nice is not nice; SCHED_ULE problem?

From: Yoshihiro Ota <ota_at_j.email.ne.jp>
Date: Wed, 4 Mar 2009 00:38:18 -0500
Hi, all.

I have a laptop that has 2 AMD CPUs.
I started 2 lzma compression processes with nice +10 and +20.
Then, I started to compile a port, qemu.

Back in SCHED_4BSD time, whenever I build a port while compression
was running with nice +20, build used take most of CPUs and very little
to nice +20 processes.

The build process was slow, for example, when it was running configure
where lots of processes were created.  However, when a large C file was
being compiled, it took most of CPU away from nice +20 processes.
When I suspended one of the compression, build speeds up significantly.

It looked process generation had less priorities than nice +20s.
I tested on 7.1-RELEASE and it happens there, too.
Could someone take a look?

Thanks,
Hiro


Copyright (c) 1992-2009 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
        The Regents of the University of California. All rights reserved.
FreeBSD is a registered trademark of The FreeBSD Foundation.
FreeBSD 8.0-CURRENT #7 r189249M: Tue Mar  3 11:00:21 EST 2009
    root_at_XXX:/usr/obj/amd64/usr/src/head/sys/GENERIC
WARNING: WITNESS option enabled, expect reduced performance.
Timecounter "i8254" frequency 1193182 Hz quality 0
CPU: AMD Turion(tm) 64 X2 Mobile Technology TL-58 (1908.70-MHz K8-class CPU)
  Origin = "AuthenticAMD"  Id = 0x60f81  Stepping = 1
  Features=0x178bfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,MMX,FXSR,SSE,SSE2,HTT>
  Features2=0x2001<SSE3,CX16>
  AMD Features=0xea500800<SYSCALL,NX,MMX+,FFXSR,RDTSCP,LM,3DNow!+,3DNow!>
  AMD Features2=0x11f<LAHF,CMP,SVM,ExtAPIC,CR8,Prefetch>
  Cores per package: 2
usable memory = 1525399552 (1454 MB)
avail memory  = 1472196608 (1403 MB)
ACPI APIC Table: <HP             APIC  >
FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs
 cpu0 (BSP): APIC ID:  0
 cpu1 (AP): APIC ID:  1
Received on Wed Mar 04 2009 - 04:38:23 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:43 UTC