#Kernel_gcc_patch Kernel patch adds additional CPU options to the Linux kernel accessible under: Processor type and features ---> Processor family --->
Why a specific patch? The kernel uses its own set of CFLAGS, KCFLAGS. For exmaple, see:
- arch/x86/Makefile
- arch/x86/Makefile_32.cpu
- arch/x86/Kconfig.cpu
| CPU Family | GCC Optimization |
|---|---|
| Native optimizations autodetected by GCC | -march=native |
| AMD K10-family | -march=amdfam10 |
| AMD Family 10h (Barcelona) | -march=barcelona |
| AMD Family 14h (Bobcat) | -march=brver1 |
| AMD Family 15h (Bulldozer) | -march=bdver1 |
| Piledriver Family 15h (Piledriver) | -march=bdver2 |
| Intel 1st Gen Core i3/i5/i7-family (Nehalem) | -march=corei7 |
| Intel 2nd Gen Core i3/i5/i7-family (Sandybridge) | -march=corei7-avx |
| Intel 3rd Gen Core i3/i5/i7-family (Ivybridge) | -march=core-avx-i |
| Intel 4th Gen Core i3/i5/i7-family (Haswell) | -march=core-avx2 |
#Benchmarks ##Intro Three different machines running a generic x86-64 kernel and an otherwise identical kernel running with the optimized gcc options were tested using a make based endpoint.
##Conclusion There are small but real speed increases using a make endpoint to running with this patch. They are on par with the speed increase that the upstream sanctioned core2 option gives users so not including additional options seems somewhat arbitrary to me.
##Details
- Three test machines: Intel Xeon X3360, Intel i7-2620M, Intel Core i7-3660K.
- All ran the make benchmark (linked below) 35 times while booted into a 'generic' kernel. Then all ran the same make benchmark 35 times after booting into an optimized kernel. Below are the optimizations chosen for each machine.
- X3360 = core2
- i7-2620M = corei7-avx
- i7-3660K = core-avx-i
- Analyzed resulting distributions for statistical significance via ANOVA plots that clearly show statistically significant albeit small differences.
##Discussion
- All the assumptions for ANOVA are met:
- Data are normally distributed as show in the normal quantile plots.
- The population variances are fairly equal (Levene and Barlett tests).
- The ANOVA plots clearly show significance.
- Pair-wise analysis by Tukey-Kramer shows significance at the 0.05 level for all CPUs compared.
Below are the differences in median values:
| CPU | Difference in median value |
|---|---|
| core2 | +87.5 ms |
| core7-avx | +79.7 ms |
| core-avx-i | +257.2 ms |
##References
- Bash script that controls the benchmark: https://github.com/graysky2/bin/blob/master/bench
- Log file generated by script: http://repo-ck.com/bench/compile_time_optimization.txt.gz
##Credit
- Original author: jeroen AT linuxforge DOT net
- Link to original version: http://www.linuxforge.net/docs/linux/linux-gcc.php
##Data
###Corei7-avx vs. Generic
###Core-avx-i vs. Generic
###Core2 vs. Generic
