Your software recommendations are always top notch. When you recommended ncdu I installed it just for fun and it really changed the way I deal with storage. ncdu is FAST, and it is so helpful that I had to install it on all my machines.
thanks, I suppose a good selection of tools. Actually I never used any of these. Quite an unusual case as the author often talks about MY favorite software 😀. That's why I think that's a great channel.
Phoronix test suite. However it has some flaws. For example linux kernel build benchmark runs faster on stable/outdated distros with GCC7 than on new distros with GCC13 and it's not always how many factors do affect the results.
0:27 to 0:40 I noticed the speed readings I get with speedtest-cli are vastly different than ookla on their site 1Gbls connection here... speedtest-cli claimed around 300 ish Mbps while ookla was in the mid 900 ish Mbps range for download...uploads were precisely the same.om speedtest-cli and ookla.. and yes..I made sure no webpages nor downloads were going on other than the obvious. also cat 6 ethernet cable straight to a damn Arris s33 modem ( supports up to 2.5Gbps by the way)...on a PC with a solid Intel 1Gbps Ethernet jack)
I did an rmlint on my laptop. In total 15085 files (tock over 5 hours)whereof 9221 are duplicates in 1627 groups. This equals 353.72 GB of duplicates which could be removed. 2828 other suspicious item(s) found, which may vary in size. Scanning tock in total 4h 32m 33.995s Can you help
@@tech_craft excel support a number of open file formats like ods and csv. Microsofts proprietary file formats can have compatibility issues between different versions of excel and is generally unadvisable to use if you don't have a specific reason to do so. Also becomes a lot harder to work with in cli applications. I suspected you had some legacy hardware that you wanted to use that only accepted xls but that didn't seem to be the case.
It will depend on the exact nature of the benchmark. In general no, but I've found MIPS ratios and speed ratios to be consistent across platforms when running the 7zip benchmarks. For example, the 5950X vs. M1 10-core has a ratio between 1.5-1.7 for both speed and MIPS. You can calculate your own normalised rating of Speed / Utilisation if you want to be absolutely certain, but for the 7zip benchmarks across x86 and M1 that I've done the Speed/Utilisation and MIPS/Utilisation ratings are always pretty much the exact same. Summary for my setup - the 5950x still beats the M1 in raw throughput for multi-core workloads (it has 32 hardware threads so it should!) but the cores on the M1 are twice as fast (both MIPS and throughput) than the cores on the 5950x.
Thanks for the feedback. This video was intended to be a quick intro to the possibilities, but I was a little worried about doing since I had to assume a lot of prior knowledge. Is there a particular topic that you'd like a deeper dive into?
Whoever's reading this, i pray that whatever you're going through gets better and whatever you're struggling with or worrying about is going to be fine and that everyone has a fantastic day! Amen
Benchmarking is one kind of performance testing. Many of these tools can be use for baseline testing and then back testing changes. We also have a presentation of stress testing here which is a completely different kind of performance test. Your comment is misleading at best.
@@tech_craft the only tool in the list where you can control the load is iperf. Having only stress test and benchmark are simply not enough to verify system performance requirements.
You are incorrect. Each of iperf, stress-ng sysbench and iozone provide the ability to control the load. So four of the six tools. I also never once said that this list is definitive. Again, your comment is misleading at best.
What are your favourite command line tools?
Tree, and fzf for fuzzy searching
htop
currently => `nala` for upgrading all my system easily ^^
This was another great video. Thanks. I hadn't heard of some of these utilities before.
Your software recommendations are always top notch. When you recommended ncdu I installed it just for fun and it really changed the way I deal with storage. ncdu is FAST, and it is so helpful that I had to install it on all my machines.
It's great, I'm using it more and more since I seem to accumulate massive video files at a ridiculous rate now.
Amazing work man!
Excellent video. Thank you.
Thanks for your sharing
thanks, I suppose a good selection of tools. Actually I never used any of these. Quite an unusual case as the author often talks about MY favorite software 😀. That's why I think that's a great channel.
I use networkQuality -v to test my connection test on Mac. I believe this is a built in command.
Phoronix test suite. However it has some flaws. For example linux kernel build benchmark runs faster on stable/outdated distros with GCC7 than on new distros with GCC13 and it's not always how many factors do affect the results.
nice tools. Can iozone also be used to test network storage nfs/smb performace?
Absolutely. If you cd into your mount directory and run iozone from there everything should work as expected.
