Test Build: Supermicro X13SAE-F Intel Core i9-14900KF 6.0 GHz

In this article we will share the results of testing a computer assembly based on the Supermicro X13SAE-F motherboard and the new Intel Core i9-14900KF processor overclockable up to 6.0 GHz.

HOSTKEYis a hosting company, and as such it is important for us to update our equipment in a timely manner in order to provide quality service and reduce costs. New equipment always involves a testing phase, as this is a good way to build up our own expertise on new platforms, to study the specifics of their operation, customization and configuration for client tasks.

The latest novelty in our company was a build with an SM X13SAE-F motherboard and a Core i9-14900KF 6.0GHz processor. Let's take a look at what exactly this hardware is, and then we can consider the results of our testing of this configuration, especially given the lack of official results from some test programs.

SM X13SAE-F and Intel Core i9-14900KF specifications

SM X13SAE-F motherboard

SM X13SAE-F is suitable for use in high-performance servers. Firstly, it supports the latest powerful 14th generation Intel processors with a large number of cores and high clock speeds.

Secondly, the board has a high memory bandwidth of up to 192Gb using DDR5 modules. Fast memory access is critical for server workloads.

SM X13SAE-F is equipped with a fast storage subsystem: 8 SATA3 ports and 3 M.2 PCIe 4.0 x4 slots with RAID support. High-speed PCIe 5.0 and PCIe 4.0 interfaces allow you to connect high-performance additional components.

The board also supports virtualization and remote management technologies, which is important for servers. Monitoring functions track the status of all components in real time.

An additional advantage when working with the SM X13SAE-F motherboard is the ability to directly control the hardware via an HTML5 console:

Intel Core i9-14900KF 6.0GHz processor

This processor is well suited for use in server systems. Its hybrid architecture with 24 cores (8 high-performance and 16 energy-efficient cores) provides an excellent balance of performance and power consumption. The presence of cutting-edge Intel technologies is also an important advantage.

A sufficient number of cores and high clock frequencies (up to 6 GHz) allow it to efficiently cope with multi-threaded loads typical for web servers and other server segment tasks. Support for 32 threads and 36 MB of cache also scales well for server tasks.

The large amount of supported RAM (up to 192 GB), high bandwidth (up to 89.6 Gbps) and modern DDR5 memory interface are also important advantages for creating a productive server.

Testing

Description of the test environment

We prepared two identical builds for the tests. By running the same tests on each, we can track anomalies and get relevant results.

Test

nench.sh

curl —s wget.racing/nench.sh | bash; curl —s wget.racing/nench.sh | bash) 2>&1 | tee nench.log

Server 1

CPU (less is better)

Ioping (less is better)

dd (less is better, Mbps)

IPv4 speed tests (less is better, Mbps)

Server 2

CPU (less is better)

Ioping (less is better)

dd (less is better, Mbps)

IPv4 speed tests (less is better, Mbps)

Iperf3 test results

An iperf3 test was run between a local client and a remote server. The test included 10 parallel threads and was executed for 10 seconds on each thread.

The first test

The second test

The third test

The fourth test

In the first test, the total amount of data transferred was 2.03 GB with an average throughput of 1.75 Gbps. The highest throughput of 435 Mbps was observed on stream 7. There were also over 100 retransmissions on multiple streams indicating packet loss.

In the second reverse mode test, the total amount of data received was 1.23 GB and the total throughput was 1.05 Gbps. The average throughput varied between streams from 51.5 to 174 Mbps. Maximum retransmissions (5) on streams 5 and 19.

In the third test, 10.7 GB of data were transferred at an average rate of 9.22 Gbps. The throughput of the streams ranged from 855 to 978 Mbps with few retransmissions.

The fourth test transferred 11 GB at an average rate of 9.45 Gbps. A single stream achieved a record throughput of 2.99 Gbps. The number of retransmissions was up to 22 on individual streams.

librespeed

./librespeed-cli --server-json https://speedtest.hostkey.com/servers.json --server 1	

Using the librespeed-cli utility, we tested our internet speed on the speedtest.imd.com server located in the Netherlands.

The download and upload speeds were measured. The download speed was 306.6 Mbps - a pretty decent result, indicating fast and stable downloading of content. The upload speed was 309.23 Mbps - also above average, which is important for transferring large amounts of data. Another plus is that the download and upload speeds are about the same.

In general, the test showed excellent results - fast and stable connection without obvious bottlenecks when downloading and uploading data

Stress- ng

stress-ng --cpu 24 --io 4 --vm 1 --vm-bytes 16G --timeout 60s --metrics-brief

stress-ng --cpu 32 --io 4 --vm 1 --vm-bytes 32G --timeout 60s --metrics-brief

stress-ng --cpu 32 --io 8 --vm 4 --vm-bytes 128G --timeout 60s --metrics-brief

According to the results of the CPU performance test (bogo ops), the best result was shown by AMD Ryzen 9 7950X processor - 3240575 bogo ops. The second place went to the Intel Core i9-14900KF at 2074846 bogo ops. AMD Ryzen 9 5950X showed a result of 2509636 bogo ops.

