PHENOM II X4 940 BE Experience (was Re: booting from SATA DVD-RAM, no go SOLVED)

[Mars G Miro]

On Sun, Mar 29, 2009 at 8:59 PM, Gavin Atkinson
<gavin.atkinson at ury.york.ac.uk> wrote:
> On Sat, 28 Mar 2009, Mars G Miro wrote:
[snip]
>> rrrrrrrrrr. plugging in an old CD-ROM still gives me the same thing.
>>
>> Next stop: PXE, or just install FreeBSD on the HD in a separate
>> machine altogether ...
>
> This was probably fixed by the most recent commits to
> src/sys/boot/i386/libi386/bioscd.c - they were done within the last couple
> of weeks, so the March snapshot may well have these changes included.
>


I'll check on that once new snapshot ISOs are available.


> Gavin


Ok, I managed to bootstrap most recent CURRENT on the SSD (OCZ Core
Series V2 SATA II 2.5") via its USB, from another CURRENT box (I have
a set of scripts to do this). It's good to know the new USB2 stuff
don't crash plugging a USB device anymore!! After I bootstrapped
FreeBSD on it, I rsync'd /usr/src and /usr/obj and everything went
fine. Previous experiences would crash (7.X). Also I notice this in
dmesg:

   ad4: FAILURE - SET_MULTI status=51<READY,DSC,ERROR> error=4<ABORTED>

 but it seems to be harmless.

Full verbose dmesg: http://pastebin.com/f2ae007e6

Rebuild stats:
 make -j8 buildworld
        3379.225u 463.832s 19:27.34 329.2%      6637+2112k
21241+9092io 17130pf+0w
 make -j16 buildworld
        3386.927u 467.825s 20:09.77 318.6%      6649+2115k
21301+9121io 17440pf+0w
 make -j8 buildkernel KERNCONF=BLACKBOX
        728.654u 77.502s 6:35.60 203.7% 6446+2024k 7116+10170io 4pf+0w

w/c is somehow at par w/ the 6-CORE DUNNINGTON stats I posted last
October: http://markmail.org/message/vtadzdlvakflw33u#query:+page:1+mid:g64htyb5bxhfcbrq+state:results

diskinfo -ctv:

ad4
        512             # sectorsize
        64105742336     # mediasize in bytes (60G)
        125206528       # mediasize in sectors
        124212          # Cylinders according to firmware.
        16              # Heads according to firmware.
        63              # Sectors according to firmware.
        ad:MK06090307483000F    # Disk ident.

I/O command overhead:
        time to read 10MB block      0.078007 sec       =    0.004 msec/sector
        time to read 20480 sectors   3.852412 sec       =    0.188 msec/sector
        calculated command overhead                     =    0.184 msec/sector

Seek times:
        Full stroke:      250 iter in   0.048635 sec =    0.195 msec
        Half stroke:      250 iter in   0.046734 sec =    0.187 msec
        Quarter stroke:   500 iter in   0.091953 sec =    0.184 msec
        Short forward:    400 iter in   0.072450 sec =    0.181 msec
        Short backward:   400 iter in   0.072721 sec =    0.182 msec
        Seq outer:       2048 iter in   0.401060 sec =    0.196 msec
        Seq inner:       2048 iter in   0.409373 sec =    0.200 msec
Transfer rates:
        outside:       102400 kbytes in   0.726566 sec =   140937 kbytes/sec
        middle:        102400 kbytes in   0.729195 sec =   140429 kbytes/sec
        inside:        102400 kbytes in   0.742233 sec =   137962 kbytes/sec

/me goes back building stuff on it ;-)

Thanks guys!!!





-- 
cheers
mars
-----
Yogi Berra  - "A nickel ain't worth a dime anymore."