sprintd Basel GreenTech Edition · Pitch 2 Room 2.22 Team Invincibles · Interactive 3D Atlas
CalorGrid logo

CalorGrid

Useful Load => Useful Heat · Basel 2025 Thermal-Compute 3D Atlas

Can 4 x 1 kW liquid-cooled servers heat a Basel home while doing real Digital Humanities compute?

Isaac Sim 6.2 RTX

Basel Klingental 2025
swisstopo + SpeedTree · 31.2 MWh/yr demand

Native SimReady USD

Liquid-Cooled Plant Room
4 x 1 kW rack · Liebert CDU · 35 kWh buffer

Summer Verbund RTX

Gartenbad St. Jakob
Summer sink archetype · 96.2% utilisation

Vaibhav Motwani · Prakash Aryan · Vilas Harapanahalli

Live looping videos in HTML · press Space / Arrow keys · press N for speaker notes

1 / 8

Open with CalorGrid: three real ray-traced scenes of our Basel reference neighborhood, liquid-cooled plant room and summer Verbund pool sink. Every number in this talk comes from one verified 8,760-hour Python conservation engine.

OverviewLive DemoSummer VerbundPilot RealityEvidenceVerdict

Basel 2025 Reference House: Physics & 7 KPIs

4 x 1 kW servers · 35 kWh lossless tank · IWB H4 0.3325 CHF/kWh excl. VAT · scored DH-Lab trace (12 srv-h/srv/day)

Basel house liquid cooling cutaway diagram

80% recovered to hydronic loop (3.2 kW_th) · 20% to 1-node lumped basement air · no heat pump

1. Heat Demand Covered 34.1% 10,643 / 31,231 kWh (baseline: 32.7%)
2. Annual Heat Dumped 0.0 kWh Baseline dumps 3,790.1 kWh (27.0% wasted)
3. Server Utilisation (N x 8,760 h) 38.0% 13,304 srv-h sold (baseline: 50.0%)
4. Missed Deadlines (Unserved) 26.3% 9,689 / 13,140 jobs (N=1 house: 0.0%)
5. Electricity vs Compute Income CHF 4,424 / 3,326 At 0.25 CHF/srv-h · heat credit CHF 1,416
6. OPEX Break-Even Floor 0.2261 CHF/srv-h · baseline 0.2549 · res. 3.97e-10
Economics: Smart cuts 3,790 kWh dump & saves CHF 1,401 in power (beats baseline below ~0.346 CHF/srv-h incl. heat, before SLA loss). Add CAPEX (1,400 CHF/srv-yr) and this house needs ~0.65 CHF/srv-h (0.39-0.43 with a summer sink).

Sources: Open-Meteo Basel 2025 UTC->Europe/Zurich (8,760 h, 206 heating days, -5.6 C min, 31,231 kWh); ElCom 2026 IWB H4 (0.3325 CHF/kWh excl. VAT; displaced heat value 0.133 CHF/kWh_th).

2 / 8

All 7 required reference-house numbers upfront: 4x1 kW servers, 34.1% heat covered, 0.0 kWh dumped vs 3,790 kWh baseline, 38.0% util, 26.3% missed deadlines, CHF 4,424 elec vs CHF 3,326 income and 0.2261 CHF/srv-h OPEX break-even.

OverviewLive DemoSummer VerbundPilot RealityEvidenceVerdict

Live 3D Atlas: Winter, Summer & OS Pause

The exact same Python function (decide_live_step()) drives the 8,760 h balance and real CPU Digital Humanities jobs

1. Jan 15 Peak (-4.5 C)
212.6 kWh peak demand
4 servers recover 38.4 kWh with 0.0 kWh dumped in X-Ray basement view.
2. Jul 15 Dump (12 kWh DHW)
26.4 kWh wasted in 1 day
Baseline fires red bypass; Smart runs 15.0 srv-h with 0.0 kWh dumped.
3. Live CPU OCR Worker
Real OS SIGSTOP / SIGCONT
Full tank pauses the tesseract PID; heat demand resumes the same PID.
35 kWh Buffer Tanklive SoC + hourly demand
decide_live_step()deterministic Python policy
SIGSTOP / SIGCONTPOSIX process group signal
Live DH-Lab WorkerOCR · IIIF FFT · Audio MFCC

80% recovered server heat feeds back into the 35 kWh buffer tank (closing the hourly conservation loop)

Verified day runs: Jan 15 (212.59 kWh demand, 38.40 kWh used, 0.0 kWh dumped); Jul 15 Smart (15.0 kWh elec, 12.0 kWh used, 0.0 kWh dumped) vs Baseline (48.0 kWh elec, 26.4 kWh dumped).

