See it to Place it: Evolving Macro Placements with Vision-Language Models

Ikechukwu Uchendu; Swati Goel; Karly Hou; Ebrahim Songhori; Kuang-Huei Lee; Joe Wenjie Jiang; Vijay Janapa Reddi; Vincent Zhuang

cs.LGMar 30, 2026

See it to Place it: Evolving Macro Placements with Vision-Language Models

Authors: Ikechukwu Uchendu, Swati Goel, Karly Hou, Ebrahim Songhori, Kuang-Huei Lee, Joe Wenjie Jiang, Vijay Janapa Reddi, Vincent Zhuang

Organizations: Harvard University, Cambridge, MA, USA · Kempner Institute for the Study of Natural and Artificial Intelligence, Harvard University, Cambridge, MA, USA · Google DeepMind

Abstract

We propose using Vision-Language Models (VLMs) for macro placement in chip floorplanning, a complex optimization task that has recently shown promising advancements through machine learning methods. Because human designers rely heavily on spatial reasoning to arrange components on the chip canvas, we hypothesize that VLMs with strong spatial reasoning capabilities can effectively complement existing placement algorithms. We introduce VeoPlace (Visual Evolutionary Optimization Placement), a novel framework that uses a VLM, without any fine-tuning, to guide the actions of a base placer by constraining it to subregions of the chip canvas. The VLM proposals are iteratively optimized through an evolutionary search strategy with respect to resulting placement quality. Using global half-perimeter wirelength (gHPWL) after standard-cell placement and legalization as the evaluation metric, VeoPlace boosts ChiPFormer performance on 9 of 10 benchmarks with peak reductions exceeding 32%. We further demonstrate that VeoPlace generalizes to analytical placers, improving DREAMPlace on all 8 evaluated Superblue benchmarks with gains up to 4.4%. Our approach opens new possibilities for electronic design automation tools that leverage foundation models to solve complex physical design problems.

Figures & tables

Panel of Figure 1 — Figure 1: VeoPlace-guided DREAMPlace on (a) superblue5 , (b) superblue16 , and (c) superblue18 . VP+DP denotes VeoPlace with DREAMPlace; DP 4.3.0 is standalone DREAMPlace. We report mean and standard error of the best gHPWL observed ( $\times 10^{7}$ , lower is better) across three random seeds with 2,000 rollouts each, using Gemini 2.5 Flash. The continued decrease in gHPWL demonstrates how VeoPlace improves placement quality over the fixed rollout budget, whereas standalone DREAMPlace converges early.

Figure 2: VeoPlace framework overview. The VLM suggests placement regions (1-2) to constrain a low-level placer (3) for macro placement (4). A history buffer that stores the existing population of placements (5) facilitates evolutionary in-context improvement, creating a feedback loop to improve placement quality.

Method	Category
DREAMPlace [ 25 ]	Analytical
NTUPlace3 [ 7 ]	Analytical
RePlace [ 8 ]	Analytical
SP-SA [ 29 ]	Packing
ChiPFormer [ 20 ]	IL
DeepPR [ 9 ]	RL

Table 1: Comparison of macro placement approaches. RL: Reinforcement Learning, MCTS: Monte Carlo Tree Search, IL: Imitation Learning, BBO: Blackbox Optimization, VLM: Vision-Language Model, LLM: Large Language Model.

Figure 3: VLM guidance visualization on superblue1 . Left: Final macro placement produced by VeoPlace + DREAMPlace 4.3.0. Right: The same placement with VLM-suggested regions overlaid. The VLM proposes target positions for selected macros, and DREAMPlace’s loss function is modified with anchor weights to keep macros close to VLM suggestions while allowing standard cells to optimize freely (standard cells removed for visual clarity).

Component	Description
Grid Repr.	• 84 $\times$ 84 grid for ChiPFormer experiments (matching [ 20 ] ); 512 $\times$ 512 grid for DREAMPlace experiments. • Macros positioned by bottom-left corner coordinates. • Origin (0,0) at the bottom-left.
Visual Repr.	• Image of canvas showing placed macros and colors. • Visual context for spatial relationships and patterns.
Context	• Grid specs and macro properties (dimensions, color). • History of prior placements with performance metrics. • Current state (locations) of placed macros.

