension ( ) • Entry size = Q-value width ( b) | 24 KB |
| EQ | • # entries = 256 • Entry size = state ( b) + action index ( b) + reward ( b) + filled-bit ( b) + address ( b) | 1.5 KB |
| Total | 25.5 KB |
Table 5.4: Storage overhead of Pythia.
| Core | 1-12 cores, 4-wide OoO, 256-entry ROB, 72/56-entry LQ/SQ, Perceptron-based branch predictor [ 411 ] , 20-cycle misprediction penalty |
|---|---|
| L1/L2 Caches | Private, 32KB/256KB, 64B line, 8 way, LRU, 16/32 MSHRs, 4-cycle/14-cycle round-trip latency |
| LLC | 2MB/core, 64B line, 16 way, SHiP [ 876 ] , 64 MSHRs per LLC Bank, 34-cycle round-trip latency |
| Main Memory | 1C: Single channel, 1 rank/channel; 4C: Dual channel, 2 ranks/channel; 8C: Quad channel, 2 ranks/channel; 8 banks/rank, 2400 MTPS, 64b data-bus per channel, 2KB row buffer/bank, tRCD = tRP = 15ns, tCAS = 12.5ns |
Table 5.5: Simulated system parameters.
| Suite | # Workloads | # Traces | Example Workloads |
|---|---|---|---|
| SPEC06 | 16 | 28 | gcc, mcf, cactusADM, lbm, … |
| SPEC17 | 12 | 18 | gcc, mcf, pop2, fotonik3d, … |
| PARSEC | 5 | 11 | canneal, facesim, raytrace, … |
| Ligra | 13 | 40 | BFS, PageRank, Bellman-ford, … |
| Cloudsuite | 4 | 53 | cassandra, cloud9, nutch, … |
Table 5.6: Workloads used for evaluation.
| SPP [ 465 ] | 256-entry ST, 512-entry 4-way PT, 8-entry GHR | 6.2 KB |
| Bingo [ 120 ] | 2KB region, 64/128/4K-entry FT/AT/PHT | 46 KB |
| MLOP [ 761 ] | 128-entry AMT, 500-update, 16-degree | 8 KB |
| DSPatch [ 141 ] | Same configuration as in [ 141 ] | 3.6 KB |
| PPF [ 148 ] | Same configuration as in [ 148 ] | 39.3 KB |
Table 5.7: Configuration of evaluated prefetchers.
Figure 5.6 : Coverage and overprediction with respect to the baseline LLC misses in the single-core system.
Figure 5.7 : Performance improvement in single-core workloads. St=Stride, S=SPP, B=Bingo, D=DSPatch, and M=MLOP.
Figure 5.8 : Performance line graph of 150 single-core traces.
Figure 5.9 : Performance in the four-core system.
Figure 5.10 : Performance line graph of 272 four-core trace mixes.
Figure 5.11 : Performance on unseen traces.
Figure 5.12 : Performance of memory bandwidth-oblivious Pythia versus the basic Pythia.
Figure 5.13 : Geomean performance improvement of prefetchers in systems with varying (a) number of cores, (b) DRAM million transfers per second (MTPS), and (c) LLC size. Each DRAM bandwidth configuration roughly matches MTPS/core of various commercial processors [ 29 , 5 , 4 ] . The baseline bandwidth/LLC configuration is marked in red.
Figure 5.14 : Performance sensitivity of Pythia towards (a) the exploration rate ( ), and (b) the learning rate ( ) hyperparameter values. The values in basic Pythia configuration are marked in red.
Figure 5.15 : Performance, coverage, and overprediction of Pythia with different feature combinations. The x-axis shows experiments with different feature combinations.
Figure 5.16 : Performance sensitivity of all prefetchers to number of warmup instructions.
Figure 5.17 : Performance of multi-level prefetching schemes (a) in the baseline 2400-MTPS configuration, and (b) with varying main memory bandwidth.
Figure 5.18 : Performance of Pythia vs. the context prefetcher [ 660 ] using hardware contexts.
Figure 5.19 : Performance comparison against IBM POWER7 prefetcher [ 416 ] .
Figure 5.20 : Q-value curves of PC+Delta feature values (a) 0x436a81+0 and (b) 0x4377c5+0 in 459.GemsFDTD-1320B .
