Re: Hyperthreading/SMT and NetBSD; how well does it work?

[Andreas Wrede <andreas%wrede.ca@localhost>]

On Jan 30, 2009, at 17:38 , Andrew Doran wrote:

> Hi,
>
> I think it would be interesting to know how well -current / 5.0 does  
> on SMT
> chips like the Pentium 4. I don't have such a setup currently, so..  
> Anyone
> share my interest and have the necessary kit? I'd love to see how  
> build.sh
> or a kernel compile does with HT enabled and HT disabled.


Building a GENERIC kernel from a -current source tree on a dual-core  
Atom 330 cpu and 2GB of memory running 5.0_RC1/amd64

build.sh -j 8 kernel=GENERIC run times:

With all 4 cpus:      8 min 41 sec
With 0 and 1 online: 11 min 07 sec
With 0 and 2 online: 16 min 23 sec
With only 0 online:  21 min 39 sec

Looks like HT gets us a 25% reduction in runtime.

cpu0: Intel Pentium Pro, II or III (686-class), 1596.16 MHz, id 0x106c2
cpu0: features  
0xbfe9fbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR>
cpu0: features 0xbfe9fbff<PGE,MCA,CMOV,PAT,CFLUSH,DS,ACPI,MMX>
cpu0: features 0xbfe9fbff<FXSR,SSE,SSE2,SS,HTT,TM,SBF>
cpu0: features2 0x40e31d<SSE3,DTES64,MONITOR,DS- 
CPL,TM2,SSSE3,CX16,xTPR,PDCM>
cpu0: features3 0x20100800<SYSCALL/SYSRET,XD,EM64T>
cpu0: "Intel(R) Atom(TM) CPU  330   @ 1.60GHz"
cpu0: I-cache 32KB 64B/line 8-way
cpu0: Initial APIC ID 3
cpu0: Cluster/Package ID 0
cpu0: Core ID 0
cpu0: SMT ID 0

cpu1: Core ID 1
cpu1: SMT ID 0

cpu2: Core ID 0
cpu2: SMT ID 1

cpu3: Core ID 1
cpu3: SMT ID 1

--
    aew