Table 2: Components of the structured prompt for VeoPlace.

Benchmark	VP+DP	DP 4.3.0	DP 4.0	AutoDMP	WireMask-EA	MaskPlace	MaskRegulate
superblue1	37.29 $\pm$ 0.03 ( $-$ 1.0%)	37.66 $\pm$ 0.01	89.6 $\pm$ 8.4	81.3 $\pm$ 1.7	80.7 $\pm$ 3.8	79.3 $\pm$ 0.6	57.7 $\pm$ 0.5
superblue3	43.81 $\pm$ 0.04 ( $-$ 3.4%)	45.38 $\pm$ 0.03	128.7 $\pm$ 17.3	81.3 $\pm$ 6.9	93.7 $\pm$ 8.1	89.0 $\pm$ 1.7	70.5 $\pm$ 0.3
superblue4	28.68 $\pm$ 0.06 ( $-$ 1.4%)	29.09 $\pm$ 0.03	68.1 $\pm$ 2.3	45.7 $\pm$ 7.8	55.1 $\pm$ 0.7	52.8 $\pm$ 0.3	41.5 $\pm$ 0.6
superblue5	39.79 $\pm$ 0.04 ( $-$ 0.7%)	40.09 $\pm$ 0.03	87.8 $\pm$ 14.7	126.7 $\pm$ 40.9	102.3 $\pm$ 6.8	98.1 $\pm$ 0.3	69.4 $\pm$ 0.0
superblue7	53.88 $\pm$ 0.09 ( $-$ 1.0%)	54.45 $\pm$ 0.04	227.0 $\pm$ 9.1	100.4 $\pm$ 16.3	100.5 $\pm$ 3.4	99.9 $\pm$ 0.5	79.0 $\pm$ 0.3
superblue10	68.33 $\pm$ 0.12 ( $-$ 4.4%)	71.45 $\pm$ 0.12	148.1 $\pm$ 13.3	114.8 $\pm$ 17.8	135.2 $\pm$ 22.5	109.4 $\pm$ 2.1	112.3 $\pm$ 3.8

Table 3: VeoPlace-guided DREAMPlace (VP+DP) vs. DREAMPlace 4.3.0 (DP) and Superblue baselines reported by Xue et al. [46] . All columns report gHPWL (

\times 10^{7}

, lower is better). VP+DP uses Gemini 3.6 Flash and one global anchor weight

\lambda_{A}{=}0.01

on all designs. VP+DP/DP report mean

\pm

standard error of the best gHPWL observed across five seeds with 2,000 rollouts. Percentages compare VP+DP to matched DP 4.3.0.

	VP+CF	CF
Benchmark	$(C{=}1)$	No-FT	FT
adaptec1	15.97 $\pm$ 0.80 (+12.2%)	14.23 $\pm$ 0.77	14.99 $\pm$ 0.55
adaptec2	11.82 $\pm$ 0.89 ( $-$ 13.8%)	15.03 $\pm$ 0.59	13.71 $\pm$ 0.33
adaptec3	24.10 $\pm$ 0.21 ( $-$ 6.0%)	25.64 $\pm$ 0.43	26.31 $\pm$ 0.97
adaptec4	19.63 $\pm$ 0.20 ( $-$ 8.7%)	22.65 $\pm$ 1.69	21.50 $\pm$ 0.50
ibm01	0.31 $\pm$ 0.01 ( $-$ 22.5%)	0.44 $\pm$ 0.04	0.40 $\pm$ 0.01

Table 4: VeoPlace (VP) with ChiPFormer (CF). VP uses Gemini 2.5 Flash. Mean

\pm

SE of best gHPWL (

\times 10^{7}

, lower is better) across three random seeds with 2,000 rollouts each. Percentages compare VP+CF to the better CF baseline. Best results are bolded.