3 / 8

All three videos loop live right inside the slide: winter peak cutaway, summer dump bypass and our live local CPU tesseract process pausing on SIGSTOP and resuming on SIGCONT.

OverviewLive DemoSummer VerbundPilot RealityEvidenceVerdict

Why Single Houses Hit a Summer CAPEX Cliff

Smart scheduling stops summer heat dumping, but 3,816 warm hours leave expensive server hardware idle

Winter play vs summer SIGSTOP pause whiteboard sketch
Winter (206 heating days)
Cold house + low tank -> SIGCONT runs batch compute at full load.
Summer (3,816 warm h)
Full 35 kWh tank -> SIGSTOP pauses servers; idle time raises CAPEX/h.

Continuous Annual Heat Dumped (N=4)

Basel Single House
9,910 kWh
Dual-Sink Erdsonde
1,052 (-89.4%)
Sole Uno / ARA 24/7
0 kWh (-100%)

Heat-Led Server Utilisation (Unlimited Queue)

Standalone Pool (May-Sep)
42.2%
Basel Single House
64.6%
Erdsonde / Paired Pool
96.2%
Year-Round Thermal Host
100.0%

Comparator note: Unlimited-queue runs isolate each thermal sink's absorption ceiling (64.6% single-house util vs 38.0% on the arrival-limited Slide-2 DH-Lab trace).

4 / 8

This visual explains the central bottleneck: pausing servers solves dumping, not hardware amortisation. Adding a summer sink lifts utilisation from 64.6% to 96.2%-100%.

OverviewLive DemoSummer VerbundPilot RealityEvidenceVerdict

Summer Verbund: Jobs Move, Heat Stays Local

3-way diverter sends July surplus to borehole recharge or Swiss warm-water sites (central CAPEX 1,400 CHF/srv-yr)

Summer Verbund 3-way valve to Erdsonde borehole and pool
Archetype & Valuation Catch: St. Jakob, Sole Uno Rheinfelden and ARA are thermal archetypes (no partnership). 0.392-0.43 CHF/srv-h spans full vs zero secondary-sink heat credit.
St. Jakob Pool · 32/26 C
Davos Alpine · 52.1 MWh
Archetype (N=4, unlimited queue)LoopW45UtilCHF/srv-h
Basel Single House (comparator)55/40 CNo64.6%0.473
Dual-Sink Erdsonde (SIA 384/6)40/30 CSec96.2%0.392-0.43*
Paired Freibad (St. Jakob)32/26 CYes96.2%0.392-0.43*
Sole Uno Bath / ARA Biogas (24/7)42-45 CYes100.0%0.386
Davos Alpine House (291 heat days)55/40 CNo84.1%0.416

All 7 Verbund scenes close hourly conservation (residual <= 7.3e-9 kWh). Laundry (60/45 C) and brewery CIP (62/47 C) are preheat-only outside W45; grid feed-in stays out of scope.

5 / 8

Explain why pairing one house with an underground Erdsonde borehole or placing servers at year-round 35-45 C warm-water sites lifts utilisation to 96%-100% and brings total break-even down to 0.386-0.43 CHF/srv-h.

OverviewLive DemoSummer VerbundPilot RealityEvidenceVerdict

Pilot Reality: Where Assumptions Break

Five measured limits kept outside the lossless benchmark

1. HEATWISE Idle-Power Reversal (21.0-26.8% idle)

Ideal Smart (0% idle)
0.226 CHF/h
Baseline (21-27% idle)
0.325-0.344
Smart SIGSTOP (paused)
0.340-0.372!

Long pauses need deep sleep or power-off, not OS SIGSTOP.

2. ElCom 2026 Retail H4 vs Industrial C6 Tariff

Industrial C6 (excl. VAT)
0.2527 CHF/kWh
Retail H4 (excl. VAT)
0.3325 (+0.0798)

H4 gap erases 75% of the 0.1064 heat credit; COP 3.5 lifts OPEX floor to 0.2565.

Climate prerequisite: Useful compute and displaced heating are prerequisites for our claimed climate benefit.
Live Pilot Reality Drawer
CAPEX Cliff
N=12 drops util to 37.2% (0.66 CHF/h)
ASHRAE W45
Mixed tank halves to 17.5 kWh (<=45 C)
MLPerf v5.1
8x B200: 0.0063 vs 0.0093 CHF/1M tok

HEATWISE (Zenodo 23237930): hybrid liquid 54.9 C vs 91.2 C air (9.4% slowdown); 80% capture at 50-60 C needs Aquasar full-board cooling. MLPerf v5.1: derived energy-only bounds (not verified by MLCommons).

6 / 8

Covers five hard engineering limits: paused idle power reversal, retail vs industrial tariffs, N=12 CAPEX cliff, ASHRAE W45 temperature limits and MLPerf v5.1 energy-only marginal cost.

OverviewLive DemoSummer VerbundPilot RealityEvidenceVerdict

Provenance Mode: Data & Model Sources

One FastAPI contract · zero browser math · 4-color 3D overlay labelling each mesh by source

4-Category 3D Provenance Overlay
Measured vs. Modelled: Open-Meteo is reanalysis; live worker reads OS PID & CPU time (estimates workload joules, not wall-plug); IFC 4.3 house & SimReady racks are visual stand-ins.
1. Live Observed OS process ID, CPU %, estimated workload joules and SIGSTOP/SIGCONT transitions.
2. Benchmark SHA-256 pinned Open-Meteo 2025, ElCom 2026 H4 tariff and 35 kWh lossless tank.
3. Geodata / CAD 2048px swisstopo, OpenStreetMap, SpeedTree vegetation, IFC 4.3 shell & SimReady hardware.
4. Sensitivity HEATWISE, MLPerf v5.1, ASHRAE W45, CAPEX cliff and ElCom C2-C6 tariffs.
77/77 Verified Unit Tests 125/125 Chrome WebGL E2E 8.5 MB Fresh CPU Clone <= 7.3e-9 kWh Residual

Attributions: (c) swisstopo; (c) OpenStreetMap contributors, ODbL; (C) buildingSMART International Ltd. (CC BY 4.0); NVIDIA, Vertiv and ForgeCAD SimReady (CC BY 4.0); Three.js (MIT); Tesseract (Apache-2.0).

7 / 8

Provenance Mode answers visual skepticism in one click: green shows live OS worker telemetry, amber marks benchmark inputs, blue marks Swiss geodata and open CAD stand-ins inside an 8.5 MB CPU clone.

OverviewLive DemoSummer VerbundPilot RealityEvidenceVerdict

Pilot Verdict: Where to Test First

Basel pilot targets from 8,760 h physics, measured idle power and ElCom tariffs

Skip: Oversized Winter Homes
N=12 drops util to 37.2% and pushes CAPEX+OPEX to 0.66+ CHF/srv-h.
Skip: Retail H4 Without C6 Rate
0.0798 CHF/kWh gap erases 75% of residential heat credit.
Skip: SIGSTOP Without Sleep
HEATWISE 21-27% idle draw reverses Smart's OPEX edge.
Pilot decision crossroads sketch
1. Year-Round 35-45 C Hosts
Thermal baths, biogas digesters & laundry preheat: 100% util · 0.386 CHF/h.
2. Dual-Sink Erdsonde Homes
SIA 384/6 borehole recharge: 96.2% util · -89.4% dump · 0.392-0.43 CHF/h.
3. Pre-Hardware Checklist
Verify commercial tariff, usable loop temp & contracted interruptible batch demand.
Reproduce (CPU Linux): uv sync && uv run digitalheat verify && uv run digitalheat serve
77/77 Unit · 125/125 E2E · 8.5 MB Clone

Literature: ETH Zurich Aquasar (Zimmermann et al., Energy 2012); Data Furnace (Liu et al., USENIX HotCloud 2011). Datasets: Open-Meteo, ElCom/LINDAS 2026, HEATWISE, MLPerf v5.1.

8 / 8

Close on two first pilot targets: year-round 35-45 C warm-water hosts and same-server Erdsonde borehole homes, gated by the 3-point pre-hardware checklist.