0:27 to 0:40 I noticed the speed readings I get with speedtest-cli are vastly different than ookla on their site
1Gbls connection here... speedtest-cli claimed around 300 ish Mbps while ookla was in the mid 900 ish Mbps range for download...uploads were precisely the same.om speedtest-cli and ookla.. and yes..I made sure no webpages nor downloads were going on other than the obvious.
also cat 6 ethernet cable straight to a damn Arris s33 modem ( supports up to 2.5Gbps by the way)...on a PC with a solid Intel 1Gbps Ethernet jack)
Yeah, you will see that because speedtest is using different servers than ookla and others.
I did an rmlint on my laptop. In total 15085 files (tock over 5 hours)whereof 9221 are duplicates in 1627 groups. This equals 353.72 GB of duplicates which could be removed. 2828 other suspicious item(s) found, which may vary in size. Scanning tock in total 4h 32m 33.995s
Can you help
think you could make a video covering your emacs setup?
I'm planning a complete overhaul of my Emacs setup and will definitely do a video on it.
Im curious why you would us a proprietary file format for the spreadsheet. Is there a special reason or just a habit?
From my reading of the docs, Excel is the only option. If that’s wrong I’d love to know because I don’t like having Office on all my machines
@@tech_craft excel support a number of open file formats like ods and csv. Microsofts proprietary file formats can have compatibility issues between different versions of excel and is generally unadvisable to use if you don't have a specific reason to do so. Also becomes a lot harder to work with in cli applications. I suspected you had some legacy hardware that you wanted to use that only accepted xls but that didn't seem to be the case.
I don't think the issue here is Excel, I think it's iozone. If you try to output a CSV for example, you still just get an XLS file.
Do you think running these tools from windows 11 virtualized Linux would provide good results?
For the network tests yes, for the CPU, IO and memory I'm not sure - I imagine it would depend on the VM provider you are using.
Rob, what btop theme are you using?
I left it configured to use the TTY theme so that it picks up the Dracula theme I have in my Kitty terminal.
what is the app name that in start showing cpu memory so on info
That is btop.
Which monospace font are you using in the terminal?
It is JetBrains Mono.
@@tech_craft woah man, thanks! I love it! Just applied it in vs code, eclipse, xfce terminal and it looks stellar!
Can you tell us what font u use in cli ??
It's Jetbrains Mono: www.jetbrains.com/lp/mono/
Is it fair to compare MIPS on two different architectures?
It will depend on the exact nature of the benchmark. In general no, but I've found MIPS ratios and speed ratios to be consistent across platforms when running the 7zip benchmarks.
For example, the 5950X vs. M1 10-core has a ratio between 1.5-1.7 for both speed and MIPS. You can calculate your own normalised rating of Speed / Utilisation if you want to be absolutely certain, but for the 7zip benchmarks across x86 and M1 that I've done the Speed/Utilisation and MIPS/Utilisation ratings are always pretty much the exact same.
Summary for my setup - the 5950x still beats the M1 in raw throughput for multi-core workloads (it has 32 hardware threads so it should!) but the cores on the M1 are twice as fast (both MIPS and throughput) than the cores on the 5950x.
somehow I got a 403 running speedtest-cli on my Mac…
which is really weird 😂
That's a new one! I have never seen that before 😂
@@tech_craft used it again, works fine now.
I guess it's a random thing. Still really weired 😂
Computers are just so fickle!
supply unit op92caed
ong
tho
Comment
dude just look it up, there are free providers out there, it s goddaym expensive to buy, especially when you re a pro soft maker.
Did you comment on the wrong video? All these tools are free.
not dumbed down enuff for me
Thanks for the feedback. This video was intended to be a quick intro to the possibilities, but I was a little worried about doing since I had to assume a lot of prior knowledge. Is there a particular topic that you'd like a deeper dive into?
@@tech_craft just the actual command lines to copy and paste for install and run in terminal would do wonders
Whoever's reading this, i pray that whatever you're going through gets better and whatever you're struggling with or worrying about is going to be fine and that everyone has a fantastic day! Amen
I have seen you in bunch of videos
Seriously you're starting to become the bad thing I'm going through, so _get better_
I love pirating your software nom onom onomnom
you can not simply replace benchmarking with "performance testing" this is misleading at best
Benchmarking is one kind of performance testing. Many of these tools can be use for baseline testing and then back testing changes. We also have a presentation of stress testing here which is a completely different kind of performance test.
Your comment is misleading at best.
@@tech_craft the only tool in the list where you can control the load is iperf. Having only stress test and benchmark are simply not enough to verify system performance requirements.
You are incorrect. Each of iperf, stress-ng sysbench and iozone provide the ability to control the load. So four of the six tools.
I also never once said that this list is definitive. Again, your comment is misleading at best.
China