Bridges, funeral access and population in waves

On 3 October the work centred less on new city area than on whether the existing Northstar functions. The building agent replaced level crossings with bridges, moved funeral services closer to problem residential areas and corrected buildings with poor access. New households kept arriving in jumps. At midday the operator stopped the construction work while the simulation kept running; that is a pause in the building agent's work, not a demonstrated drop in computing effort.

The day's population curve

The lowest recorded level of the day, about 1.58 million inhabitants, came in the early hours. The last qualified reading of the evening was also the maximum, at around 1.84 million. A dip and later growth belong to the same day. In the qualified readings the population had not yet reached two million. The midday hand-over figure is an interim state and not a day-end value.

The readings belong to the Berlin reporting day, from 22:00 UTC on 2 October to 22:00 UTC on 3 October.

Funeral service: capacity is not enough

In the early morning, nearly all hearses in one examined group were on the road, yet almost two thirds of those observed on the road were standing. The only access to newly built crematoria was heavily congested. The building agent added a second road connection and later placed facilities nearer the affected housing.

In the following samples the share of standing hearses fell step by step. Hearse reports fell from roughly 550 in the early morning to a low later in the morning. These are not a city-wide, continuous vehicle measurement. New facilities, roads, capacity extensions and changing move-in overlapped, and the share of each measure in the decline is not isolated.

After another wave of move-in the reports rose again. Full cemeteries were replaced, hearse garages added and three more crematoria with furnace and cold room built in the north-west. The later hand-over recorded about 30 reports. The bottleneck was reduced; it is not declared permanently removed.

An access check also found five previously placed hearse garages without a working connection. They were removed, and only at one of the sites did a connected replacement succeed. The nominal fleet size would otherwise have counted these vehicles. The placement routine now checks access after building.

Four crossings, four bridges, one open road bottleneck

An analysis of the stored network snapshot found four level road-rail crossings. Bridges were built at these places in the west. The logged attempts show the limits of the planning: the first height was not enough for the road bridge's clearance and pier layout, and a different design was accepted. For the rail bridges, ramps and small lateral offsets had to be shaped so that the new sections stood on their own.

Follow-up checks in two affected areas showed rising internal traffic-flow values. This is a local finding in the running game, not a city-wide throughput gain and not a controlled CPU comparison.

In the saturated street grids the result differed. Removing further traffic lights brought no visible gain on two streets; flow values stayed near 0.08. The building agent corrected its earlier, more favourable assessment. The remaining bottleneck was identified as the outflow chain to the motorway junction. A simple lane upgrade failed on an overlap. A larger rebuild is noted but was not carried out that day.

Utilities, parks and move-in

A burnt-out substation in the south was replaced on the same site. Later checks showed power reports falling from 13 to zero, and a fire station was added nearby. The zero applies to that check; it does not show a lasting, fault-free power supply for the whole city. A further recycling centre with power, water and sewage connections was built in the west, and an industrial business had to make way for its access road. The intended shortening of waste routes remains an assumption until the trips are re-measured comparably.

The night's wave of parks needed rework too. Most of the newly placed small parks first reported missing pedestrian access. A corrected orientation lowered that number in the first follow-up check; more parks were improved later. The build preview alone had not caught the error. The recorded upkeep factor improved, but that gives no isolated effect of the whole park measure on satisfaction.

Empty commercial and industrial areas were rezoned to high-density offices in several batches, in the north-west and the south-east. Zoned cells are neither finished offices nor actual jobs, which is why the daily supplement shows measured jobs separately. A further batch of unbuilt high-density residential cells was switched to cheaper housing; first buildings appeared, and a lasting growth effect is open.

The residential tax was adjusted according to the observed relation of move-ins, waiting households and housing demand. During standstill it stood at twelve percent for a time and was lowered step by step during move-in phases. The measure is meant to slow the supply of waiting households; it does not remove the existing stock directly. At 12:28 UTC a standstill that had lasted about 25 minutes ended without a process restart, so a restart is not a necessary condition for every new move-in surge.

More offices can worsen another bottleneck. The goods analysis assigned large electronics imports to office industries. According to the network check, the eastern freight yard lies in its own rail network, and linking it to the main network would be a larger project. An additional terminal in the west was set aside after a fruitless search for a site.

Computing load

The hand-over at 12:40 UTC recorded moving averages of game CPU and system-wide CPU for the preceding six hours. These are logged by the building agent, not a daily mean we recomputed from the performance series. The CPU time per unit of simulation work recorded there kept fluctuating, and no reliable efficiency gain from the day's city measures was shown. Large rezoning waves, move-in, traffic and construction change the load at once.

The separate performance series holds raw readings with documented gaps, including the day's edges. Some readings are marked paused. The counts of process and world identifiers it contains are not a reliable number of restarts. CPU and frame differences across such boundaries must not count as normal performance windows.

Status and open points

The operator stopped construction at about 12:37 UTC; the game stayed active at speed 1, so later population development still belongs to the reporting day. An interruption of the agent's work is neither a game pause nor an optimisation success. After 19:52 UTC no further qualified population value exists for the rest of the day, and none can be derived from the performance series. The larger rebuild at the motorway junction and the link for the eastern freight yard remain open. Project overview: CS2.