Forum Home
Press F1
 
Thread ID: 128515 2012-12-24 10:24:00 Upgrade to SSD bk T (215) Press F1
Post ID Timestamp Content User
1320107 2012-12-27 22:00:00 Nah they don't always do the "max" speed, it's more of a "peak" rather than a "get this all the time" Chilling_Silence (9)
1320108 2012-12-28 00:17:00 So, you reckon that benchmark test is acceptable?

But the difference so much from what they claim, i.e. "Seq Performance (MB/sec) Up to 550 / 520" ; only achieve ~60%. 		bk T (215)
1320109 2012-12-28 00:51:00 I dunno, I reckon you'll never get the full amount, so just how much of a hit on that you want to accept is up to you. Chilling_Silence (9)
1320110 2012-12-28 02:52:00 Looks like a defective disk, (to me).

Is there any experts on this issue in this forum? 		bk T (215)
1320111 2012-12-28 03:57:00 The "up to" statement means just that, to test you'd have to find out the exact set of conditions and test used to determine the numbers. I tried my samsung 830 out of interest and it's in between your two results but I have no Idea if crystalmark is a reliable program to test with.
Samsung 830 256GB
-----------------------------------------------------------------------
CrystalDiskMark 3.0.2 Shizuku Edition (C) 2007-2012 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 361.911 MB/s
Sequential Write : 363.626 MB/s
Random Read 512KB : 284.484 MB/s
Random Write 512KB : 343.233 MB/s
Random Read 4KB (QD=1) : 20.878 MB/s [ 5097.1 IOPS]
Random Write 4KB (QD=1) : 56.450 MB/s [ 13781.7 IOPS]
Random Read 4KB (QD=32) : 188.466 MB/s [ 46012.2 IOPS]
Random Write 4KB (QD=32) : 128.151 MB/s [ 31286.8 IOPS]

Test : 1000 MB [C: 75.8% (180.6/238.4 GB)] (x5)
Date : 2012/12/28 16:33:26
OS : Windows 7 Home Premium Edition SP1 [6.1 Build 7601] (x64)

And just out of curiosity, an intel 320 series 120GB in a cheap USB 3.0 enclosure
-----------------------------------------------------------------------
CrystalDiskMark 3.0.2 Shizuku Edition (C) 2007-2012 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 216.201 MB/s
Sequential Write : 137.952 MB/s
Random Read 512KB : 93.747 MB/s
Random Write 512KB : 139.971 MB/s
Random Read 4KB (QD=1) : 17.473 MB/s [ 4266.0 IOPS]
Random Write 4KB (QD=1) : 40.280 MB/s [ 9833.9 IOPS]
Random Read 4KB (QD=32) : 18.763 MB/s [ 4580.8 IOPS]
Random Write 4KB (QD=32) : 60.400 MB/s [ 14746.1 IOPS]

Test : 1000 MB [J: 95.2% (106.4/111.8 GB)] (x5)
Date : 2012/12/28 16:44:35
OS : Windows 7 Home Premium Edition SP1 [6.1 Build 7601] (x64)

The intel looks much slower on paper, in practice though as an OS disk the two feel pretty much the same. Benchmarks aren't everything. 		dugimodo (138)
1320112 2012-12-29 00:42:00 Did some research and found that the issue could be the Marvell SATA 6.0 Gb/s controller which apparently doesn't support the full speed of SSD drives. Why Marvell couldn't resolve this with a firmware/driver to address this issue and why SSD manufacturers don't disclose this to the public?

Or, there is a fix available already?

Cheers 		bk T (215)
1320113 2012-12-29 20:36:00 dugimoto, you have a Marvell 9128 SATA III 6.0 Gb/s disk controller too? If your MB has a Intel SATA3 , try on the Intel one and see any improvement? bk T (215)
1320114 2012-12-29 21:07:00 Might do not sure, there are 4 SATA 3 ports and I think the chipset only has 2.
However I don't want to mess with it, I have 7 hard drives and rearranging them screws things up. 		dugimodo (138)
1320115 2012-12-29 22:58:00 Your motherboard user manual should tell you that.

Don't try it if you're not confident in doing so. 		bk T (215)
1 2