How to keep pushing ML accelerator performance? Know your rooflines! | Paper Citation Record

Observation c44066fd-fa82-459e-ac9a-e92d9805bfd9 · outbound

This paper cites Visualizing size of large language models,.

How to keep pushing ML accelerator performance? Know your rooflines! Visualizing size of large language models,

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:29.398472Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 14d25844-d249-4f4c-977e-e25c5c79aa5c · outbound

This paper cites Trends in deep learning hardware,.

How to keep pushing ML accelerator performance? Know your rooflines! Trends in deep learning hardware,

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:32.138137Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 7086185c-8b13-4bc9-9546-2fd5afc2a679 · outbound

This paper cites Roofline: an insightful visual performance model for multicore architectures,.

How to keep pushing ML accelerator performance? Know your rooflines! Roofline: an insightful visual performance model for multicore architectures,

Reference 3

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:32.113461Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 83c26e29-7cd7-4049-b5c9-4c43790a0f65 · outbound

This paper cites Eyeriss: A apatial architecture for energy-efficient dataflow for convolutional neural networks,.

How to keep pushing ML accelerator performance? Know your rooflines! Eyeriss: A apatial architecture for energy-efficient dataflow for convolutional neural networks,

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:32.096760Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation dba73b76-278c-45fe-8064-49a7104e97cc · outbound

This paper cites Roofline performance analysis of dnn architectures on cpu and gpu systems,.

How to keep pushing ML accelerator performance? Know your rooflines! Roofline performance analysis of dnn architectures on cpu and gpu systems,

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:32.079775Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 3bc34efd-f60a-415e-b5a8-0faad797ffe0 · outbound

This paper cites an unresolved cited work.

How to keep pushing ML accelerator performance? Know your rooflines! Unresolved cited work

Reference 6

Resolution: unresolved; raw_fallback, observed 2026-08-07T15:06:32.048841Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 5ebbde3f-f00f-4f35-a464-fe03ea9f0e38 · outbound

This paper cites Understanding reuse, performance, and hardware cost of dnn dataflow: A data-centric approach,.

How to keep pushing ML accelerator performance? Know your rooflines! Understanding reuse, performance, and hardware cost of dnn dataflow: A data-centric approach,

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:32.028318Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 191aa578-ee30-4289-a75c-2836b4ec5a45 · outbound

This paper cites A roofline model of energy,.

How to keep pushing ML accelerator performance? Know your rooflines! A roofline model of energy,

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:32.008663Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 2db46215-a0bb-4c48-84f7-8a8d5ad0049d · outbound

This paper cites Symphony: Orchestrating sparse and dense tensors with hierarchical heterogeneous processing,.

How to keep pushing ML accelerator performance? Know your rooflines! Symphony: Orchestrating sparse and dense tensors with hierarchical heterogeneous processing,

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.991855Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 79867109-ebbf-4d1d-8b49-b0c9761378ec · outbound

This paper cites Lots of questions on Google’s “Trillium.

How to keep pushing ML accelerator performance? Know your rooflines! Lots of questions on Google’s “Trillium

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.976054Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation fcbb83b6-24a5-426b-b612-a27c077aa0f7 · outbound

This paper cites Envi- sion: A 0.26-to-10tops/w subword-parallel dynamic-voltage-accuracy- frequency-scalable convolutional neural network processor in 28nm fdsoi,.

How to keep pushing ML accelerator performance? Know your rooflines! Envi- sion: A 0.26-to-10tops/w subword-parallel dynamic-voltage-accuracy- frequency-scalable convolutional neural network processor in 28nm fdsoi,

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.959854Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 03919dc5-ba5d-4c84-a46c-846a3f0ba0cc · outbound

This paper cites 9.5 a 6k-mac feature-map-sparsity-aware neural processing unit in 5nm flagship mobile soc,.

How to keep pushing ML accelerator performance? Know your rooflines! 9.5 a 6k-mac feature-map-sparsity-aware neural processing unit in 5nm flagship mobile soc,

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.939247Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 563c3ace-a04a-40e0-b028-5fc467f91985 · outbound

This paper cites Compute solution for tesla’s full self-driving computer,.