Figure 5.21 : Performance and main memory bandwidth usage of prefetchers in Ligra-CC .
Figure 5.22 : Performance of the basic and strict Pythia configurations in the Ligra workloads.
Figure 5.23 : Performance of the basic and feature-optimized Pythia on the SPEC CPU2006 suite.
Figure 5.24 : Performance improvement in 483 single-core DPC4 traces.
Figure 5.25 : Performance improvement in 966 four-core DPC4 traces.
| Pythia’s area : mm2/core; Pythia’s power : mW/core | ||
|---|---|---|
| Overhead compared to real systems | Area | Power |
| 4-core Skylake D-2123IT, 60W TDP [ 27 ] | 1.03% | 0.37% |
| 18-core Skylake 6150, 165W TDP [ 28 ] | 1.24% | 0.60% |
| 28-core Skylake 8180M, 205W TDP [ 30 ] | 1.33% | 0.75% |
Table 5.8: Area and power overhead of Pythia.
Figure 6.1 : The distribution of ROB-blocking and non-blocking load requests (on the left y-axis), and LLC misses per kilo instructions (on the right y-axis) in the absence and presence of a state-of-the-art hardware data prefetcher [ 136 ] .
Figure 6.2 : The average number of cycles a core stalls due to an off-chip load blocking any instruction from retiring from the ROB across all workload categories. The dark portion in each bar shows the cycles that can be completely eliminated by removing the on-chip cache access latency from an off-chip load’s critical path.
Figure 6.3 : Comparison of the execution timeline of an off-chip load request in a conventional processor and in Hermes.
Figure 6.4 : (a) Speedup of Ideal Hermes by itself and when combined with Pythia in single-core workloads. (b) Speedup of Ideal Hermes when combined with four recently-proposed prefetchers: Bingo [ 120 ] , SPP [ 465 , 148 ] , MLOP [ 761 ] , and SMS [ 792 ] .
Figure 6.5 : Percentage of loads that miss the LLC and goes off-chip (on the left y-axis) and the LLC MPKI (on the right y-axis) in the baseline system with Pythia.
Figure 6.6 : Overview of Hermes.
Figure 6.7 : Stages to make a prediction by POPET.
| Features without control-flow information | Features with control-flow information |
| 1. Load virtual address 2. Virtual page number 3. Cacheline offset in page 4. First access 5. Cacheline offset + first access 6. Byte offset in cacheline 7. Word offset in cacheline | 8. Load PC 9. PC load virtual address 10. PC virtual page number 11. PC cacheline offset 12. PC + first access 13. PC byte offset 14. PC word offset 15. Last- load PCs 16. Last- PCs |
Table 6.1: The initial set of program features used for automated feature selection. represents a bitwise XOR operation.
| Selected features | • PC cacheline offset • PC byte offset • PC + first access • Cacheline offset + first access • Last- load PCs |
| Threshold values | , , |
Table 6.2: POPET configuration parameters.
| Structure | Description | Size |
| POPET | • Perceptron weight tables – PC cacheline offset: – PC byte offset: – PC + first access: – Cacheline offset + first access: – Last- load PCs: • Page buffer: | 3.2 KB |
| LQ Metadata | Hashed PC: ; Last-4 PC: ; First access: ; perceptron weight: ; prediction: | 0.8 KB |
| Total | 4.0 KB | |
Table 6.3: Storage overhead of Hermes.
| Core | 1 and 8 cores, 6-wide fetch/execute/commit, 512-entry ROB, 128/72-entry LQ/SQ, Perceptron branch predictor [ 411 ] with 17-cycle misprediction penalty |
|---|---|
| L1/L2 Caches | Private, 48KB/1.25MB, 64B line, 12/20-way, 16/48 MSHRs, LRU, 5/15-cycle round-trip latency [ 60 ] |
| LLC | 3MB/core, 64B line, 12 way, 64 MSHRs/slice, SHiP [ 876 ] , 55-cycle round-trip latency [ 56 , 60 ] , Pythia prefetcher [ 136 ] |
| Main Memory | 1C: 1 channel, 1 rank per channel; 8C: 4 channels, 2 ranks per channel; 8 banks per rank, DDR4-3200 MTPS, 64b data-bus per channel, 2KB row buffer per bank, tRCD=12.5ns, tRP=12.5ns, tCAS=12.5ns |
| Hermes | Hermes-O/P : / -cycle Hermes request issue latency |
Table 6.4: Simulated system parameters.
