On 4/16/07, Joerg t <tjoerg_at_yahoo.com> wrote: > > My system shows significant worse performance with ZFS - I suspect that this > > is because the devices are part of a ATAPI-RAID 0 already > > > but here is the story - A > > > UFS + Softupdates - > > > make -j 4 buildworld - app. 24 Min (Debugging kernel) > > > ZFS same command but with usr/obj on ZFS > > > make -j 4 buildworld - app. 1h12m > > > The weird thing is that one sees some fantastic peak transfer numbers e.g. > > up to 112MB/s - Compression is off > > I'm also seeing a very odd behavior regarding memory consumption, > with src and obj under zfs (different filesystems, same pool, compression off) the system begin to wire all memory and swap tops 1GB. Maybe there's something trashing memory on zfs ? > > this is a 1GB pentium 4 HT enabled box. > building a debug enabled kernel to help figure things out. > UFS with SoftUpdates on a ad0 <SAMSUNG SV3012H/RD100-06> ATA/ATAPI revision 6 20GB make buildkernel : 23.5 mins ZFS with compresson=gzip on 4x 320GB SATA RAIDZ make buildkernel: 27.5 mins systat -vm output after making the kernel on the ZFS volume: Mem:KB REAL VIRTUAL VN PAGER SWAP PAGER Tot Share Tot Share Free in out in out Act 21884 2928 91768 4968 818472 count All 189320 4484 2230492 8592 pages This is on: FreeBSD 7.0-CURRENT #5: Tue Apr 10 00:06:35 WEST 2007 root_at_xeon.bsdtech.org:/usr/obj/usr/src/sys/GENERIC WARNING: WITNESS option enabled, expect reduced performance. acpi_alloc_wakeup_handler: can't alloc wake memory ACPI APIC Table: <IntelR AWRDACPI> Timecounter "i8254" frequency 1193182 Hz quality 0 CPU: Intel(R) Xeon(TM) CPU 3.06GHz (3073.65-MHz 686-class CPU) Origin = "GenuineIntel" Id = 0xf29 Stepping = 9 Features=0xbfebfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE ,SSE2,SS,HTT,TM,PBE> Features2=0x4400<CNXT-ID,xTPR> Logical CPUs per core: 2 real memory = 1072562176 (1022 MB) FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs cpu0 (BSP): APIC ID: 0 cpu1 (AP): APIC ID: 1 With HTT enabled -- Joao BarrosReceived on Mon Apr 16 2007 - 20:46:53 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:08 UTC