How to keep pushing ML accelerator performance? Know your rooflines! Compute solution for tesla’s full self-driving computer,

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.921995Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation a7c94c0d-2efb-4268-8c84-19ebc1307c38 · outbound

This paper cites 7.2 a 12nm programmable convolution-efficient neural- processing-unit chip achieving 825tops,.

How to keep pushing ML accelerator performance? Know your rooflines! 7.2 a 12nm programmable convolution-efficient neural- processing-unit chip achieving 825tops,

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.902241Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation d74fe782-4893-4413-82b4-4950a2811a22 · outbound

This paper cites Groq rocks neural networks,.

How to keep pushing ML accelerator performance? Know your rooflines! Groq rocks neural networks,

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.883652Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation db943f5f-3a0c-4ed3-b6d0-bdf3d7e97261 · outbound

This paper cites 9.1 a 7nm 4-core ai chip with 25.6tflops hybrid fp8 training, 102.4tops int4 inference and workload-aware throttling,.

How to keep pushing ML accelerator performance? Know your rooflines! 9.1 a 7nm 4-core ai chip with 25.6tflops hybrid fp8 training, 102.4tops int4 inference and workload-aware throttling,

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.859940Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 21688fde-3472-4885-9133-5e67b0f4bbc7 · outbound

This paper cites 16.7 a 40-310tops/w sram-based all-digital up to 4b in-memory computing multi-tiled nn accelerator in fd-soi 18nm for deep-learning edge applications,.

How to keep pushing ML accelerator performance? Know your rooflines! 16.7 a 40-310tops/w sram-based all-digital up to 4b in-memory computing multi-tiled nn accelerator in fd-soi 18nm for deep-learning edge applications,

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.841139Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 64e534e9-7d1c-4f72-aba6-e2a83c65919d · outbound

This paper cites Charm: Composing heterogeneous accelerators for matrix multiply on versal acap architecture,.

How to keep pushing ML accelerator performance? Know your rooflines! Charm: Composing heterogeneous accelerators for matrix multiply on versal acap architecture,

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:29.512744Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 074e189b-9dd1-4e3d-bd31-b2e6bb4362eb · outbound

This paper cites Davinci: A scalable architecture for neural network computing,.

How to keep pushing ML accelerator performance? Know your rooflines! Davinci: A scalable architecture for neural network computing,

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.821177Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 70b5df40-95fa-47b8-b99e-193403ae7e95 · outbound

This paper cites Nvidia tensor core programmability, performance & precision,.

How to keep pushing ML accelerator performance? Know your rooflines! Nvidia tensor core programmability, performance & precision,

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:29.538744Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a7876e8c-e647-4b9b-b903-617d5ecc35ae · outbound

This paper cites A charge domain sram compute-in-memory macro with c-2c ladder- based 8-bit mac unit in 22-nm finfet process for edge inference,.

How to keep pushing ML accelerator performance? Know your rooflines! A charge domain sram compute-in-memory macro with c-2c ladder- based 8-bit mac unit in 22-nm finfet process for edge inference,

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:29.557111Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8aa39835-d1be-4fdd-968a-736471b5bba6 · outbound

This paper cites A 22 nm, 1540 top/s/w, 12.1 top/s/mm 2 in-memory analog matrix-vector-multiplier for dnn acceleration,.

How to keep pushing ML accelerator performance? Know your rooflines! A 22 nm, 1540 top/s/w, 12.1 top/s/mm 2 in-memory analog matrix-vector-multiplier for dnn acceleration,

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.775674Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 060dfa3f-6df2-41e0-8d36-efaafad155e2 · outbound

This paper cites A 64-tile 2.4- mb in-memory-computing cnn accelerator employing charge-domain compute,.

How to keep pushing ML accelerator performance? Know your rooflines! A 64-tile 2.4- mb in-memory-computing cnn accelerator employing charge-domain compute,

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.752647Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 0b0afbc5-c8bf-4721-b4fb-82278323b147 · outbound

This paper cites Compute Solution for Tesla’s Full Self-Driving Computer,.

How to keep pushing ML accelerator performance? Know your rooflines! Compute Solution for Tesla’s Full Self-Driving Computer,

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.733368Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation fae8ddfc-2b2d-4e3a-9f0c-497ee270e85d · outbound

This paper cites Hardware for deep learning,.