VLM (VP+DP)	sb1	sb3	sb4	sb5	sb7	sb10	sb16	sb18
DP 4.3.0	376.63 $\pm$ 0.13	453.75 $\pm$ 0.33	290.92 $\pm$ 0.26	400.88 $\pm$ 0.30	544.45 $\pm$ 0.35	714.47 $\pm$ 1.18	366.34 $\pm$ 0.12	219.24 $\pm$ 0.07
Gemini 3.6 Flash	372.85 $\pm$ 0.34	438.14 $\pm$ 0.42	286.82 $\pm$ 0.55	397.90 $\pm$ 0.41	538.76 $\pm$ 0.86	683.25 $\pm$ 1.22	364.43 $\pm$ 0.09	217.50 $\pm$ 1.09
Qwen3-VL 8B Instruct	374.52 $\pm$ 0.41	437.04 $\pm$ 0.93	284.03 $\pm$ 0.75	400.50 $\pm$ 0.26	539.85 $\pm$ 0.38	686.85 $\pm$ 0.51	364.46 $\pm$ 0.38	217.25 $\pm$ 0.28
Qwen3-VL 30B A3B Instruct	375.54 $\pm$ 0.59	432.98 $\pm$ 1.35	283.40 $\pm$ 0.84	400.29 $\pm$ 0.34	539.58 $\pm$ 0.74	686.25 $\pm$ 1.26	364.40 $\pm$ 0.39	217.59 $\pm$ 0.32

Table 5: VP+DP with open-weight VLMs on Superblue: gHPWL (

\times 10^{6}

, mean

\pm

standard error over five seeds, lower is better). All runs use 2,000 rollouts. Bold is best and underlined is second best. Both Qwen models are served locally with vLLM and use top-stratified history with

C{=}25

and the global anchor weight

\lambda_{A}{=}0.01

Method	Metric	Ariane-133	Ariane-136	BlackParrot	BP-Quad
Hier-RTLMP	Routed WL	6.91	7.46	6.11	52.56
	WNS	-0.411	0.432	-2.991	-150.056
	TNS	-0.719	0.000	-0.003	-38511.600
	Power	254.3	162.1	147.0	2008.7
DP 4.3.0	Routed WL	5.93 $\pm$ 0.19	6.43 $\pm$ 0.05	5.71 $\pm$ 0.02	45.91 $\pm$ 0.16
	WNS	-0.368 $\pm$ 0.041	0.268 $\pm$ 0.055	-2.990 $\pm$ 0.028	-149.712 $\pm$ 0.087

Table 6: Post-route PPA on four Nangate45 designs after the full OpenROAD flow. The columns are routed wirelength (WL,

10^{6}\,\mu

m), worst negative slack (WNS, ns), total negative slack (TNS,

10^{3}

ns), and power (mW). Table 15 reports the full set of metrics. VP+DP uses Gemini 3.5 Flash. Bold is best and underlined is second best. Error terms are standard deviations over three seeds. An asterisk ( ∗ ) marks a mean over two seeds, because the third seed failed to route in OpenROAD. Plain values are single runs.

Appendix figures & tables18 assets

Supplementary material from the paper’s appendix.

Appendix

Benchmark	# Objects	# Macros	# Nets	# Pins
ISPD 2005
adaptec1	211K	63	221K	944K
adaptec2	255K	127	266K	1.1M
adaptec3	452K	58	467K	1.9M
adaptec4	496K	69	516K	1.9M
ICCAD 2004

Appendix

Table 7: Benchmark statistics for all evaluated circuits. IBM benchmarks do not distinguish macros from standard cells; their macro counts come from DREAMPlace, which treats a movable object as a macro when its area exceeds 10 times the mean movable-cell area and its height spans more than two standard-cell rows.