| Suite | # Workloads | # Traces | Example Workloads |
|---|---|---|---|
| SPEC06 | 14 | 22 | gcc, mcf, cactusADM, lbm, … |
| SPEC17 | 11 | 23 | gcc, mcf, pop2, fotonik3d, … |
| PARSEC | 4 | 12 | canneal, facesim, raytrace, … |
| Ligra | 11 | 20 | BFS, PageRank, Radii, … |
| CVP | 33 | 33 | integer, floating-point, server, … |
Table 6.5: Workloads used for evaluation.
| HMP [ 905 ] with local, gshare, and gskew predictors | 11 KB |
| TTP with a metadata budget similar to the L2 cache | 1536 KB |
| Pythia [ 136 ] with the same configuration in [ 136 ] | 25.5 KB |
| Bingo [ 120 ] with the same configuration in [ 120 ] | 46 KB |
| SPP [ 465 ] with perceptron-based prefetch filter [ 148 ] | 39.3 KB |
| MLOP [ 761 ] with the same configuration in [ 761 ] | 8 KB |
Table 6.6: Storage overhead of all evaluated mechanisms.
Figure 6.8 : Comparison of (a) accuracy and (b) coverage of POPET against those of HMP [ 905 ] and TTP.
Figure 6.9 : The accuracy and coverage of POPET using each program feature individually and in various combinations.
Figure 6.10 : Line graph of POPET’s (a) accuracy and (b) coverage using each of the five program features individually across all 110 single-core workloads. No single feature can provide the best accuracy or coverage across all workloads.
Figure 6.11 : Speedup in single-core workloads.
Figure 6.12 : Single-core performance of all 110 workloads.
Figure 6.13 : Speedup of Hermes with three off-chip load predictors (HMP, TTP, and POPET) and the Ideal Hermes.
Figure 6.14 : (a) Reduction in stall cycles caused by off-chip loads. (b) Overhead in the main memory requests.
Figure 6.15 : Speedup in eight-core workloads.
Figure 6.16 : Performance sensitivity to (a) main memory bandwidth and (b) baseline prefetcher. The baseline system configuration is highlighted in green. Other highlighted configurations closely match with various commercial processors [ 29 , 5 , 4 ] .
Figure 6.17 : Performance sensitivity to (a) Hermes request issue latency and (b) on-chip cache hierarchy access latency. The baseline system configuration is highlighted in green.
Figure 6.18 : Performance sensitivity to (a) reorder buffer size and (b) LLC size. The baseline system configuration is highlighted in green.
Figure 6.19 : Effect of the activation threshold on POPET’s accuracy and coverage (on the left y-axis) and Hermes’s speedup (on the right y-axis) for all single-core workloads.
Figure 6.20 : (a) Variation of off-chip load prediction accuracy and coverage and (b) the increase in the main memory requests with different data prefetchers.
Figure 6.21 : Performance of Hermes and Pythia+Hermes while varying Hermes request issue latency.
Figure 6.22 : Speedup in 483 single-core DPC4 workload traces.
Figure 6.23 : (a) Speedup of Hermes with three different off-chip predictors, HMP, TTP, and POPET in single-core DPC4 workloads. Comparison of (b) accuracy and (c) coverage of POPET against those of HMP and TTP in single-core DPC4 workloads.
Figure 6.24 : Performance sensitivity to baseline prefetcher in single-core DPC4 workloads.
Figure 6.25 : Processor power consumption of Hermes, Pythia, and Hermes combined with Pythia.
Figure 7.1 : Performance line graph of a state-of-the-art off-chip predictor (OCP), POPET [ 133 ] , and a state-of-the-art prefetcher, Pythia [ 136 ] , across 100 workloads.
Figure 7.2 : Geomean speedup of POPET, Pythia, Naive, and StaticBest combinations across all workloads.
Figure 7.3 : Fraction of prefetch fills from off-chip main memory that are inaccurate.