How to keep pushing ML accelerator performance? Know your rooflines! Hardware for deep learning,

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.715995Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation da16470b-2d3e-4fe0-84a9-b071dd26d719 · outbound

This paper cites Lincoln ai computing survey (laics) update,.

How to keep pushing ML accelerator performance? Know your rooflines! Lincoln ai computing survey (laics) update,

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.701147Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 901acd9a-d4f1-4ea0-ac01-b5ca945a06b6 · outbound

This paper cites Neural network accelerator comparison.

How to keep pushing ML accelerator performance? Know your rooflines! Neural network accelerator comparison

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.682603Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 97d55eaf-e1ea-4233-aa9f-aea58d0b400a · outbound

This paper cites LLM Inference Unveiled: Survey and Roofline Model Insights.

How to keep pushing ML accelerator performance? Know your rooflines! LLM Inference Unveiled: Survey and Roofline Model Insights

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:29.707975Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 08353473-84cd-4cb4-bd49-537fa186db6a · outbound

This paper cites Minifloats on risc-v cores: Isa extensions with mixed- precision short dot products,.

How to keep pushing ML accelerator performance? Know your rooflines! Minifloats on risc-v cores: Isa extensions with mixed- precision short dot products,

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.666883Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 8555f8ba-09e2-4b4a-af27-409f6d015b63 · outbound

This paper cites Cutie: Beyond petaop/s/w ternary dnn inference acceleration with better-than-binary energy efficiency,.

How to keep pushing ML accelerator performance? Know your rooflines! Cutie: Beyond petaop/s/w ternary dnn inference acceleration with better-than-binary energy efficiency,

Reference 31

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.651276Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 0e670c07-2879-40bf-b013-1f9a3a083047 · outbound

This paper cites Binareye: An always-on energy-accuracy-scalable binary cnn processor with all memory on chip in 28nm cmos,.

How to keep pushing ML accelerator performance? Know your rooflines! Binareye: An always-on energy-accuracy-scalable binary cnn processor with all memory on chip in 28nm cmos,

Reference 32

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.635016Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 0a3287c9-08c5-4d5d-a841-4c814fd7ab59 · outbound

This paper cites BitNet: Scaling 1-bit Transformers for Large Language Models.

How to keep pushing ML accelerator performance? Know your rooflines! BitNet: Scaling 1-bit Transformers for Large Language Models

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:29.752659Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1519180e-97b3-4911-bfc5-9e5d87d183b3 · outbound

This paper cites A 3 tops/w risc-v parallel cluster for inference of fine-grain mixed-precision quantized neural networks,.

How to keep pushing ML accelerator performance? Know your rooflines! A 3 tops/w risc-v parallel cluster for inference of fine-grain mixed-precision quantized neural networks,

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.615689Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 0ba24f90-6cda-4cec-98bc-13dd8c9fa289 · outbound

This paper cites Marsellus: A heterogeneous risc-v ai-iot end-node soc with 2–8 b dnn acceleration and 30%-boost adaptive body biasing,.

How to keep pushing ML accelerator performance? Know your rooflines! Marsellus: A heterogeneous risc-v ai-iot end-node soc with 2–8 b dnn acceleration and 30%-boost adaptive body biasing,

Reference 35

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.601169Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 99a00f12-8207-439c-b6f2-350163c31afb · outbound

This paper cites Microscaling Data Formats for Deep Learning.

How to keep pushing ML accelerator performance? Know your rooflines! Microscaling Data Formats for Deep Learning

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.167879Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2cf05c9b-d6e0-4f88-98ba-393fd718c14d · outbound

This paper cites Nvidia blackwell platform: Advancing generative ai and accelerated computing,.

How to keep pushing ML accelerator performance? Know your rooflines! Nvidia blackwell platform: Advancing generative ai and accelerated computing,

Reference 37

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.585901Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 1680c9ec-81fd-4344-8309-0f0dbff0fc78 · outbound

This paper cites Siracusa: A 16 nm heterogenous risc-v soc for extended reality with at-mram neural engine,.

