1 |
I've been trying to up the performance of the disks in my Ultra10, and |
2 |
I'm running across something I can't quite figure out... |
3 |
|
4 |
I have 2 identical 120GB IDE drives. One is installed on the primary |
5 |
IDE interface of the motherboard. The other is installed on the primary |
6 |
IDE interface of a Promise Ultra66 PCI/IDE card (just recently flashed |
7 |
to the most-recent bios). The only other drives in the box are the stock |
8 |
CDROM (which is on the secondary IDE interface of the motherboard |
9 |
[hdc]), and a SCSI tape backup (which is plugged into a PCI/SCSI card). |
10 |
|
11 |
Here are the details: |
12 |
|
13 |
sun ~ # hdparm -i /dev/hda |
14 |
|
15 |
/dev/hda: |
16 |
|
17 |
Model=WDC WD1200JB-00GVA0, FwRev=08.02D08, SerialNo=WD-WCALA2043613 |
18 |
Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq } |
19 |
RawCHS=16383/16/63, TrkSize=57600, SectSize=600, ECCbytes=74 |
20 |
BuffType=DualPortCache, BuffSize=8192kB, MaxMultSect=16, MultSect=16 |
21 |
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=234441648 |
22 |
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120} |
23 |
PIO modes: pio0 pio1 pio2 pio3 pio4 |
24 |
DMA modes: mdma0 mdma1 *mdma2 |
25 |
UDMA modes: udma0 udma1 udma2 udma3 udma4 udma5 |
26 |
AdvancedPM=no WriteCache=enabled |
27 |
Drive conforms to: device does not report version: |
28 |
|
29 |
* signifies the current active mode |
30 |
|
31 |
sun ~ # hdparm -i /dev/hde |
32 |
|
33 |
/dev/hde: |
34 |
|
35 |
Model=WDC WD1200JB-00GVA0, FwRev=08.02D08, SerialNo=WD-WCALK1317943 |
36 |
Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq } |
37 |
RawCHS=16383/16/63, TrkSize=57600, SectSize=600, ECCbytes=74 |
38 |
BuffType=DualPortCache, BuffSize=8192kB, MaxMultSect=16, MultSect=16 |
39 |
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=234441648 |
40 |
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120} |
41 |
PIO modes: pio0 pio1 pio2 pio3 pio4 |
42 |
DMA modes: mdma0 mdma1 mdma2 |
43 |
UDMA modes: udma0 udma1 udma2 udma3 *udma4 udma5 |
44 |
AdvancedPM=no WriteCache=enabled |
45 |
Drive conforms to: device does not report version: |
46 |
|
47 |
* signifies the current active mode |
48 |
|
49 |
sun log # hdparm /dev/hda |
50 |
|
51 |
/dev/hda: |
52 |
multcount = 16 (on) |
53 |
IO_support = 0 (default 16-bit) |
54 |
unmaskirq = 0 (off) |
55 |
using_dma = 1 (on) |
56 |
keepsettings = 0 (off) |
57 |
readonly = 0 (off) |
58 |
readahead = 8 (on) |
59 |
geometry = 14593/255/63, sectors = 120034123776, start = 0 |
60 |
|
61 |
sun log # hdparm /dev/hde |
62 |
|
63 |
/dev/hde: |
64 |
multcount = 16 (on) |
65 |
IO_support = 0 (default 16-bit) |
66 |
unmaskirq = 0 (off) |
67 |
using_dma = 1 (on) |
68 |
keepsettings = 0 (off) |
69 |
readonly = 0 (off) |
70 |
readahead = 8 (on) |
71 |
geometry = 14593/255/63, sectors = 120034123776, start = 0 |
72 |
|
73 |
As you can see, the drives are identical. The first one comes up in |
74 |
mdma2 mode, and the second one comes up in udma4 mode. This makes |
75 |
sense, since the motherboard doesn't support the faster interface (only |
76 |
up to 33, right?). What doesn't make sense is this: |
77 |
|
78 |
sun ~ # hdparm -Tt /dev/hda |
79 |
|
80 |
/dev/hda: |
81 |
Timing O_DIRECT cached reads: 564 MB in 2.01 seconds = 280.60 MB/sec |
82 |
Timing O_DIRECT disk reads: 46 MB in 3.04 seconds = 15.13 MB/sec |
83 |
|
84 |
sun ~ # hdparm -Tt /dev/hde |
85 |
|
86 |
/dev/hde: |
87 |
Timing O_DIRECT cached reads: 576 MB in 2.00 seconds = 288.00 MB/sec |
88 |
Timing O_DIRECT disk reads: 30 MB in 3.13 seconds = 9.58 MB/sec |
89 |
|
90 |
I've run them multiple times, and it always comes out similar to above. |
91 |
This is on a relatively dormant system. Can anyone tell me why a drive |
92 |
plugged into the motherboard's IDE interface (running in mdma2 mode) |
93 |
would outperform the same drive plugged into an Ultra66 card (running in |
94 |
udma4 mode)? With these results, I'd be better off putting the second |
95 |
drive back on the chain with the CDROM. |
96 |
|
97 |
Thanks in advance... |
98 |
|
99 |
-- |
100 |
gentoo-sparc@g.o mailing list |