Panel of Figure 4 — Figure 4: Visual comparison of placements on selected Superblue benchmarks. Top row: DREAMPlace 4.3.0. Bottom row: VeoPlace-guided DREAMPlace. Blue clouds are individual standard cells; colored rectangles are macros. VeoPlace guidance improves gHPWL on all eight Superblue benchmarks.

Panel of Figure 5 — Figure 5: VeoPlace’s VLM-guided placement on adaptec4. (a) VLM proposes initial regions ( $t=0$ ); policy is unconstrained for macros without valid suggestions. (b) Mid-placement ( $t=T/2$ ). (c) Final placement ( $t=T$ ), with the policy operating within VLM constraints.

Panel of Figure 6 — Figure 6: Visual comparison of placements at different anchor weights on superblue1 . Lower anchor weights allow the analytical placer more flexibility to optimize standard cell positions around the macro placements.

Panel of Figure 7 — Figure 7: Visual comparison demonstrating why a fixed-macro flow is essential for a fair evaluation of VeoPlace. (a) The final layout from our two-stage flow, where macro placements generated by VeoPlace are fixed . (b) The result of an alternative three-stage flow, which takes the exact same initial placement from (a) as input but allows DREAMPlace to move all cells. The analytical placer’s drastic rearrangement of macros in (b) shows that it effectively ignores the initial solution, making its final metrics an invalid measure of VeoPlace’s contribution. Our fixed-macro approach ensures a direct and unconfounded evaluation. The flipped macro IDs are a DREAMPlace rendering artifact. We did not modify the DREAMPlace code.

Benchmark	DP rollout (s)	DP (h)	VP+DP (h)	Overhead
superblue1	83	46	48	+5%
superblue3	168	93	95	+2%
superblue4	127	70	72	+3%
superblue5	95	53	55	+4%
superblue7	122	68	70	+3%
superblue10	190	106	108	+2%

Appendix

Table 8: Wall-clock overhead for VeoPlace-guided DREAMPlace on Superblue. DP rollout is one vanilla DREAMPlace run; DP/VP+DP report total time for one 2,000-rollout run on one A100, with VP+DP including 125 Gemini 3.6 Flash queries at the measured 62 s per query.

Parameter	Value
Temperature	0.7
Top- $k$	64
Top- $p$	0.95
Candidates	8

Appendix

Table 9: Gemini API hyperparameters.

Parameter	Value	Description
Number of Groups	2000	The fixed number of clusters to group all standard cells into.
Cell Area Utilization	1.25	A target for the density of cells within a cluster.
Enable Group Breakup	True	A boolean flag that allows the algorithm to split larger groups.

Appendix

Table 10: Hyperparameters for the Standard Cell Grouping Algorithm.

Netlist	Return-to-go
adaptec1	-2.86E+06
adaptec2	-2.91E+06
adaptec3	-5.90E+06
adaptec4	-6.37E+06
ibm01	-7.00E+04
ibm02	-1.67E+05

Appendix

Table 11: Target Returns-to-Go for Different Benchmark Netlists

ICCAD 2015 Superblue
Benchmark	Target Density	Stop Overflow	Density Weight	Num Bins (X)	Num Bins (Y)	Iterations
superblue1	1.00	0.10	$8\times 10^{-5}$	1024	1024	1000
superblue3	1.00	0.10	$8\times 10^{-5}$	2048	2048	1000
superblue4	1.00	0.10	$8\times 10^{-5}$	512	512	1000
superblue5	1.00	0.10	$8\times 10^{-5}$	1024	1024	1000
superblue7	1.00	0.10	$8\times 10^{-5}$	512	512	1000

Appendix

Table 12: DREAMPlace 4.3.0 Hyperparameters for All Benchmarks

Benchmark	VP+DP 2.5 Flash	DP 4.3.0
superblue1	0.93 $\pm$ 0.01	0.92 $\pm$ 0.00
superblue3	1.06 $\pm$ 0.02	1.04 $\pm$ 0.00
superblue4	0.91 $\pm$ 0.02	0.93 $\pm$ 0.00
superblue5	0.85 $\pm$ 0.02	0.80 $\pm$ 0.00
superblue7	1.03 $\pm$ 0.00	1.06 $\pm$ 0.00
superblue10	1.08 $\pm$ 0.02	1.09 $\pm$ 0.01

Appendix

Table 13: Congestion proxy (RUDY, lower is better) for VeoPlace-guided DREAMPlace vs. DREAMPlace 4.3.0 on Superblue. VP+DP uses Gemini 2.5 Flash; mean

\pm

standard error across three seeds. Differences are small and mixed (4–4 split), indicating that VeoPlace’s gHPWL improvements do not materially change routability.