How to keep pushing ML accelerator performance? Know your rooflines! Siracusa: A 16 nm heterogenous risc-v soc for extended reality with at-mram neural engine,

Reference 38

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.571402Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 3fe2bff3-6cba-4d4f-9df5-377dcb967997 · outbound

This paper cites Onyx: A 12nm 756 gops/w coarse-grained reconfigurable array for accelerating dense and sparse applications,.

How to keep pushing ML accelerator performance? Know your rooflines! Onyx: A 12nm 756 gops/w coarse-grained reconfigurable array for accelerating dense and sparse applications,

Reference 39

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.557095Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 811baf5b-cbc8-40a8-b6b8-58e74ad81a01 · outbound

This paper cites Learning N:M Fine-grained Structured Sparse Neural Networks From Scratch.

How to keep pushing ML accelerator performance? Know your rooflines! Learning N:M Fine-grained Structured Sparse Neural Networks From Scratch

Reference 40

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.316475Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fd4078a9-32f3-41e9-9fcb-d49e353faaa0 · outbound

This paper cites 3.2 the a100 datacenter gpu and ampere architecture,.

How to keep pushing ML accelerator performance? Know your rooflines! 3.2 the a100 datacenter gpu and ampere architecture,

Reference 41

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.543279Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 1b9ff632-a182-4763-8a18-bfe3f114406f · outbound

This paper cites Venom: A vectorized n: M format for unleashing the power of sparse tensor cores,.

How to keep pushing ML accelerator performance? Know your rooflines! Venom: A vectorized n: M format for unleashing the power of sparse tensor cores,

Reference 42

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.527977Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 442b8fd4-7f49-4dfd-bff1-240ef09e13e4 · outbound

This paper cites Occamy: A 432-core dual-chiplet dual-hbm2e 768-dp-gflop/s risc-v system for 8- to-64-bit dense and sparse computing in 12-nm finfet,.

How to keep pushing ML accelerator performance? Know your rooflines! Occamy: A 432-core dual-chiplet dual-hbm2e 768-dp-gflop/s risc-v system for 8- to-64-bit dense and sparse computing in 12-nm finfet,

Reference 43

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.511828Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation a916b1ac-aac4-455c-b223-d85dd214cb28 · outbound

This paper cites Neupims: Npu-pim heterogeneous acceleration for batched llm inferencing,.

How to keep pushing ML accelerator performance? Know your rooflines! Neupims: Npu-pim heterogeneous acceleration for batched llm inferencing,

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.338151Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 17058cdc-295b-4200-86ac-6371834add2f · outbound

This paper cites Inclusive-PIM: Hardware-Software Co-design for Broad Acceleration on Commercial PIM Architectures.

How to keep pushing ML accelerator performance? Know your rooflines! Inclusive-PIM: Hardware-Software Co-design for Broad Acceleration on Commercial PIM Architectures

Reference 45

Resolution: verified exact; local_arxiv, observed 2026-08-07T15:06:30.773144Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation c504234d-02c1-4508-a794-fdaf31d7348c · outbound

This paper cites In-memory computation of a machine-learning classifier in a standard 6t sram array,.

How to keep pushing ML accelerator performance? Know your rooflines! In-memory computation of a machine-learning classifier in a standard 6t sram array,

Reference 46

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.496919Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation eabae999-f60c-4fe5-bc52-73e6730bdc6a · outbound

This paper cites An energy-efficient memory-based high-throughput vlsi architecture for convolutional networks,.

How to keep pushing ML accelerator performance? Know your rooflines! An energy-efficient memory-based high-throughput vlsi architecture for convolutional networks,

Reference 47

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.479682Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation dc024c9b-6911-468e-a603-4c7ed3dba6b1 · outbound

This paper cites Fast, energy-efficient, robust, and reproducible mixed-signal neuromorphic classifier based on embedded nor flash memory technology,.

How to keep pushing ML accelerator performance? Know your rooflines! Fast, energy-efficient, robust, and reproducible mixed-signal neuromorphic classifier based on embedded nor flash memory technology,

Reference 48

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.457417Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 1983e91c-efc9-4b7e-9918-6b05794fdb74 · outbound

This paper cites Analog in-memory subthreshold deep neural network accelerator,.

