Summary
This knowledge base summarizes the results of performance tests on four TRI050S color cameras.
In the test, the four cameras were connected to a nominal-specification PC via a 4-port network interface card (NIC). The primary objective was to assess the cameras’ ability to maintain stable video streams to console while the host system was under significant computational load. The CPU usage was monitored during the test, and the system was stressed using IntelBurn. Even under the second-highest stress setting in IntelBurn, all four camera streams remained stable in both tests with no frame drops observed. When idle (no stress testing), CPU utilization remained at about 3%.
This testing scenario simulates use cases where cameras are expected to perform consistently in environments with limited computing resources or during periods of high system demand.
Setup
Cameras tested
- Models: 4 × TRI050S-CC
- Serial Numbers: 240604017, 240302125, 234600016, 234600018
- Firmware Version: v1.127.0.0
Arena SDK version
- Arena SDK 1.0.49.3
Host system
The nominal-spec PC was used as the host machine for these tests.
| CPU | Intel® Core™ i7-700 CPU @ 3.60GHz | |
| GPU | NVIDIA GeForce GTX 1050 Ti | |
| Memory | Total RAM | 18 GB |
| Committed | 5.1/18.2 GB | |
| Cached | 8.8 GB | |
| Speed | 2400 MHz | |
| Slots used | 3 of 4 | |
| Form Factor | DIMM | |
| Hardware reserved | 142 MB | |
| Operating System | Windows 10 Pro | |
| Disk Space | Local Disk (C drive): 110 GB | |
| System Firmware | Driver Version 10.0.19041.1 | |
| Network Interface Card | 4-port Realtek PCLe 2.5Gbe Family Controller | |
| Network Adapter Settings | MTU | 9014 Bytes |
| Receive Buffers | 512 | |
| Transmit Buffers | 4096 | |
| Speed Duplex | Auto Negotiate | |
| Interrupt Moderation | Enabled/Adaptive | |
| IPv4 Checksum Offload | Rx & Tx Enabled | |
| Energy-Efficient Ethernet | Disabled |
Node values
| Node | Value |
|---|---|
| AcquisitionMode | Continuous |
| ExposureAuto | Off |
| ExposureTime | 500.0 |
| StreamBufferHandlingMode | bufferHandlingMode |
| TriggerMode | Off |
| PacketSize | 9000 |
| GevSCPD | 80 |
| DeviceLinkThroughputReserve | 10 |
| StreamPacketResendEnable | True |
| GevGVSPExtendedIDMode | On |
Other test details
| Test feature | Details |
|---|---|
| Power source | PoE |
| Cables | M12 to RJ45 IP67 Cat6a Cable – 2m, Amphenol |
| Monitoring tools | Basic system performance monitor (CPU usage only) |
| CPUID DWMonitor | |
| Streaming target | Console |
| Stress tool | IntelBurn |
- IntelBurn details: Tests were conducted up to the second-highest stress level. The highest stress level was not executed due to computer hardware constraints unrelated to camera performance.
- Test conditions: Only the CPU performance monitor and camera streaming processes were running during testing to ensure results reflected conditions without background interference.
Figure 1: Four Triton cameras connected to a Realtek controller.
Figure 2: Hardware monitor for Triton test: Voltage, temperatures, and fan speeds before the stress test.
Figure 3. Hardware monitor for Triton test: Voltage, current, power, temperatures, fan speeds, and clock speeds before the stress test.
Results
The test results are shown below:
- CPU usage increased significantly during each level of the IntelBurn stress test, as expected. The test successfully loaded the CPU without errors or instability.
- No frame drops were observed in any of the four TRI050S0 camera streams during any phase of the test.
- The system remained responsive enough to maintain continuous console logging, even as it approached 100% CPU load thresholds.
Figure 4: Very-high-stress test with 4 Triton streams active; 100% CPU utilization; no frames dropped.
Figure 5. Completed IntelBurn test on host computer for Triton test with stress level of “Very High”.
Discussion
This test assessed whether four TRI050S cameras could maintain stable console streams on a low-spec PC under significant CPU stress. The results exceeded expectations. In each test, all four camera streams remained stable with no observable frame drops even at the second-highest IntelBurn stress level. Based on the results, it’s likely the system used was more powerful than necessary, and an even lower-spec PC could potentially deliver the same results.
Conclusion
Testing showed that four TRI050S cameras can each stream reliably with no frame drops on a low-spec PC, even under high CPU stress. These results exceeded expectations.