Well i used them on the Ryzen and it worked? I have little knowledge of bash shell. I just use what i know and what i can.
We all start somewhere, and nowadays Linux doesnât need as much familiarity with the shell as you used to. At least this was a fun little dive into the shell for you 
Yes, I like learning new things. I tend to stick to more hardware. But i only know what i know and thatâs pretty tiny in this realm!
I thought Iâd try this too - but I thought it was supposed to be a slow process? Guess not these days
real 1m41.006s
user 20m50.037s
sys 2m8.172s
This is a Ryzen 7 3700x. Is it good? 
3 Likes
The Ryzenâs are pretty darn good. I was hoping to get a 4700G but i see on the AMD site itâs OEM only.
Any modern processor will compile the kernel within a few minutes so long as you use all the threads you have available. AMDs will do better than Intel because they tend to have more cores. A more demanding test would be something like compiling the chromium source code which would take like an hour on a half decent machine, and a lot longer on an older one; but who wants to spend an hour plus on something done just for a bit of fun and no real benefit? 
real 3m28,307s
user 23m23,515s
sys 2m17,948s
this is a 6700K 4 cores / 8 threads
16GB DDR4
SSD
on performance scheduling ( check this )
i have to create menuconfig , no old config present
Havenât got a clue. Last time i went to the kitchen for coffee and the bathroom, and it was done when I came back.
Three minutes tops and thatâs on an old processor.
Hmm iâm getting this:
$ time(make -j$(nproc))
zsh: number expected
I could probably pass the core number directly, but what would that be? this is a 6core/12thread cpu. So is nproc 6 or 12?
Nevermind, I simply ran bash and ran the test under bash instead of zsh
real 3m0,329s
user 30m40,902s
sys 2m53,927s
On an Intel i7-10710U
- Logical cores are counted as well.
Doesnât there have to be a space between time and the parenthesis?
This is on my i5-9700k
real 3m34,096s
user 16m32,421s
sys 1m54,819s
1 Like
real 6m9,419s
user 38m17,208s
sys 3m33,684s
$ inxi -Cx
CPU: Topology: Quad Core model: Intel Core i7-4702MQ bits: 64 type: MT MCP arch: Haswell rev: 3 L2 cache: 6144 KiB
flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 35132
Speed: 2258 MHz min/max: 800/3200 MHz Core speeds (MHz): 1: 1487 2: 1214 3: 2168 4: 1621 5: 1644 6: 1184 7: 2072
8: 1772
nproc is a command itself that gives you the number of threads your CPU has. Running it by itself would give you 12, so the command without calling nproc would be time(make -j12).
I used time from the time package instead, only printing out real elapsed time. Also used the -s switch for the make command to suppress compilation output. (Note: built in tmpfs):
$ /usr/bin/time -f%E make -j24 -s
[...]
1:06.52
CPU: 12-Core AMD Ryzen 9 3900X (-MT MCP-) speed/min/max: 2195/2200/3800 MHz
Building with clang-11 is much slower:
$ /usr/bin/time -f%E make LLVM=1 -j24 -s
1:23.36
CPU: Dual Core Intel Pentium G2020 (-MCP-) speed/min/max: 1597/1600/2900 MHz
Kernel: 5.4.58-1-lts x86_64 Up: 9m Mem: 1344.4/3902.7 MiB (34.4%)
Storage: 465.76 GiB (7.4% used) Procs: 124 Shell: Zsh 5.8 inxi: 3.1.06
I donât even want to start the compilation
4 Likes
Just do it when you go to bedâŚ
2 Likes
I ran this on three different machines âŚ
Current Laptop (ASUS UX461UN): ~11m 42s
CPU: Intel i7-8550U (8) @ 4.000GHz
GPU: Intel UHD Graphics 620
GPU: NVIDIA GeForce MX150
Memory: 15866MiB
time make -j$(nproc) âŚ
real 11m41.908s
user 82m28.841s
sys 8m12.556s
Old Laptop (Clevo P650RG): ~4m 44s
CPU: Intel i7-6700HQ (8) @ 3.500GHz
GPU: NVIDIA GeForce GTX 980M
GPU: Intel HD Graphics 530
Memory: 64141MiB
time make -j$(nproc) âŚ
real 4m44.188s
user 32m54.099s
sys 3m3.072s
Home File Server (DIY frankenputer): ~84m 2s
CPU: AMD C-60 APU (2) @ 1.000GHz
GPU: AMD ATI Radeon HD 6290
Memory: 7562MiB
time make -j$(nproc) âŚ
real 84m2.196s
user 148m3.326s
sys 15m22.278s
So I expected the frankenputer to take forever, but I wasnât expecting the old laptop to be faster (significantly faster) than my current one. To be honest, I havenât done any tuning on the current laptop and I know I have some issues with how I have power management set up. There could be some serious throttling going on there ⌠thatâll be a project for another day! 
1 Like
Good to see you putting old hardware to use in that file server. Too often people chuck em out and buy something new.
It is interesting to see these super-slim laptops with i7 CPUs. I would guess that they could not properly cool it down and there has to be some serious throttling.
For example I was looking at some reviews of Lenovo Yoga X380 (my work computer) and it turned our that my version with i5 CPU is in every benchmark significantly faster compared to a version with i7. Every other component was the same. That is not something that could be patched after a release. But then again, maybe the designer didnât anticipate that someone like me would run FEM analysis on it. 
2 Likes
Two possible reasons:
The i7-s7xx is a higher-performing CPU than the i7-s5xx, and
The U-series are low-voltage units designed for power saving, HQ-series are relatively high-performing mobile parts.
https://ark.intel.com/content/www/us/en/ark/compare.html?productIds=88967,122589
https://www.cpubenchmark.net/compare/Intel-i7-6700HQ-vs-Intel-i7-8550U/2586vs3064
3 Likes