How to keep pushing ML accelerator performance? Know your rooflines! Analog in-memory subthreshold deep neural network accelerator,

Reference 49

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.441396Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation a97abc10-7d55-49d8-8883-98a46308033d · outbound

This paper cites A 5-nm 254-tops/w 221-tops/mm2 fully-digital computing-in-memory macro supporting wide-range dynamic-voltage- frequency scaling and simultaneous mac and write operations,.

How to keep pushing ML accelerator performance? Know your rooflines! A 5-nm 254-tops/w 221-tops/mm2 fully-digital computing-in-memory macro supporting wide-range dynamic-voltage- frequency scaling and simultaneous mac and write operations,

Reference 50

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.425267Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation dc0712c1-5f9f-4572-8f47-94f1c6f54a44 · outbound

This paper cites 16.4 an 89tops/w and 16.3tops/mm2 all-digital sram-based full-precision compute-in memory macro in 22nm for machine-learning edge applications,.

How to keep pushing ML accelerator performance? Know your rooflines! 16.4 an 89tops/w and 16.3tops/mm2 all-digital sram-based full-precision compute-in memory macro in 22nm for machine-learning edge applications,

Reference 51

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.410366Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation db9547c9-63ff-48aa-9301-808349548048 · outbound

This paper cites A maximally row- parallel mram in-memory-computing macro addressing readout circuit sensitivity and area,.

How to keep pushing ML accelerator performance? Know your rooflines! A maximally row- parallel mram in-memory-computing macro addressing readout circuit sensitivity and area,

Reference 52

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.395141Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation fd9d4dde-1f9e-452b-994a-62a1a717dc71 · outbound

This paper cites A programmable heterogeneous microprocessor based on bit-scalable in-memory comput- ing,.

How to keep pushing ML accelerator performance? Know your rooflines! A programmable heterogeneous microprocessor based on bit-scalable in-memory comput- ing,

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.383254Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6393fdbe-e573-4a73-bd6a-fc229cb2005d · outbound

This paper cites A crossbar array of magnetoresistive memory devices for in-memory computing,.

How to keep pushing ML accelerator performance? Know your rooflines! A crossbar array of magnetoresistive memory devices for in-memory computing,

Reference 54

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.387652Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7cd66dde-bc47-49df-a0eb-f1c9e5eba12e · outbound

This paper cites In-memory computing: Advances and prospects,.

How to keep pushing ML accelerator performance? Know your rooflines! In-memory computing: Advances and prospects,

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.392346Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 37ddb158-edd6-4020-8d8a-5accaacf7207 · outbound

This paper cites 14.2 a compute sram with bit-serial integer/floating-point operations for programmable in-memory vector acceleration,.

How to keep pushing ML accelerator performance? Know your rooflines! 14.2 a compute sram with bit-serial integer/floating-point operations for programmable in-memory vector acceleration,

Reference 56

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.354015Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation c4466db2-cdb4-4a65-b9c0-9ca6dfc39acd · outbound

This paper cites A 40nm 64kb 26.56tops/w 2.37mb/mm2rram binary/compute-in-memory macro with 4.23x im- provement in density and > 75% use of sensing dynamic range,.

How to keep pushing ML accelerator performance? Know your rooflines! A 40nm 64kb 26.56tops/w 2.37mb/mm2rram binary/compute-in-memory macro with 4.23x im- provement in density and > 75% use of sensing dynamic range,

Reference 57

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.337684Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 04b325b6-8759-454e-9434-474b467266b1 · outbound

This paper cites Funda- mental limits on energy-delay-accuracy of in-memory architectures in inference applications,.

How to keep pushing ML accelerator performance? Know your rooflines! Funda- mental limits on energy-delay-accuracy of in-memory architectures in inference applications,

Reference 58

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.318288Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 64a7278b-04c8-4782-a5a6-d54aab5e3df7 · outbound

This paper cites 11.3 metis aipu: A 12nm 15tops/w 209.6tops soc for cost- and energy-efficient inference at the edge,.

