Apache OpenOffice (AOO) Bugzilla – Issue 47609
m95 and m93 uses up to 96% cpu
Last modified: 2005-06-07 12:29:48 UTC
M93 with two writer docs open, one a manual and the other a blank text doc, and one drawing window open uses up to 70% of cpu as shown in the following top() output: top - 13:28:10 up 4 days, 2:35, 6 users, load average: 2.25, 1.95, 1.57 Tasks: 149 total, 4 running, 145 sleeping, 0 stopped, 0 zombie Cpu(s): 79.8% us, 1.7% sy, 0.0% ni, 18.5% id, 0.0% wa, 0.0% hi, 0.0% si Mem: 775436k total, 711292k used, 64144k free, 63768k buffers Swap: 1028152k total, 235832k used, 792320k free, 266864k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7009 gerry 16 0 237m 108m 61m S 70.8 14.4 171:24.16 soffice.bin THis is excessive in comparison to 1.1.x.
As I have worked with my document, I notice that up to 96% of cpu cycles are being used during a save. This is serious. I have not been able to determine why.
Did an strace of soffice from a blank new document, loading a recently used doc, updated the TOC, inserted a graphic, saved the document and closed it.
Created attachment 25148 [details] zip of strace session.
Tried m94 from Pavel's repostitory. This release only pegs the cpu during a save. I suspect this is due to compression occurring during the save. Nonetheless, on large documents this approaches 98% cpu usgage. Is there a way to nice the priority of the save such that cpu usage is reduced?
Hi grsingleton, is there a way to get your "sample documents"?
You are going to hate me. Use http://documentation.openoffice.org/manuals/OOo1.x.x/user_guide.sxw It doesn't matter about the format I have found. i.e. I started with this and saved it as odt. BTW m94 does not exhibit the same problem of pegging the cpu while idle. Only m93.
Regression for m94. m95 is consuming upto 98% of cpu. At idle with user guide loaded cpu is 62%.
Resetting version. We need to to when this issue has been found. When there has been no fixing iit would also occure in m95.
That is fine but m94 does not exhibit this behaviour. Since changing the release if verboten, this will me an new issue.
Hi grsingleton, I opened your sample document using OOom99 on FC3. Installation of OOo caused rpm to use 82.8% CPU (peak!)... Opening your document caused Office to use up to 90%CPU (peak!) (==>decompressing...!) ... While watching ksysgard it used up to 12%CPU (peak!)... Scrolling through the document caused OOo to use up to 10% CPU (peak!)... Due to continous formatting of the document I think this is ok (?)...
WHat is strange about this behaviour is that m94 does not exhibit the same cpu hunger. Yes I agree that compression does affect things; however, m94 only consumes cpu cycles during the save and not afterward. m97 is slightly better but still consumes more cpu at idle than m94. To me this is a regression that should be addressed and don't you agree?
I have another discovery. OOo1.1.4 with the user guide loaded consumes less than 5% cpu. I also straced the 1.1.4 instance and see that gettimeofday is executed in the same manner as in 1.9m??. So now the question is how to track down why 1.9 uses 33% cpu whilst 1.1.4 does not. Any ideas that I can try?
Hi grsingleton, I just tried your document using OOom103 on FedoraCore3 and didn't get your problems... Office needs some time to load, unzip and format the document... I think that's normal... When leaving Office idle I get a CPU usage up to 5% if I bring Office back on top of all opened windows... Maybe it's useful to know something about your system... (Distribution, CPU, RAM, DesktopEnvironment,...)
Well that is good that m103 works. Is this release publically available? I would like to try it. I am currently using m99 which uses about 34% cpu at idle. Not the best. You have asked for system specs: # cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 2 model name : Intel(R) Pentium(R) 4 CPU 2.40GHz stepping : 4 cpu MHz : 2400.613 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm bogomips : 4751.36 # cat /proc/meminfo MemTotal: 775436 kB MemFree: 41364 kB Buffers: 59272 kB Cached: 222736 kB SwapCached: 80784 kB Active: 566772 kB Inactive: 85436 kB HighTotal: 0 kB HighFree: 0 kB LowTotal: 775436 kB LowFree: 41364 kB SwapTotal: 1028152 kB SwapFree: 743036 kB Dirty: 8 kB Writeback: 0 kB Mapped: 456680 kB Slab: 65920 kB CommitLimit: 1415868 kB Committed_AS: 1611988 kB PageTables: 5368 kB VmallocTotal: 245752 kB VmallocUsed: 6124 kB VmallocChunk: 239444 kB HugePages_Total: 0 HugePages_Free: 0 Hugepagesize: 4096 kB Desktop is Gnome 2.8 All FC3 is current with Up2date(). I have re-opened it because the problem exists with the current beta candidate and I recommend that it stay that way until m103 is released on download.openoffice.org. I am pleased that progress is being made.
Tried m102 from pavel's repository and am testing. Things are definitely improving as idle cpu usage with this version and the 2.0 user guide loaded is consuming between 5% and 7% cpu. Looking good. I will continue testing with this release and look forward to the next official beta candidate. If the bc is m103 or later and behaves as you have reported then we can close the issue.
Hi grsingleton, m104 is available.... May you test with that build?
*** Issue 47944 has been marked as a duplicate of this issue. ***
I will certainly test m104 when it shows up at http://download.openoffice.org/680/index.html or whatever is posted there.
Closing. m104 works better than m102.
Hi grsingleton, thanks for your response... I'm closing this issue now... :)