Benchmark	Seed 1	Seed 2	Seed 3	Seed 4	Seed 5
superblue1	4.26	4.37	2.41	4.01	3.86
superblue3	15.93	13.53	17.84	16.00	14.72
superblue4	2.47	3.91	4.36	4.55	5.18
superblue5	4.93	3.39	2.27	2.51	1.80
superblue7	6.39	9.14	6.59	3.09	3.25
superblue10	33.92	32.16	27.10	32.07	30.86

Appendix

Table 14: Per-seed paired gHPWL differences (

\times 10^{6}

), DREAMPlace 4.3.0 minus VP+DP, on Superblue. Positive means VP+DP is better. Green marks pairs where VP+DP is better, and red marks pairs where DREAMPlace 4.3.0 is better. Each column pairs the two runs that share that seed. The per-design means are the Table 3 gaps.

Method	Metric	Ariane-133	Ariane-136	BlackParrot	BP-Quad
Hier-RTLMP	Routed WL	6.91	7.46	6.11	52.56
	WNS	-0.411	0.432	-2.991	-150.056
	TNS	-0.719	0.000	-0.003	-38511.600
	NVP	2894	0	1	259747
	Power	254.3	162.1	147.0	2008.7
	Area	727.7	748.6	645.2	5402.6

Appendix

Table 15: Post-route PPA on four Nangate45 designs after the full OpenROAD flow. The metrics are routed wirelength (WL,

10^{6}\,\mu

m), worst negative slack (WNS, ns), total negative slack (TNS,

10^{3}

ns), number of violating pins (NVP), power (mW), area (

10^{3}\,\mu

m 2 ), and DRC violations. VP+DP uses Gemini 3.5 Flash. Bold is best and underlined is second best. Error terms are standard deviations over three seeds. An asterisk ( ∗ ) marks a mean over two seeds, because the third seed failed to route in OpenROAD. Plain values are single runs.

Panel of Figure 8 — Figure 8: Ablations on superblue1 with VeoPlace-guided DREAMPlace (Gemini 2.5 Flash, three seeds). Defaults: top stratified (TS), $C{=}10$ , $\lambda_{A}{=}0.01$ ; main Superblue results use $C{=}25$ . Lower anchor weights and longer context perform best; all context strategies improve over rollouts, with invalid suggestions converging to ${\sim}20\%$ .

Method (rollouts)	superblue1	superblue3	superblue4	superblue5	superblue7	superblue10	superblue16	superblue18
DP 4.3.0 (4,000)	376.61 $\pm$ 0.12	453.64 $\pm$ 0.27	290.90 $\pm$ 0.25	399.94 $\pm$ 0.25	544.36 $\pm$ 0.31	714.05 $\pm$ 1.30	366.21 $\pm$ 0.07	219.13 $\pm$ 0.05
VP+DP 3.6 Flash (2,000)	372.85 $\pm$ 0.34	438.14 $\pm$ 0.42	286.82 $\pm$ 0.55	397.90 $\pm$ 0.41	538.76 $\pm$ 0.86	683.25 $\pm$ 1.22	364.43 $\pm$ 0.09	217.50 $\pm$ 1.09
Random regions (4,000)	496.75 $\pm$ 3.99	584.81 $\pm$ 2.88	352.41 $\pm$ 0.34	599.84 $\pm$ 2.41	712.07 $\pm$ 2.53	748.64 $\pm$ 2.23	406.88 $\pm$ 1.27	236.62 $\pm$ 0.49
Best-so-far anchor (4,000)	382.79 $\pm$ 0.72	451.09 $\pm$ 1.20	289.09 $\pm$ 0.43	406.20 $\pm$ 0.74	544.93 $\pm$ 0.28	702.34 $\pm$ 2.69	366.55 $\pm$ 0.11	221.35 $\pm$ 0.31