Figure 7.4 : Geomean speedup of Naive, HPAC, MAB, and StaticBest combinations across all workloads.
Figure 7.5 : High-level overview of Athena as an RL agent.
| Feature | Measurement | Rationale |
|---|---|---|
| Prefetcher accuracy | Effectiveness of prefetching | |
| OCP accuracy | Effectiveness of OCP | |
| Bandwidth usage | Memory bus pressure | |
| Cache pollution | Interference caused by prefetches | |
Table 7.1: Candidate features considered for Athena’s state representation.
Algorithm 2 Q-value-driven prefetcher aggressiveness control
| Reward Component | Constituent Metric | Weight |
|---|---|---|
| # Cycles | ||
| # LLC misses | ||
| LLC miss latency | ||
| # Load instructions | ||
Table 7.2: Constituent metrics of Athena’s reward.
Figure 7.6 : Organization of QVStore.
| Category | Final Values |
|---|---|
| Selected Features | (1) prefetcher accuracy, (2) OCP accuracy, (3) bandwidth usage, (4) prefetch-induced cache pollution |
| Reward Weights | , , , , |
| Hyperparameters | , , , , Epoch length ( ) = instructions |
Table 7.3: Final Athena configuration derived through automated design-space exploration.
| Structure | Description | Size |
| QVStore | # planes = , # rows = | 2 KB |
| # columns = , entry size = bits | ||
| Accuracy Tracker | -bit Bloom filter, hashes | 0.5 KB |
| Pollution Tracker | -bit Bloom filter, hashes | 0.5 KB |
Table 7.4: Storage overhead of Athena.
| Core | 6-wide fetch/issue/commit, 512-entry ROB, 128-entry LQ, 72-entry SQ, perceptron branch predictor [ 411 ] , 17-cycle misprediction penalty |
|---|---|
| L1I/D | Private, 48KB, 64B line, 12-way, 16 MSHRs, LRU, 4/5-cycle round-trip latency |
| L2C | Private, 1.25MB, 64B line, 20-way, 48 MSHRs, LRU, 15-cycle round-trip latency [ 60 ] |
| LLC | Shared, 3MB/core, 64B line, 12-way, 64 MSHRs/slice, SHiP [ 876 ] , 55-cycle round-trip latency [ 56 , 60 ] |
| Main Memory | 1 rank per channel, 8 banks per rank, 64-bit data bus, 2KB row buffer, = = = 12.5ns; DDR4 with 3.2 GB/s per core |
Table 7.5: Simulated system parameters.
| Suite | # Workloads | Example Workloads |
|---|---|---|
| SPEC CPU06 | 29 | astar, leslie3d, libquantum, milc, omnetpp, sphinx3, soplex … |
| SPEC CPU17 | 20 | bwaves, cactuBSSN, fotonik3d, gcc, lbm, mcf, xalancbmk … |
| Ligra | 13 | BC, BFS, BFSCC, PageRankDelta, Radii, … |
| PARSEC | 13 | canneal, facesim, fluidanimate, raytrace, streamcluster … |
| CVP | 25 | integer, floating point, … |
Table 7.6: Workloads used for evaluation.
| Cache Design | Description | Comparison Points |
|---|---|---|
| CD1 | OCP + 1 L2C prefetcher | HPAC, MAB |
| CD2 | OCP + 1 L1D prefetcher | HPAC, MAB, TLP |
| CD3 | OCP + 2 L2C prefetchers | HPAC, MAB |
| CD4 | OCP + 1 L1D + 1 L2C prefetcher | HPAC, MAB, TLP |
Table 7.7: Evaluated cache designs (CD) and corresponding comparison points.
| Prefetchers | L1D | IPCP [ 628 ] , as an L1D-only prefetcher | 0.7 KB |
| Berti , as configured in [ 604 ] | 2.55 KB | ||
| L2C | Pythia , as configured in [ 136 ] | 25.5 KB | |
| SPP+PPF , as configured in [ 465 , 148 ] | 39.3 KB | ||
| MLOP , as configured in [ 761 ] | 8 KB | ||
| SMS , as configured in [ 792 ] | 20 KB | ||
| OCPs | POPET [ 133 ] , with 5 features | 4 KB | |
Table 7.8: Storage overhead of all evaluated mechanisms.
