samsung 830 256gb benchmarks way off

[ldefeo1]

hi guys got a 830 which seems great so far though in xbench my scores are really low.

im on a 2010 mbp so only sata2 its in the superdrive slot if that makes a difference.

here are the scores from another forum,

Disk Test 573.84
Sequential 398.50
Uncached Write 583.13 358.03 MB/sec [4K blocks]
Uncached Write 423.34 239.53 MB/sec [256K blocks]
Uncached Read 209.61 61.34 MB/sec [4K blocks]
Uncached Read 840.50 422.43 MB/sec [256K blocks]
Random 1024.67
Uncached Write 642.49 68.01 MB/sec [4K blocks]
Uncached Write 865.23 276.99 MB/sec [256K blocks]
Uncached Read 2897.09 20.53 MB/sec [4K blocks]
Uncached Read 1181.60 219.25 MB/sec [256K blocks]

and here are my scores

Sequential 199.65
Uncached Write 260.80 160.12 MB/sec [4K blocks]
Uncached Write 289.06 163.55 MB/sec [256K blocks]
Uncached Read 98.86 28.93 MB/sec [4K blocks]
Uncached Read 380.8 191.39 MB/sec [256K blocks]
Random 744.12—
Uncached Write 771.42—> 81.66 MB/sec [4K blocks]
Uncached Write 493.61—> 158.02 MB/sec [256K blocks]
Uncached Read 1497.45—> 10.61 MB/sec [4K blocks]
Uncached Read 721.77—> 133.93 MB/sec [256K blocks]

i have done about 4 diferent benchmarks with results fluctuating a fair bit, i have also enabled trim which made a small improvement. just a side note is trim a good idea for this drive?

whats going on, i dont see how this is a sata2 issue as i no where near saturating the link. im getting roughly half what the other guy got and no where near what mr. anand got

 

[cmdrdredd]

Use as ssd or crystal disk mark and observe your scores vs those online. I have sata 2 and a crucial m4 and I score faster writes than online scores but my reads are only half that of sata 3 which is still more than 2x faster than my raptor.

 

[Burner27]

Maybe the 'other guy' has it on a SATA 3 connection. That might explain his numbers compared to yours.

 

[ldefeo1]

the benchmark i am comparing to is done in xbench too so i thought this would be a fair test.

im not concerned about the high sequential read speeds that he gets over 400. i knew i wouldnt go more than 275mb/s on any test

its the fact im not getting the same performance when im not even close to saturating the sata2 port. eg my 256k random read and write is half of his, why would this be the case. my sata 2 port is capable of delivering ~250mb/s (what he got) but i didnt come close (~130mb/s).

 

[corkyg]

Does the CPU speed have any bearing on this?

 

[ldefeo1]

good question, im no expert on this but id assume the xbench disk test would isolate the disks performance from the rest of the system, maybe the ram clock would affect it somewhat? still 2x is an enormous difference, i was wondering if this might be a firmware/trim issue?

 

[cmdrdredd]

Not trim, firmware maybe. When I get home I'll post screenshots of my ssd performance with my system for you. Even on sata2 you should be getting 250MB/s reads.

 

[Rubycon]

Does the CPU speed have any bearing on this?

On most benchmarks not only CPU but memory latency and throughput will have effects due to the complex way the OS handles iops.

If you throw a hardware controller into the mix it gets even more complicated!

That said OP if you can post AS SSD, CDM and Anvil SSD benchmark results if you could.

 

[cmdrdredd]

I don't know if you can run AS SSD, CDM, or Anvil on your macbook. You would need to boot to windows and then I don't know the effect that will have on the benchmark results. Anyway here's mine.

AS SSD

Crystal Disk Mark

Anvil Pro

 

[Boomhowler]

I would personally not really trust the results from xbench, it has been a bit unreliable from time to time.

Use AJA System Test instead (free from: http://www.aja.com/en/products/software/) and run a "sweep binary frame sizes" to get a better, more consistent, read/write test.

 

[ldefeo1]

thanks might try that,

it seems Rubycon was right looking at this:

http://www.anandtech.com/show/4205/the-macbook-pro-review-13-and-15-inch-2011-brings-sandy-bridge/7

shows the 2011 mbp (sata 3) improving performance considerably, even within the sata II ~275mb/s threshold.

it must be due to cpu/memory clock also. disapointing i guess my simplified view of performance for random test should be equal for a sata II/III port arent entirely true.

anand even said

"You'll note that the absolute numbers are pretty low to begin with. A 128KB sequential read of the Intel SSD 510 on our desktop Sandy Bridge SSD testbed pulls nearly 400MB/s. On the new MacBook Pro we can't get more than 320MB/s.

Our sequential Iometer tests are run at a queue depth of 1 so there's no advantage there. The only explanation I can come up with (assuming Xbench's test is accurate) is that Apple may be aggressively implementing SATA controller power management under OS X. Capping the link's performance or aggressively putting it to sleep could reduce performance at the benefit of increasing battery life."

so the situation is even worse for my c2d model.

that said would still like to compare my benchmarks to you guys using that ssd tool under windows, havent got a boot camp partition but would running the benchmarks under parallels be worthwhile or would the emulation make the results pointless.

 

[cmdrdredd]

That's what I don't know. Would it be fast enough to give a real gauge.

 

[Rubycon]

I have a 512GB M4 in a 17" early MBP. (Sandy Bridge 2.2GHz).
In OSX it gets 500MB/S reads and 250MB/S writes.
Not sure of any other programs like Anvil/ASSSD for Mac.
I do have a Win7 (bootcamp) partition running and the benchmarks are much lower because the drive is running in IDE mode. D: