will install and tell the score but it doesn't say which mode it uses
(only points to likely hood of a mode usage inferred from speed). How
do you choose which mode you'r running (ata100/133)?
My matrox is the plus 9 - 80 gb 7200 rpm 2mb cache
The drive I mentioned above, the Maxtor 80GB Plus 9, 7K2 RPM, has 8MB
cache. Currently it's attached to a system with Via 8233 Southbridge
(IDE controller supporting up to ATA100), is a FAT32, formatted empty
drive. The accoustic management mode was not changed, left at factory
default which (IIRC) is set to AM-On, Fast, which is a median setting,
faster seeks than AM Slow but slower than AM-Off. It's interesting to
note that this Maxtor drive "appears" to be nearly the opposite of a
Western Digital 8MB cache drive, with the WD having much better write
scores but lower read scores in same system... unfortunately I don't
have a spare WD at the moment so benchmarks in same environment can't
be included below. I presume it's due to WD having a better caching
algorithm, as it's writes are significantly faster than the reads, at
least in a few, < ~8MB at a time.
The test system wasn't tweaked in any way, just happened to have a
Win98 installation on a difference, Maxtor Plus 8 HDD and threw it
into a case that already had the board/CPU/etc in it, then chipset
driver & Via Latency Patch PFD v1.05 installed.
ATTO peak write was 65MB/s, peak read 55 MB/s (average of 5 longest
transfers) on a board running it in ATA100 mode.
Sandra 2002 File System Benchmark for that drive was 38204 KB/s, avg.
of 3 runs which varied only 0.2% . At this point I'll mention that
the ~ 22MB/s range you reported previously from the Sandra Benchmark
isn't THAT far off from what this version of Sandra shows as a
baseline score for an ATA100 7K2 80GB NTFS drive, 29200 KB/s.
HDTach read testing showed burst rate off the scale (> 80 MB/s) and
read Max: 63042, Avg: 47890. HDTach suggests some sort of problem
because CPU utilization was at 55% (Athlon XP2000) for both the Plus 8
and Plus 9 drives... this is not typical of the chipset, the OS or
chipset driver may be misconfigured, it's usually closer to 5%. I
then uninstalled the Via Latency Patch (PFD 1.05) and reinstalled the
Via 4in1 v4.35 chipset driver.
With the Latench Patch removed, HDTach CPU Utilization dropped back to
reasonable levels, 4.5%. Max and Avg read speed in HDTach stayed same
as with Latency Patch installed. Sandra scores "might" have dropped a
few small amount after removing the Latency Patch, but I don't have
time at the moment to do enough runs with & without the patch to do a
meaningful enough comparision as the difference is small.
Another Maxtor (Plus 8) ATA133 drive (2MB cache, 7K2 RPM) is attached
to same system, FAT32, relatively highly fragmented (testbed system
with a LOT of reconfigurations) and scores a little lower, around
46MB/s read & write on ATTO and 36329 KB/s on Sandra File System
Benchmark. HDTach scores for it were Max: 62828 KB/s, Avg: 51287
KB/s.
At this point in the testing I disabled DMA for the drive in Windows,
and/or replaced the ATA133 cable with an ATA33 cable. All following
scores were with one or both of those misconfigurations present:
DMA OFF, ATA133 Cable
ATTO - Write 7 MB/s, Read 6 MB/s
HDTach - Burst 6 MB/s, Max 6 MB/s , Avg 6 MB/s
Sandra - (took so long I went and had a snack) 5299 KB/s
DMA OFF, ATA33 Cable
ATTO - Write 7MB/s, Read 6 MB/s
HDTach - Burst 6MB/s, Max 6MB/s, Avg 6 MB/s
Sandra - I don't need another snack, sorry.
DMA ON, ATA33 Cable
ATTO - Write 31 MB/s, Read 29 MB/s
HDTach - Burst 31 MB/s, Max 31 MB/s, Avg 30 MB/s
Sandra - 21603 KB/s
Based on your previous question about ATA33 cables, and that last
"21603 KB/s" score I posted, it makes sense you'd be getting 22000
MB/s with that drive if using the wrong cable, so it's running in
ATA33 mode.
Dave