Figure 7.7 : Speedup in cache design 1 (CD1).
Figure 7.8 : (a) Workload category-wise performance analysis in CD1. (b) Performance comparison with StaticBest in CD1.
Figure 7.9 : Comparison of (a) the number of main memory requests and (b) average last-level cache (LLC) load miss latency.
Figure 7.10 : Speedup in cache design 2 (CD2).
Figure 7.11 : Speedup in cache design 3 (CD3).
Figure 7.12 : Speedup in cache design 4 (CD4).
Figure 7.13 : Performance sensitivity to underlying prefetching mechanism at L2C in CD1.
Figure 7.14 : Performance sensitivity to off-chip prediction mechanisms in CD1.
Figure 7.15 : Performance sensitivity off-chip predicted request issue latency in CD1.
Figure 7.16 : Performance sensitivity to prefetching mechanism at L1D in CD4.
Figure 7.17 : Performance sensitivity to main memory bandwidth in CD4.
Figure 7.18 : Speedup in four-core workloads.
Figure 7.19 : Speedup in eight-core workloads.
Figure 7.20 : Distribution of Athena’s action in coordinating Pythia and POPET and speedup of different Pythia-POPET combinations in compute_fp_78 workload from CVP suite, while varying memory bandwidth: 3.2 GB/s and 25.6 GB/s.
Figure 7.21 : Contribution of individual state features and reward components to Athena’s geomean performance across all 100 workloads.
Figure 7.22 : Geomean performance of Athena coordinating two L2C prefetchers without OCP.
Figure 7.23 : Speedup in (a) CD1 and (b) CD2 across 483 DPC4 traces.
Figure 7.24 : Speedup in (a) CD3 and (b) CD4 across 483 DPC4 traces.
Figure 8.1 : Execution timeline of a code example in a processor (a) without a load value predictor (LVP), (b) with LVP, and (c) with LVP and load elimination.
Figure 8.2 : (a) Fraction of dynamic loads that are global-stable. Distribution of global-stable loads by their (b) addressing mode and (c) inter-occurrence distance. (d) Distribution of inter-occurrence distance of global-stable loads from each addressing mode.
Figure 8.3 : Code example and disassembly from 541.leela_r of SPEC CPU 2017 suite. The highlighted load instructions are global-stable.
Figure 8.4 : Code example and disassembly from 557.xz_r of SPEC CPU 2017 suite. The highlighted load instructions are global-stable.
Figure 8.5 : Fraction of all dynamic loads that are global-stable in workloads compiled without and with APX (on the left y-axis) and the reduction in dynamic loads by APX (on the right y-axis).
Figure 8.6 : Distribution of global-stable loads by their addressing modes in workloads without and with APX. Each number on the x-axis corresponds to the respective workload from SPEC CPU 2017 suite.
Figure 8.7 : (a) Fraction of total execution cycles where at least one load port is utilized (we call such cycles load-utilized ). (b) Categorization of load-utilized cycles based on whether or not a global-stable load utilizes a load port.
Figure 8.8 : Speedup of Ideal Constable against Ideal Stable LVP and a processor with load execution width of the baseline.
Figure 8.9 : Overview of Constable.
Figure 8.10 : (a) Average number of SLD updates per cycle during rename stage. (b) Change in performance when Constable’s structures are updated only by correct path instructions vs. all instructions without updating Constable’s structures on branch misprediction recovery.
Figure 8.11 : An illustrative example of Constable’s operation.
| Structure | Description | Size |
| SLD | • # entries: ( sets ways) • Entry size: tag ( ) + addr ( ) + val ( ) + confidence level ( ) + can_eliminate flag ( ) | 7.9 KB |
| RMT | • load PCs for each stack registers (RSP and RBP) • load PCs for each remaining architectural registers in x86-64 | 0.4 KB |
| AMT | • # entries: 256 ( sets ways) • Entry size: physical address tag ( ) + # hashed load PCs ( ) | 4.0 KB |
Table 8.1: Storage overhead of Constable.