Appendix

Table 16: VP+DP (Gemini 3.6 Flash) vs. DREAMPlace 4.3.0 (DP) and two evolutionary search controls. All columns report gHPWL (

\times 10^{6}

, lower is better), mean

\pm

standard error across five seeds. DP 4.3.0 and both controls run 4,000 rollouts. VP+DP runs 2,000 rollouts. Best results are bolded.

Greedy Prompt	Default Prompt	Exploratory Prompt
Preamble addition: “Use the previous examples as a strong prior. Stay very close to the best known historical arrangements and prefer small, low-risk refinements over novel departures.” Strategy section: “Refinement on Best Archetype: Propose exactly one small, low-risk improvement to the best known pattern. Stay very close to the successful historical arrangement and explain why this minimal change should help.”	Preamble addition: none. Strategy section: “Improvement on Best Archetype: Propose exactly one concrete change to the best pattern (e.g., shift position, adjust spacing, try different canvas region); if evidence is weak, label it ‘exploratory’ but still commit.”	Preamble addition: “Use the previous examples as context only, not as a constraint. Favor novel, diverse, and deliberately exploratory region proposals when they could plausibly lead to a better floorplan.” Strategy section: “Deliberate Departure from Best Archetype: Propose exactly one concrete way your plan intentionally differs from the best known pattern. Explain why this novel configuration could outperform it; if evidence is weak, label it an ‘exploratory hypothesis’ but still commit.”

Appendix

Figure 9: Strategy-specific prompt text extracted from the implemented prompt templates. The Greedy prompt adds stay-close instructions, the Default prompt keeps the neutral scaffold, and the Exploratory prompt adds explicit novelty-seeking guidance.

Benchmark	Method	Greedy	Default	Exploratory
superblue1	VP+DP 2.5 Flash	37.51 $\pm$ 0.09	37.47 $\pm$ 0.11	37.64 $\pm$ 0.11
	DP 4.3.0 (Baseline)	37.95 $\pm$ 0.16

Appendix

Table 17: Ablation study on the VLM prompt’s strategic guidance (TS,

C{=}25

) on the superblue1 benchmark. We report VeoPlace with DREAMPlace (VP+DP) using Gemini 2.5 Flash, mean

\pm

standard error across three seeds. The Default prompt yields the best objective (

\times 10^{7}

, lower is better). The best result is bolded .

Benchmark	DP 4.3.0	VP+DP Full	VP+DP NoImg	VP+DP NoTxt	VP+DP Shuffled
superblue1	376.63 $\pm$ 0.13	372.85 $\pm$ 0.34	374.18 $\pm$ 0.36	372.73 $\pm$ 0.39	373.54 $\pm$ 0.38
superblue3	453.75 $\pm$ 0.33	438.14 $\pm$ 0.42	438.31 $\pm$ 0.73	436.81 $\pm$ 0.22	438.01 $\pm$ 0.72
superblue4	290.92 $\pm$ 0.26	286.82 $\pm$ 0.55	285.88 $\pm$ 0.22	284.64 $\pm$ 0.59	285.91 $\pm$ 0.21
superblue5	400.88 $\pm$ 0.30	397.90 $\pm$ 0.41	397.91 $\pm$ 0.53	397.99 $\pm$ 0.96	398.42 $\pm$ 0.39
superblue7	544.45 $\pm$ 0.35	538.76 $\pm$ 0.86	537.35 $\pm$ 0.57	541.32 $\pm$ 0.57	537.64 $\pm$ 0.90
superblue10	714.47 $\pm$ 1.18	683.25 $\pm$ 1.22	684.01 $\pm$ 1.31	682.72 $\pm$ 0.51	681.75 $\pm$ 1.06

