On 2003-05-06 22:26 +0000, W. Josephson wrote: > I recently cvsup'd (yesterday afternoon EDT) and again this morning. > I'm seeing a major decrease in performance and wondering if anyone > else is seeing the same. Compiliation and interactive X11 use seems > to be very slow (long times to compile individual files and to repaint > windows) as compared to -CURRENT as of about two weeks ago. In fact, > it took all afternoon and early evening to build world and the kernel. > The only obvious things were ACPI complaining about zero-length > allocations, new problems assigning interrupts at boot time, and a > reported clock rate about one third of what it is and what the kernel > has reported in the past. ACPI has never really worked quite right > for me anyway on this machine (a Dell Inspiron 4000), so I doubt that > is the problem and the message about sio1 is nothing new; 4.x said the > same thing. It feels a lot like I'm losing interrupts, but I haven't > had a chance to do much poking around and nothing has changed > (hardware or bios) in a year. Trimmed boot messages follow... I see the same thing. I'd been running -CURRENT on this box for around 18 months, and decided to reinstall with a fresh JPSNAP (20030506) this morning, since things had gotten messy. Even after a full rebuild of world/kernel, things are quite slow and unresponsive. No fancy flags or anything passed to GCC. I also see the ACPI complaints in dmesg. My last world/kernel was from April 20, at which point performance was quite decent (relative to other points in the 5.x series). -- Munish ChopraReceived on Tue May 06 2003 - 17:54:50 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:06 UTC