Thread: Samsung 32GB SATA SSD tested
For background, please read the thread "Fusion-io ioDrive", archived at http://archives.postgresql.org/pgsql-performance/2008-07/msg00010.php To recap, I tested an ioDrive versus a 6-disk RAID with pgbench on an ordinary PC. I now also have a 32GB Samsung SATA SSD, and I have tested it in the same machine with the same software and configuration. I tested it connected to the NVIDIA CK804 SATA controller on the motherboard, and as a pass-through disk on the Areca RAID controller, with write-back caching enabled. Service Time Percentile, millis R/W TPS R-O TPS 50th 80th 90th 95th RAID 182 673 18 32 42 64 Fusion 971 4792 8 9 10 11 SSD+NV 442 4399 12 18 36 43 SSD+Areca 252 5937 12 15 17 21 As you can see, there are tradeoffs. The motherboard's ports are substantially faster on the TPC-B type of workload. This little, cheap SSD achieves almost half the performance of the ioDrive (i.e. similar performance to a 50-disk SAS array.) The RAID controller does a better job on the read-only workload, surpassing the ioDrive by 20%. Strangely the RAID controller behaves badly on the TPC-B workload. It is faster than disk, but not by a lot, and it's much slower than the other flash configurations. The read/write benchmark did not vary when changing the number of clients between 1 and 8. I suspect this is some kind of problem with Areca's kernel driver or firmware. On the bright side, the Samsung+Areca configuration offers excellent service time distribution, comparable to that achieved by the ioDrive. Using the motherboard's SATA ports gave service times comparable to the disk RAID. The performance is respectable for a $400 device. You get about half the tps and half the capacity of the ioDrive, but for one fifth the price and in the much more convenient SATA form factor. Your faithful investigator, jwb
On Tue, Jul 22, 2008 at 6:04 PM, Jeffrey W. Baker <jwbaker@gmail.com> wrote: > Strangely the RAID controller behaves badly on the TPC-B workload. It > is faster than disk, but not by a lot, and it's much slower than the > other flash configurations. The read/write benchmark did not vary when > changing the number of clients between 1 and 8. I suspect this is some > kind of problem with Areca's kernel driver or firmware. Are you still using the 2.6.18 kernel for testing, or have you upgraded to something like 2.6.22. I've heard many good things about the areca driver in that kernel version. This sounds like an interesting development I'll have to keep track of. In a year or two I might be replacing 16 disk arrays with SSD drives...
On Tue, Jul 22, 2008 at 5:32 PM, Scott Marlowe <scott.marlowe@gmail.com> wrote: > On Tue, Jul 22, 2008 at 6:04 PM, Jeffrey W. Baker <jwbaker@gmail.com> wrote: > >> Strangely the RAID controller behaves badly on the TPC-B workload. It >> is faster than disk, but not by a lot, and it's much slower than the >> other flash configurations. The read/write benchmark did not vary when >> changing the number of clients between 1 and 8. I suspect this is some >> kind of problem with Areca's kernel driver or firmware. > > Are you still using the 2.6.18 kernel for testing, or have you > upgraded to something like 2.6.22. I've heard many good things about > the areca driver in that kernel version. These tests are being run with the CentOS 5 kernel, which is 2.6.18. The ioDrive driver is available for that kernel, and I want to keep the software constant to get comparable results. I put the Samsung SSD in my laptop, which is a Core 2 Duo @ 2.2GHz with ICH9 SATA port and kernel 2.6.24, and it scored about 525 on R/W pgbench. > This sounds like an interesting development I'll have to keep track > of. In a year or two I might be replacing 16 disk arrays with SSD > drives... I agree, it's definitely an exciting development. I have yet to determine whether the SSDs have good properties for production operations, but I'm learning. -jwb
