OCZ Vertex 2 60gb [redone]

[skubany]

system: M2N32-SLI Dlx, WinXP SP2 64bit
benchmark software: HDTune 4.50, test block size 64KB (default)

firmware: OCZ Vertex 2 60gb 1.24
diskpart.exe: clean, align=64kb, active

I've done read tests first, followed by write tests and at the end I've done an actual file transfer test finishing off with Radom Access read test. I was switching the SSD between On-Board and Areca after each individual test (ex: Random Access read test)

----

On-Board SATA II:
Sequential read: 177-222MB/s, 213MB/s avg, 208MB/s burst, smooth graph
- note: the 177MB transfer occurs over the part of the SSD which I've done a write test on prior to starting this series of tests
Random Access read: 512b 16256 IOPS, 4kb 13174 IOPS, 64kb 3301 IOPS, 1mb 234 IOPS, Random size 458 IOPS
Random Access read after write test: 512b 7216 IOPS, 4kb 6662 IOPS, 64kb 2613 IOPS, 1mb 208 IOPS, Random size 408 IOPS

Sequential write: 113-194MB/s, first 24gb therefore no avg and no burst, jagged graph but horizontal looking
- note: the 113MB transfer occurs over the part of the SSD which I've dona write test on
Random Access write: 512b 11266 IOPS, 4kb 9430 IOPS, 64kb 2557 IOPS, 1mb 178 IOPS, Random size 346 IOPS

After all sequential write tests, final test:
Used SSD state file transfer test, speeds reported by "disk monitor" function of HDTune: write 97MB/s, read 125MB/s
Random Access read: 512b 5605 IOPS, 4kb 5193 IOPS, 64k 2297 IOPS, 1mb 196 IOPS, Random size 384 IOPS

---------

Areca ARC-1220 hardware RAID controller on PCI-E 8x bus:
Sequential read: 246-268MB/s, 259MB/s avg, 542MB/s burst (this is the controller's on-board cache speed I think), jagged but with same peeks and bottoms throughout horizontal line graph
- note: interesting thing is that the graph is pretty much a horizontal line without the drop in read speed over the previously write tested portion of the SSD as was present in the On-Board test (the 177MB/s min read)
Random Access read: 512b 11400 IOPS, 4k 9593 IOPS, 64k 2528 IOPS, 1mb 240 IOPS, Random size 455 IOPS
Random Access read after write test: similar to On-Board after write test

Sequential write: 223MB/s, first 36gb therefore no avg and no burst, very slightly jagged very nice overall graph much better than the On-Board version
Random Access write: 512b 23906 IOPS, 4kb 19778 IOPS, 64kb 4871 IOPS, 1mb 258 IOPS, Random size 476 IOPS
- note: don't think these write access speeds are reliable, they might be for the controller's cache not the drive itself

After all sequential write tests, final test:
Used SSD state file transfer, speeds reported by "disk monitor" function of HDTune: write 97MB/s, read 125MB/s
Random Access read: similar to On-Board

Is this about right? Or did I get some subpar part?

thanks

 

[lmccrary]

I wouldn't get hung up on test numbers. I have the same drive and it exceeded every expectation. Just enjoy it.

 

[skubany]

I've redone the post...
lmccrary you are right, I'm just curious as to what others benched as the numbers are quite off from what OCZ markets.

I do own a Vertex 1 30GB (the original working [no stutter] SSD) in my server and I'm very happy with how fast the OS loads and how fast the applications open up.

Having said that, from my tests it appears that WD Raptors (I have 2nd gen, 150GB ones) in RAID 0 are the way to go for sequential access compared to one SSD. File read/write was faster with 2x Raptors in RAID 0 than Vertex 2, but they are light years behind the SSD on random access.

I will set up my system with OS/apps on the SSD and games on Raptors, not sure yet if I'll make them striped or mirrored.

Further opinions welcome.

 

[TC91]

Yeah those numbers are pretty normal. OCZ uses ATTO for their numbers so maybe run that if you want to see the numbers they are advertising.

 

[Zap]

did I get some subpar part?

Yeah it is subpar. :twisted:

No, seriously though, just stop running benchmarks and just go out and... enjoy using your computer. Oh yeah, I also do not recommend you check your PSU voltages or your CPU temperatures either, because we got enough threads going "OMG my +12v is at +11.89V I'm doomed BrandX suxxors I'm gonna RMA." :twisted:

I'm joking... or am I? :sneaky:

See what happens when you troubleshoot something that isn't broken? "My CPU idles in the 30°C range and must be burning up! OMG stock heatsink must be bad!"

 

[skubany]

Well, we ask to get points of reference. How else will we know if value X is correct or something we should be worried about.

Luckily my PSU voltages and CPU temperature are all in order Although these are different, comparing to hard drive benchmarks, as the voltage and temperature ranges are stated by the manufacturer/specifications group. So, as long as my values fall within the range then no worries.

With hard drives we only get the upper bounds and we need to reach out to others to get the transfer rates that we can realistically expect.

 

[Zap]

LOL, fair enough. Your numbers seem fine.