Appendix

Table 18: Input modality ablation on Superblue. The values are gHPWL (

\times 10^{6}

, mean

\pm

standard error over five seeds, lower is better). VP+DP uses Gemini 3.6 Flash. ‘Full’ is the default multimodal prompt with both image and text. ‘NoImg’ removes the canvas image from every in-context example. ‘NoTxt’ removes the textual context from every in-context example. ‘Shuffled’ keeps the images but pairs each one with a placement it did not come from. DREAMPlace 4.3.0 runs 2,000 rollouts. The best result per design is in bold.

Macro Placement

See all 8 papers

PlaceReasoner-Beta: Reasoning-Driven Macro Placement and Benchmarking
Sep 18, 2026Qiufeng Li, Chengxuan Wang, Rongqian Chen +6Macro Placement Electronic Design Automation
The George Washington University, USA · University of California, Los Angeles, USA · Brown University, USA +3
Order Matters: Unveiling the Hidden Impact of Macro Placement Sequences via Proxy-Guided LLM Evolution
Jun 8, 2026Shibing Mo, Jing Liu, Jianchu Xu +1Automated Algorithm Discovery Macro Placement
the School of Artifcial Intelligence, Xidian University · the Guangzhou Institute of Technology, Xidian University.
FlowPlace: Flow Matching for Chip Placement
Apr 26, 2026Peng Xie, Ke Xue, Yunqi Shi +6Constrained Generative Modeling Flow Matching
1State Key Laboratory of Novel Software Technology, Nanjing University, China · School of Artificial Intelligence, Nanjing University, China · 3Huawei Noah’s Ark Lab, China

Sep 18, 2026J/K move · Enter open · S save

PlaceReasoner-Beta: Reasoning-Driven Macro Placement and Benchmarking

Qiufeng Li, Chengxuan Wang, Rongqian Chen, Quan Cheng+5

The George Washington University, USA · University of California, Los Angeles, USA · Brown University, USA+3

Macro Placement Electronic Design Automation

Open paper AskPDF arXiv

Automated macro placement remains a fundamental challenge in VLSI physical design. Despite decades of research, existing approaches predominantly optimize hand-crafted proxy objectives, such as estimated wirelength, and typically produce placements through one-shot numerical optimization, limiting their ability to incorporate visual layout context, codified design expertise, and downstream physical-design feedback in a unified loop. We present PlaceReasoner-Beta, a verifier-guided multi-agent framework that reformulates macro placement as a closed-loop reasoning problem rather than black-box optimization. A vision-language model (VLM) planner generates candidate placements from the floorplan image, macro specifications, and connectivity structure; a geometric verifier enforces physical legality and expert placement principles; a physical verifier refines candidates using early implementation feedback; and a post-route optimizer further improves promising layouts using final PPA. To enable reproducible evaluation, we introduce PlaceReasoner-Bench, a fully open end-to-end benchmark built from open RTL designs, EDA tools, and technology libraries. It comprises 8 designs at two aspect ratios, yielding 16 tasks with fixed floorplans and I/O assignments, so methods differ only in macro positions and orientations and are evaluated using routed PPA and DRC rather than pre-route proxies. Across the benchmark, PlaceReasoner-Beta achieves the best timing among DRC-clean methods on all square tasks, reducing post-route TNS by 61.2% at 1:1 and 53.0% at 2:1 relative to the classical baseline field. It also shortens routed wirelength on most designs despite never explicitly optimizing it, demonstrating that reasoning over spatial structure under physical-design feedback can improve end-to-end layout quality beyond proxy-objective optimization.

Abstract

Macro Placement

Explore similar work

PlaceReasoner-Beta: Reasoning-Driven Macro Placement and Benchmarking

Order Matters: Unveiling the Hidden Impact of Macro Placement Sequences via Proxy-Guided LLM Evolution

FlowPlace: Flow Matching for Chip Placement