On Wed, Jul 23, 2008 at 1:57 PM, Jeffrey Baker <jwbaker@gmail.com> wrote: > On Tue, Jul 22, 2008 at 5:32 PM, Scott Marlowe <scott.marlowe@gmail.com> wrote: >> On Tue, Jul 22, 2008 at 6:04 PM, Jeffrey W. Baker <jwbaker@gmail.com> wrote: >> >>> Strangely the RAID controller behaves badly on the TPC-B workload. It >>> is faster than disk, but not by a lot, and it's much slower than the >>> other flash configurations. The read/write benchmark did not vary when >>> changing the number of clients between 1 and 8. I suspect this is some >>> kind of problem with Areca's kernel driver or firmware. >> >> Are you still using the 2.6.18 kernel for testing, or have you >> upgraded to something like 2.6.22. I've heard many good things about >> the areca driver in that kernel version. > > These tests are being run with the CentOS 5 kernel, which is 2.6.18. > The ioDrive driver is available for that kernel, and I want to keep > the software constant to get comparable results. > > I put the Samsung SSD in my laptop, which is a Core 2 Duo @ 2.2GHz > with ICH9 SATA port and kernel 2.6.24, and it scored about 525 on R/W > pgbench. From what I've read the scheduler in 2.6.24 has some performance issues under pgsql. Given that the 2.6.18 kernel driver for the areca card was also mentioned as being questionable, that's the reason I'd asked about the 2.6.22 kernel, which is the one I'll be running in about a month on our big db servers. Ahh, but I won't be running on 32 Gig SATA / Flash drives. :) Wouldn't mind testing an array of 16 or so of them at once though.
Jeff, Some off topic questions: Is it possible to boot the OS from the ioDrive? If so, is the difference in boot up time noticeable? Also, how does ioDrive impact compilation time for a moderately large code base? What about application startup times? Cheers, Behrang Jeffrey Baker wrote: > > For background, please read the thread "Fusion-io ioDrive", archived at > > http://archives.postgresql.org/pgsql-performance/2008-07/msg00010.php > > To recap, I tested an ioDrive versus a 6-disk RAID with pgbench on an > ordinary PC. I now also have a 32GB Samsung SATA SSD, and I have tested > it in the same machine with the same software and configuration. I > tested it connected to the NVIDIA CK804 SATA controller on the > motherboard, and as a pass-through disk on the Areca RAID controller, > with write-back caching enabled. > > Service Time Percentile, millis > R/W TPS R-O TPS 50th 80th 90th 95th > RAID 182 673 18 32 42 64 > Fusion 971 4792 8 9 10 11 > SSD+NV 442 4399 12 18 36 43 > SSD+Areca 252 5937 12 15 17 21 > > As you can see, there are tradeoffs. The motherboard's ports are > substantially faster on the TPC-B type of workload. This little, cheap > SSD achieves almost half the performance of the ioDrive (i.e. similar > performance to a 50-disk SAS array.) The RAID controller does a better > job on the read-only workload, surpassing the ioDrive by 20%. > > Strangely the RAID controller behaves badly on the TPC-B workload. It > is faster than disk, but not by a lot, and it's much slower than the > other flash configurations. The read/write benchmark did not vary when > changing the number of clients between 1 and 8. I suspect this is some > kind of problem with Areca's kernel driver or firmware. > > On the bright side, the Samsung+Areca configuration offers excellent > service time distribution, comparable to that achieved by the ioDrive. > Using the motherboard's SATA ports gave service times comparable to the > disk RAID. > > The performance is respectable for a $400 device. You get about half > the tps and half the capacity of the ioDrive, but for one fifth the > price and in the much more convenient SATA form factor. > > Your faithful investigator, > jwb > > > -- > Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org) > To make changes to your subscription: > http://www.postgresql.org/mailpref/pgsql-performance > > -- View this message in context: http://www.nabble.com/Samsung-32GB-SATA-SSD-tested-tp18601508p19282698.html Sent from the PostgreSQL - performance mailing list archive at Nabble.com.