Hello
Are there programs in Linux like in Windows?
PCMark 10
PassMark PerformanceTest
Hello,
If you are wanting to test memory performance in Linux, there are certainly options to be sure. One that comes to mind is RAMspeed/SMP. This provides the ramsmp command which can be used for running a variety of different memory benchmarking tests.
If you're interested in memory latency testing (and you're using a system with a recent Intel CPU), Intel actually provide a utility for that which can be downloaded for free from their Web site. You can find out more about this utility (MLC) here:
Hope this helps ! If neither of these cover what you're after, then if you can give us a bit more detail about what exactly it is you want to test, we may be able to assist further.
2 Likes
Just noticed you were also asking about disk - I've used iozone for filesystem performance testing in the past. For general-purpose system benchmarking, including I/O performance, you can check out Sysbench. There's also still the old-fashioned and no-longer-updated (but still available and working) UnixBench, which again is a general-purpose benchmark covering a wide variety of things (though its results may be less meaningful on some modern systems).
2 Likes
Thank for advice. my server test result passmark for linux
How should I interpret these results? Are the values good relative to the server?
I also don't know the purpose of the tests. Can I get some brief information about them?
Intel(R) Xeon(R) Gold 5317 CPU @ 3.00GHz (x86_64)
24 cores @ 3600 MHz | 1007.7 GiB RAM
Number of Processes: 48 | Test Iterations: 3 | Test Duration: Short
Iteration: 3/3
Tests Status Result
Integer Math Complete 196114 Million Operations/s
Floating Point Math Complete 108300 Million Operations/s
Prime Numbers Complete 232 Million Primes/s
Sorting Complete 91357 Thousand Strings/s
Encryption Complete 42756 MB/s
Compression Complete 687 MB/s
CPU Single Threaded Complete 2518 Million Operations/s
Physics Complete 4848 Frames/s
Extended Instructions (SSE) Complete 35364 Million Matrices/s
CPU Mark
Complete 48243.85
Results not submitted
If you are not aware of either the purpose or the meaning of these tests, could I ask why it is that you are inquiring about them ? Presumably there is some actual underlying problem you are trying to solve, or some matter that you are attempting to investigate.
Rather than focussing here on test results which appear not to be very meaningful to you, it might be a better idea to give us some more background information on what the final goal you are attempting to accomplish actually is. Then we will know for sure if we are giving you good advice, or if in fact what we are telling you is not helping you get closer to where you ultimately need to be.
1 Like
to test if there is an error in the read/write operations on the server in general
Ah, now testing if there is an error (diagnostics) is a different thing from testing the performance of a server (benchmarking). If you want to know if the memory in your server is working correctly, then you probably want to look at checks like MemTest86+. This is available both in a bootable form, and as a stand-alone piece of software that can be installed on most Linux distributions.
For disk testing, there is no one straightforward answer. This all depends on what kind of storage controller your server has, what type of disks, and how those disks are set up (stand-alone disks, hardware RAID, software RAID, etc.). Most likely if you are using a dedicated storage or RAID controller it will be able to tell you the health of the drives in the system by one means or another; and if you are using stand-alone disks and a straightforward disk controller you can use utilities such as smartctl to view the SMART status of your hard drives.
Again for CPU testing, there is no straightforward one answer here. Generally if your CPU is not working correctly this is very apparent, as the system will either crash when it would not be expected to, or perhaps fail to power on or boot up altogether. Again depending on your server's hardware it may have ways of monitoring this if it has a dedicated Baseboard Management Controller (BMC) or Integrated Lights-Out Management (iLO)-style system, in which case you will be able to view your CPU status (and plenty more besides) by looking at the interface to this.
Generally speaking, my advice would be: unless you have a specific reason to worry about the health of your hardware, don't worry too much about it. If there is a problem, this will become apparent in normal operation, and if you are using modern server hardware, you are likely to receive plenty of warning before something goes truly wrong. As long as you have mechanisms in place to monitor the health of your hardware, this is almost always sufficient.
Now there are exceptions to this of course - and note that "don't worry too much about hardware being dodgy" is by no means the same thing as saying "don't worry about your system's stability, or redundancy, or backups". Regardless of how good (or bad) the quality of your system's hardware or software is, all those other things are always just as important at all times.
1 Like
Hi @soolan,
in addition to @drysdalk's detailed explanations, even SMART is already mentioned:
To check disks for errors, there is https://www.smartmontools.org. It's available on almost all Un*x platforms. It can run as a service (smartd) and can send messages on error events, e.g. via email or to system logs (which in turn can be monitored with https://logcheck.org). Or it can be started as a standalone proc (smartctl), e.g. as short or long self-test via cron. It has tons of options. Almost alll (internal) devices support Self-Monitoring, Analysis and Reporting Technology - Wikipedia.
1 Like