How to keep pushing ML accelerator performance? Know your rooflines! 11.3 metis aipu: A 12nm 15tops/w 209.6tops soc for cost- and energy-efficient inference at the edge,

Reference 59

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.302813Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 1f3e82c2-4fd0-45b6-be87-3f584cd7db5a · outbound

This paper cites Benchmarking in-memory computing architectures,.

How to keep pushing ML accelerator performance? Know your rooflines! Benchmarking in-memory computing architectures,

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.415453Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c2fb4c15-99b8-4560-b5be-7029727d1d2e · outbound

This paper cites A 22nm 128-kb mram row/column-parallel in-memory computing macro with memory- resistance boosting and multi-column adc readout,.

How to keep pushing ML accelerator performance? Know your rooflines! A 22nm 128-kb mram row/column-parallel in-memory computing macro with memory- resistance boosting and multi-column adc readout,

Reference 61

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.278613Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 8af9691b-3084-4a53-9b91-dc7aaad9beaf · outbound

This paper cites A 64-core mixed-signal in-memory compute chip based on phase-change memory for deep neural network inference,.

How to keep pushing ML accelerator performance? Know your rooflines! A 64-core mixed-signal in-memory compute chip based on phase-change memory for deep neural network inference,

Reference 62

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.425731Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e721fac0-ee09-4f4e-b3ae-d4b9c7e2f8d7 · outbound

This paper cites An n40 256k×44 embedded rram macro with sl-precharge sa and low-voltage current limiter to improve read and write performance,.

How to keep pushing ML accelerator performance? Know your rooflines! An n40 256k×44 embedded rram macro with sl-precharge sa and low-voltage current limiter to improve read and write performance,

Reference 63

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.250233Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 57ef3e98-bfbe-45af-9fe0-48b39b3e2f75 · outbound

This paper cites Cmos- embedded stt-mram arrays in 2x nm nodes for gp-mcu applications,.

How to keep pushing ML accelerator performance? Know your rooflines! Cmos- embedded stt-mram arrays in 2x nm nodes for gp-mcu applications,

Reference 64

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.234950Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 847322d6-5d32-4d10-b9f9-6b0dceb45256 · outbound

This paper cites A switched-capacitor sram in-memory computing macro with high-precision, high-efficiency differential archi- tecture,.

How to keep pushing ML accelerator performance? Know your rooflines! A switched-capacitor sram in-memory computing macro with high-precision, high-efficiency differential archi- tecture,

Reference 65

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.220998Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 499a3c28-1233-4957-b69f-4fa88b284e65 · outbound

This paper cites Scalable and Programmable Neural Network Inference Accelerator Based on In-Memory Computing,.

How to keep pushing ML accelerator performance? Know your rooflines! Scalable and Programmable Neural Network Inference Accelerator Based on In-Memory Computing,

Reference 66

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.206563Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation a78fcd0f-2f3d-4bea-892a-66ee8ff32747 · outbound

This paper cites Interstellar: Using Halide’s Scheduling Language to Analyze DNN Accelerators,.

How to keep pushing ML accelerator performance? Know your rooflines! Interstellar: Using Halide’s Scheduling Language to Analyze DNN Accelerators,

Reference 67

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.450465Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d5106f5f-7b45-4e8f-8825-d07f2c8fa3a4 · outbound

This paper cites MAESTRO: A Data-Centric Approach to Understand Reuse, Performance, and Hardware Cost of DNN Mappings,.

How to keep pushing ML accelerator performance? Know your rooflines! MAESTRO: A Data-Centric Approach to Understand Reuse, Performance, and Hardware Cost of DNN Mappings,

Reference 68

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.192491Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation abdc799a-4fe5-46f0-9623-1de4d6bb730f · outbound

This paper cites Timeloop: A Systematic Approach to DNN Accelerator Evaluation,.

How to keep pushing ML accelerator performance? Know your rooflines! Timeloop: A Systematic Approach to DNN Accelerator Evaluation,

Reference 69

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.179380Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation b8576801-923c-4513-acf6-f513d0ae90d5 · outbound

This paper cites ZigZag: Enlarging Joint Architecture-Mapping Design Space Exploration for DNN Accelerators,.

