
## date (UTC)
2026-09-29T22:07:24Z

## strata version
v0.1.26

## os
PRETTY_NAME="Ubuntu 24.04.5 LTS"
VERSION_ID="24.04"
6.8.0-139-generic

## virtualization
kvm

## cpu
CPU(s):                                  20
Model name:                              AMD EPYC 7453 28-Core Processor
Thread(s) per core:                      1
Core(s) per socket:                      20
Socket(s):                               1
L3 cache:                                320 MiB (20 instances)
NUMA node(s):                            1

## cpu flags (relevant)
avx2 f16c fma

## numa
available: 1 nodes (0)
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
node 0 size: 168968 MB
node 0 free: 47460 MB
node distances:
node   0
  0:  10

## memory
               total        used        free      shared  buff/cache   available
Mem:             165          14          46           0         105         150
Swap:              0           0           0

## dimms
run as root for dmidecode
(in a VM these are virtual DIMMs: fill in the host section below by hand)

## memlock limit
21627928

## gpu
index, name, driver_version, memory.total [MiB], pcie.link.gen.current, pcie.link.gen.max, pcie.link.width.current, pcie.link.width.max, power.limit [W], power.default_limit [W], power.max_limit [W], power.draw [W], clocks.max.sm [MHz]
0, NVIDIA GeForce RTX 3090, 595.91.07, 24576 MiB, 1, 4, 16, 16, 280.00 W, 350.00 W, 366.00 W, 35.23 W, 2130 MHz
1, NVIDIA GeForce RTX 3090, 595.91.07, 24576 MiB, 1, 4, 16, 16, 280.00 W, 420.00 W, 450.00 W, 28.60 W, 2100 MHz

## gpu topology
	[4mGPU0	GPU1	CPU Affinity	NUMA Affinity	GPU NUMA ID[0m
GPU0	 X 	PHB	0-19	0		N/A
GPU1	PHB	 X 	0-19	0		N/A

Legend:

  X    = Self
  SYS  = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
  NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
  PHB  = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
  PXB  = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
  PIX  = Connection traversing at most a single PCIe bridge
  NV#  = Connection traversing a bonded set of # NVLinks

## gpu processes (must be empty before benchmarking)
pid, process_name, used_gpu_memory [MiB]

## docker
Docker version 29.8.1, build 4a63305
NVIDIA Container Toolkit CLI version 1.18.1

## disk (model data)
Filesystem      Size  Used Avail Use% Mounted on
/dev/sdb1       541G  409G  106G  80% <storage>

## host (fill in manually)
- Hypervisor: Proxmox VE (KVM guest; host version not visible inside the VM)
- CPU: AMD EPYC 7453 28-Core Processor (Milan); 20 vCPU exposed, AVX2, no AVX-512 visible
- RAM: 165 GiB visible to the guest; physical DIMM layout and bandwidth not measured
- VM: 20 vCPU, 1 socket x 20 cores x 1 thread, 1 NUMA node; ballooning and pinning not verified
- GPUs: 2x RTX 3090 24 GB; PCIe Gen4 x16 maximum reported for each (current idle link generation varies); configured power limit 280 W each
- Storage: ext4 guest data disk, 541 GiB; host storage layout not verified from the guest
