| Intel 311 Solid State Drive SSDSA2VP020G2E |  |
| Reviews - Featured Reviews: Storage | |
| Written by Olin Coles | |
| Tuesday, 10 May 2011 | |
|
Page 8 of 11
Iometer IOPS PerformanceIometer is an I/O subsystem measurement and characterization tool for single and clustered systems. Iometer does for a computer's I/O subsystem what a dynamometer does for an engine: it measures performance under a controlled load. Iometer was originally developed by the Intel Corporation and formerly known as "Galileo". Intel has discontinued work on Iometer, and has gifted it to the Open Source Development Lab (OSDL). There is currently a new version of Iometer in beta form, which adds several new test dimensions for SSDs. Iometer is both a workload generator (that is, it performs I/O operations in order to stress the system) and a measurement tool (that is, it examines and records the performance of its I/O operations and their impact on the system). It can be configured to emulate the disk or network I/O load of any program or benchmark, or can be used to generate entirely synthetic I/O loads. It can generate and measure loads on single or multiple (networked) systems. To measure random I/O response time as well as total I/O's per second, Iometer is set to use 4KB file size chunks over a 100% random sequential distribution at a queue depth of 32 outstanding I/O's per target. The tests are given a 50% read and 50% write distribution. While this pattern may not match traditional 'server' or 'workstation' profiles, it illustrates a single point of reference relative to our product field. The chart below illustrates combined random read and write IOPS over a 120-second Iometer test phase, where highest I/O total is preferred:
In our Iometer tests, which use 32 outstanding I/O's per target and a random 50/50 read/write distribution, SandForce SSDs clearly outperform the competition when tested which a larger queue depth. The latest SATA 6Gb/s storage solutions lead the pack, but Intel's Lasron Creek 311 Series SSD follows closely behind thanks to its SLC NAND flash components. In our next section, we test linear read and write bandwidth performance and compare its speed against several other top storage products using EVEREST Disk Benchmark. Benchmark Reviews feels that linear tests are excellent for rating SSDs, however HDDs are put at a disadvantage with these tests whenever capacity is high.
|
|
Comments
This doesn't really hold true, in my opinion. Valid competitors should only be those available at $120 or less, since Larson Creek (shortened "LC" below) is designed for low price. That leaves us with the Lite-On SSD, which doesn't appear in many tests, and the HDD based storages.
____________________ _______
Now, if instead of looking at it as a stand-alone unit I prefer validating it as used for SRT.
- Due to the nature of SRT a minimum requirement to make SRT faster than using the HDD alone is that the SSD used is considerably faster than the HDD in reading data, and *at least* as fast in writing data.
- SRT is also supposed to be a way to improve performance at a low cost, so any SSD used should be very affordable to purchase. Total price is far more relevant than the price/storage ratio.
SSDs based on MLC follow the rule that "more is merrier", meaning that the bigger the storage the faster it becomes. (As I've understood it the theoretical bandwidth is somewhat proportional to the number of storage chips, since they to some extent work in parallel.) Therefore low capacity MLC SSDs are not only cheap, but also very slow. Depending on many factors the write speed might drop to, and possibly below, that of HDDs.
SLC, as used in LC, is considerably more expensive per storage volume, but (as shown in the tests) doesn't have it's speed depend nearly as much on the storage space compared to MLC.
I think that keeping the write speed up was the main reason Intel choose to use SLC instead of MLC for this SSD.
It would be very interesting to see performance comparisons, both as stand-alone and in use with SRT, between LC and cheap MLC drives in the about $100 and 32-64GB region. Will LC outperform a cheaper and bigger but slower(?) MLC drive in SRT use?
As I noted in the article, it's difficult to reliably and repeatedly benchmark caching systems, especially "smart" ones whose internal algorithms are unknown. SImply running the benchmarks in a different order will affect the results.
Not only do several SSD tests where different storage size drives of same model show this, but also the somewhat related RAM on nVidia's graphics cards. Notice the different RAM bus width between GTX460 in the 768 vs 1024MB versions? An added RAM/NAND chip makes for more access.
As for Olle's remark about fairly comparing to price-similar competitors, he's dismissed that there are several SSDs that offer a better price per gigabyte than the Intel 311. In fact, the Patriot Torqx 2 (which was included in my results) offers a 32GB version for $90.
I'm well aware that there are plenty of MLC SSDs that offer better price per GB. Most of them are also considerably more than $100 in price.
The Torqx 2 isn't listed on page 4, so it's unclear to me (and all other readers) if it's the 32GB version that's been tested, or if it's another version with different performance.
HM67 and QM67 are notebook chipsets. Or did you mean to write H67 and Q67?