How to keep pushing ML accelerator performance? Know your rooflines! ZigZag: Enlarging Joint Architecture-Mapping Design Space Exploration for DNN Accelerators,

Reference 70

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.163852Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 2cc2e601-e74a-4511-9e51-3ef07c37610c · outbound

This paper cites CoSA: Scheduling by constrained op- timization for spatial accelerators,.

How to keep pushing ML accelerator performance? Know your rooflines! CoSA: Scheduling by constrained op- timization for spatial accelerators,

Reference 71

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.147252Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation b6ac7aae-c0da-4d9c-a770-63efc2c79417 · outbound

This paper cites Mind Mappings: Enabling Efficient Algorithm-Accelerator Mapping Space Search,.

How to keep pushing ML accelerator performance? Know your rooflines! Mind Mappings: Enabling Efficient Algorithm-Accelerator Mapping Space Search,

Reference 72

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.476277Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8ce0af98-da9c-489c-9b97-929ae3a283eb · outbound

This paper cites GAMMA: Automating the HW Mapping of DNN Models on Accelerators via Genetic Algorithm,.

How to keep pushing ML accelerator performance? Know your rooflines! GAMMA: Automating the HW Mapping of DNN Models on Accelerators via Genetic Algorithm,

Reference 73

Resolution: verified exact; doi, observed 2026-08-07T15:06:30.566889Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 2c87e16f-c313-489e-91ac-527b88f5fe22 · outbound

This paper cites Stream: Design space exploration of layer-fused dnns on hetero- geneous dataflow accelerators,.

How to keep pushing ML accelerator performance? Know your rooflines! Stream: Design space exploration of layer-fused dnns on hetero- geneous dataflow accelerators,

Reference 74

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.131537Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 8a588eea-947f-4fc9-90b8-ecc9ac56877d · outbound

This paper cites The groq software-defined scale-out tensor streaming multiprocessor : From chips-to-systems architectural overview,.

How to keep pushing ML accelerator performance? Know your rooflines! The groq software-defined scale-out tensor streaming multiprocessor : From chips-to-systems architectural overview,

Reference 75

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.117852Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 17a54891-ec19-44ba-bd33-97e9311fa29c · outbound

This paper cites Application specific instruction processor based implementation of a gnss receiver on an fpga,.

How to keep pushing ML accelerator performance? Know your rooflines! Application specific instruction processor based implementation of a gnss receiver on an fpga,

Reference 76

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.103230Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 1d680a15-8dd8-4544-a543-a003d2402db8 · outbound

This paper cites How flexible is your com- puting system?.

How to keep pushing ML accelerator performance? Know your rooflines! How flexible is your com- puting system?

Reference 77

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.088800Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation a9dedbe3-43ab-407c-aeec-1b8d2a4311e5 · outbound

This paper cites Tandem processor: Grappling with emerging operators in neural networks,.

How to keep pushing ML accelerator performance? Know your rooflines! Tandem processor: Grappling with emerging operators in neural networks,

Reference 78

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.501738Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cc7cdc0a-dd28-4d26-9de6-1a32daf336d9 · outbound

This paper cites Mec: memory-efficient convolution for deep neural network,.

How to keep pushing ML accelerator performance? Know your rooflines! Mec: memory-efficient convolution for deep neural network,

Reference 79

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.063778Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 196c9181-8b97-4077-97e0-4ae7c239699b · outbound

This paper cites A formalism of dnn accelerator flexibility,.

How to keep pushing ML accelerator performance? Know your rooflines! A formalism of dnn accelerator flexibility,

Reference 80

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.048331Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 84c6dd27-560d-4be8-a390-6d1be9a0e729 · outbound

This paper cites Mlir: Scaling compiler infrastructure for domain specific computation,.

How to keep pushing ML accelerator performance? Know your rooflines! Mlir: Scaling compiler infrastructure for domain specific computation,

Reference 81

Resolution: unresolved; no resolver link, observed 2026-08-07T15:06:30.515891Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d484f9f9-f15a-4eb2-988f-a8ac376c7f42 · outbound

This paper cites The hardware lottery,.

How to keep pushing ML accelerator performance? Know your rooflines! The hardware lottery,

Reference 82

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:06:31.023084Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Coverage vector

Reference resolution

Outbound references

Pith citing papers