- StorageReview’s physical server computed 314 trillion digits without a distributed cloud infrastructure
- The entire calculation ran continuously for 110 days without interruption
- Power consumption has dropped dramatically compared to previous cluster-based Pi records
A new benchmark for large-scale numerical computation has been set with the calculation of 314 trillion digits of pi on a single local system.
The run was completed by Storage assessmentovertaking previous cloud-based efforts, including Google Cloud’s 100 trillion-digit calculation from 2022.
Unlike large-scale approaches that relied on massive distributed resources, this record was achieved on a single physical server using tightly controlled hardware and software choices.
Runtime and system stability
The calculation was performed continuously for 110 days, which is considerably shorter than the roughly 225 days required for the previous large-scale record, even though that earlier effort produced fewer numbers.
The uninterrupted performance was attributed to operating system stability and limited background activity
It also depends on a balanced NUMA topology and careful tuning of memory and storage, designed to match the behavior of the y-cruncher application.
The workload was treated less as a demonstration and more as a long-term stress test of production systems.
At the heart of the effort was a Dell PowerEdge R7725 system equipped with two AMD EPYC 9965 processors, providing 384 CPU cores, in addition to 1.5 TB of DDR5 memory.
The storage consisted of forty 61.44 TB Micron 6550 Ion NVMe drives, delivering approximately 2.1 PB of gross capacity.
Thirty-four of those drives were allocated to y-cruncher scratch space in a JBOD format, while the remaining drives formed a software RAID volume to protect the final output.
This configuration prioritized throughput and energy efficiency over full data tolerance during computation.
The numerical workload generated significant disk activity, including approximately 132 PB of logical reads and 112 PB of logical writes during the run.
The peak usage of the logical drives reached approximately 1.43 PiB, while the largest checkpoint exceeded 774 TiB.
SSD wear stats reported about 7.3 PB written per disk, which equates to about 249 PB on the swap devices.
Internal benchmarks showed that sequential read and write performance more than doubled compared to the previous 202 trillion digit platform.
For this setup, an energy consumption of approximately 1,600 watts was reported, with a total energy consumption of approximately 4,305 kWh, or 13.70 kWh per trillion calculated figures.
This figure is much lower than estimates for the previous cluster-based record of 300 trillion figures, which reportedly consumed more than 33,000 kWh.
The result suggests that certain workloads require carefully tuned servers and workstations can outperform cloud infrastructure in terms of efficiency.
However, that assessment applies strictly to this class of calculations and does not automatically extend to all scientific or commercial use cases.
Follow TechRadar on Google News And add us as a preferred source to get our expert news, reviews and opinions in your feeds. Then be sure to click the Follow button!
And of course that is also possible Follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us WhatsApp at.
#news #sites #metal #server #rewrote #rules #extreme #numerical #computing