| Basic | x86-64 core clocked at 3.2 GHz with 2-way SMT support |
|---|---|
| Fetch & Decode | 8-wide fetch, TAGE/ITTAGE branch predictors [ 757 ] , 20-cycle misprediction penalty, 32KB 8-way L1-I cache, 4K-entry 8-way micro-op cache, 6-wide decode, 144-entry IDQ, loop-stream detector [ 150 ] |
| Rename | 6-wide, 288 integer, 220 512-bit and 320 256-bit physical registers, Memory Renaming [ 844 ] , zero elimination [ 300 ] , move elimination [ 300 , 278 ] , constant folding [ 300 , 278 ] , branch folding [ 251 ] |
| Allocate | 512-entry ROB, 240-entry LB, 112-entry SB, 248-entry RS |
| Issue & Retire | 6-wide issue to 12 execution ports; 5, 3, 2, and 2 ports for ALU, load, store-address, and store-data execution. Port 0, 1, and 5 are used for vector instructions. Aggressive out-of-order load scheduling with memory dependence prediction [ 220 , 569 ] , 6-wide retire |
| Caches | L1-D : 48KB, 12-way, 5-cycle latency, LRU, PC-based stride prefetcher [ 302 ] ; L2 : 2MB, 16-way, 12-cycle round-trip latency, LRU, stride + streamer [ 201 ] + SPP [ 465 ] ; LLC : 3MB, 12-way, 50-cycle data round trip latency [ 44 , 133 ] , dead-block-aware replacement policy [ 452 ] , streamer, MESIF [ 564 ] protocol |
Table 8.2: Simulation parameters. IDQ: Instruction Decode Queue, SB: Store Buffer.
| Component | Port count | Read access energy (pJ) | Write access energy (pJ) | Leakage power (mW) | Area (mm2) |
|---|---|---|---|---|---|
| SLD | R/ W | ||||
| RMT | R/ W | ||||
| AMT | R/ W | ||||
Table 8.3: Access energy, leakage power, and area estimates of Constable’s structures in 14nm technology. R: read port, W: write port.
| Suite | #Workloads | #Traces | Example Workloads |
| Client | 16 | 22 | DaCapo [ 12 ] , SYSmark [ 58 ] , TabletMark [ 59 ] , JetStream2 [ 34 ] |
| Enterprise | 9 | 14 | SPECjEnterprise [ 55 ] , SPECjbb [ 54 ] , LAMMPS [ 36 ] |
| FSPEC17 | 13 | 29 | All from SPECrate FP 2017 [ 53 ] |
| ISPEC17 | 10 | 11 | All from SPECrate Integer 2017 [ 53 ] |
Table 8.4: Workloads used for evaluation.
| Mechanism | Overhead |
| EVES [ 754 ] with the same configuration as the winner of CVP-1 KB storage budget track [ 15 ] | 32 KB |
| ELAR [ 131 ] with an additional adder in decode stage and a new copy of the ESP register | Adder + 64-b register |
| RFP [ 770 ] with 2K-entry prefetch table, 64-entry page address table, and 128-entry RFP-Inflight table | 12.4 KB |
| Constable (this work) | 15.4 KB |
Table 8.5: Overhead of all evaluated mechanisms.
Figure 8.12 : Speedup over the baseline (noSMT).
Figure 8.13 : Speedup of all workloads (noSMT).
Figure 8.14 : Speedup of Constable by eliminating execution of only PC-relative, stack-relative, and register-relative loads.
Figure 8.15 : Speedup over the baseline (SMT2).
Figure 8.16 : Speedup of Constable over ELAR and RFP.
Figure 8.17 : Load coverage of Constable versus EVES.
Figure 8.18 : Breakdown of eliminated and non-eliminated loads as fractions of global-stable loads.
Figure 8.19 : Reduction in (a) RS allocations and (b) L1-D accesses.
Figure 8.20 : (a) Overall core power consumption normalized to baseline. Expanded view of (b) OOO power and (c) MEU power.
Figure 8.21 : Performance sensitivity to (a) load execution width, and (b) pipeline depth.
Figure 8.22 : (a) Fraction of loads eliminated by Constable that violate memory ordering. (b) Increase in instructions allocated to ROB in presence of Constable.
Figure 8.23 : (a) Speedup and (b) coverage of Constable with AMT invalidation on L1D eviction compared to a vanilla Constable.