In the I/O (io) performance test, the AMD Ryzen 9 7950X is the leader with a result of 595084 bogo ops. The second place is held by the Intel Core i9-14900KF with 266706 bogo ops. The lowest result went to the AMD Ryzen 9 5950X at 14076 bogo ops.

In the virtual memory (vm) performance test, the best result was also the AMD Ryzen 9 7950X at 337066 bogo ops. Intel Core i9-14900 KF was in second place with 3856467 bogo ops. The AMD Ryzen 9 5950X took the third place with 5748301 bogo ops.

Thus, according to the sum of the scores of all the tests, the most productive processor is the AMD Ryzen 9 7950X. In second place is the Intel Core i9-14900KF. The AMD Ryzen 9 5950X showed the lowest performance among the tested processors.

Based on the test data, we can draw some conclusions about the preferred applications of these processors.

Firstly, the AMD Ryzen 9 7950X performed better in CPU and I/O performance tests. It is best suited for applications that require high processing power and fast data access, such as scientific computing, machine learning, and video rendering.

Secondly, although the Intel Core i9-14900KF is inferior in overall performance, it scored well in the virtual memory performance test. It is well suited for memory-intensive tasks such as compiling code or working with databases.

Thirdly, the AMD Ryzen 9 5950X lags behind its competitors in almost all indicators. Nevertheless, its performance is high enough for most tasks. In terms of price to features ratio, it can be a good solution for gaming and everyday tasks.

Drives

stress-ng --sequential 0 --class io --timeout 60s --metrics-brief

In the aio (asynchronous I/O) performance test, the best result was shown by the AMD Ryzen 9 7950X at 4216227 bogo ops. Second place went to the Intel Core i9-14900KF with a result of 3305101 bogo ops. The AMD Ryzen 9 5950X holds the third place at 1888993 bogo ops.

In the aiol (linear asynchronous I/O) test, the Intel Core i9-14900KF leads with a score of 614934 bogo ops. The AMD Ryzen 9 7950X took the second place with 614553 bogo ops. The AMD Ryzen 9 5950X came in third place with 744673 bogo ops.

On the hard drive (hdd) performance test, the Intel Core i9-14900KF posted the highest score with 3979264 bogo ops. The AMD Ryzen 9 7950X trails behind at 4275200 bogo ops. The AMD Ryzen 9 5950X took the lowest result at 985088 bogo ops.

The AMD Ryzen 9 7950X leads on the aggregate of all tests, with the Intel Core i9-14900KF in second place and the AMD Ryzen 9 5950X showed the lowest performance.

Thus, the AMD Ryzen 9 7950X is best suited for compute-intensive tasks and asynchronous I/O. The Intel Core i9-14900KF is good for linear I/O and storage work. The AMD Ryzen 9 5950X is inferior in performance but can be a viable alternative with a better price to feature ratio.

Supermicro Super Server

In the single core performance test, a result of 1608 points was obtained. This indicates a fairly high speed of execution of sequential tasks on a single core.

In the multithreaded test, which involved all processor cores, the server showed a result of 14 554 points. This indicates a good performance for the server in parallel data processing on all the cores.

Thus, the tests showed that the server has high computing power for both sequential and parallel tasks. It is suitable for a wide range of computational work requiring fast performance.

CPU Z

CPUID HWMonitor. Temperatures (C)

The results of the CPU performance tests showed the following:

• In the single-thread test, the CPU scored 980.6 points, which indicates a high speed of processing sequential tasks.
• In the multithreaded test, the result was 15587.3 points, and the ratio to the single thread test was 15.9. This indicates the efficiency of the processor in parallelizing calculations.

If we look at the official corporate results, our findings were quite close to the optimal ones obtained by the authors of CPU-Z (the i9-13900KF processor is weaker than the 14900KF by 5-10 per cent), but close to the real ones. In turn, the "test" advantage of the 14900KF over the Ryzen 9 7950X in real life is insignificant.

Temperature monitoring showed that in the Supermicro X13SAE-F system the temperatures are normal and do not exceed the maximum values. For the Intel Core i9 14900KF processor, the maximum package temperature reached 55 degrees Celsius. Temperatures of individual cores also did not exceed the norm.

Thus, the tests confirm the high performance and stability of the system under various loads. Critical overheating of components was not observed; at least the new (old) Intel processor is no worse than the 13th generation.

Conclusion

The tests of the new server assembly based on the Supermicro X13SAE-F motherboard and the Intel Core i9-14900KF processor conducted within the framework of this research allow us to draw the following conclusions:

1. The system configuration is quite successful and corresponds to the tasks of building a high-performance server solution. The motherboard provides support for the most modern components, a large amount of RAM and high-speed drive connection interfaces.
2. The new 24-core processor is capable of processing large data sets both sequentially and in parallel in a short period of time. Its performance remains consistently high both on a single core and at full utilization of all resources.
3. The results of the network and drive performance tests show a well-balanced system; i.e. there are no obvious bottlenecks that reduce overall speed.
4. The temperature behavior of the components remains nominal despite intensive loads. This is evidence of the correct selection of the cooling system components.

The testing confirms the high potential of servers based on the Supermicro X13SAE-F and the Intel Core i9-14900KF for solving the widest range of computing tasks. The infrastructure is suitable for deploying large-scale projects related to big data analysis, machine learning and other resource-intensive applications.