Intel SSD reviewed.

Really lame review there. This is old news and there are many better reviews on these things.
 
Its from the Inq - can't expect miracles.

I haven't seen too many reviews of the 32GB SLC drive. Most seem to point to the 80GB MLC drive. Intel has a lame naming system for these, the X25-E and X25-M.

I believe the SLC drive is actually crippled by the SATA interface, it should technically do much better. Whoops, my bad - they are still reviewing the 80GB on the Inq, Doh!
 
Sandra? Are you kidding me. Their disk benchmarks are a joke:

sandra.png


Suffice to say. that is not correct!
 
Not trying to defend them in any way, but what, exactly, is not correct in that screenshot?
 
I don't know of any disk setup that can do 1.38 GB/sec. Do you?

I think my clustered NetApp 3070 Filers with one, 9 shelf, 126 Fiber-Channel 15K RPM, disk aggregate cap at close to 4GB/sec. But that also cost like $650,000.
 
I don't know of any disk setup that can do 1.38 GB/sec. Do you?

Well, if you look at it, you'll see that the two high-speed drives are part of a Raid array which totals over 10terrabyte; that takes quite alot of disks, and since each disk, at least in a Raid0 array, effectively doubles the reading speed, it's not illogical that it could exceed 1gb/s.

There are also SSD disks coming, with PCI-E connectors, that can actually do reads at 900+mb/s single-handedly.
 
I think my clustered NetApp 3070 Filers with one, 9 shelf, 126 Fiber-Channel 15K RPM, disk aggregate cap at close to 4GB/sec. But that also cost like $650,000.

no.. they don't. I have tons of netapp storage at work and the 4 GB a sec is actually 4 gigabits/sec so 500 megabytes/sec which is actually less than my raid array. Of course the random I/O on those netapps with FC drives pwns my raid =P

netapp love:
http://box.houkouonchi.net/gallery/main.php?g2_itemId=535

Of course these are old ones though.
 
Well, if you look at it, you'll see that the two high-speed drives are part of a Raid array which totals over 10terrabyte; that takes quite alot of disks, and since each disk, at least in a Raid0 array, effectively doubles the reading speed, it's not illogical that it could exceed 1gb/s.

There are also SSD disks coming, with PCI-E connectors, that can actually do reads at 900+mb/s single-handedly.

Well I just don't know of any disk system really (except maybe software raid 0'ing two areca's or something) that will exceed 1GB/sec. I know of no single raid controller (by anyone). It might be possible on the newer adaptek/areca controllers with SAS disks but they are slower than the old generation with SATA disks. I haven't seen anyone try though so I don't know for sure.
 
Yeah, it is a little odd now that you mention it. But the text does say Areca on the two disks though, so maybe it is possible. Or, as you said, maybe the program just measures it incorrectly. That seems kind of odd too though, how hard can it be to make a program properly measure the read-speed of HDDs?
 
Yeah, it is a little odd now that you mention it. But the text does say Areca on the two disks though, so maybe it is possible. Or, as you said, maybe the program just measures it incorrectly. That seems kind of odd too though, how hard can it be to make a program properly measure the read-speed of HDDs?

Its definitely incorrect. The above screenshot is actually from my system back when I was running 12x1TB drives in raid5. Hdtune says 780MB/sec read, hdtach says 830MB/sec and dd on linux says 790-800 MB/sec. They are all pretty close to each other so I am pretty sure its ~800 MB/sec.

I don't see how sandra always says 1.4GB/sec when even if it was reading from cache (on the raid controller) i don't think it can do 1.4 GB/sec. Doing writes to cache seemed more to be in 1-1.1GB/sec range when I tested it before which is about what the burst is said to be on hdtach.
 
Back
Top