DRAMeXchange : Weekly Research : special report


【special report】How the SanDisk SSD SATA 5000 2.5" Boosts the PC Performance

How the SanDisk SSD SATA 5000 2.5" Boosts the PC Performance

Ever since the solid state disk (SSD) was officially unveiled, people have held a hopeful and curious outlook towards the use of DRAM or Flash chips as a new solution to the storage of massive data. Yet, at the same time, there is also a lingering doubt of whether these novel memory products can actually become mainstream.

Amid the declining NAND Flash prices, manufacturers are now more capable of manufacturing the SSD in a more cost effective way. Although still expensive, retail prices have dropped to a level that is at least more acceptable to consumers now. Afraid of losing their competitiveness, Hard disk drive (HDD) manufacturers have also devoted substantial resources to the development of SSDs. At the moment, they enjoy similar competitive advantages with the Flash product manufacturers, but the latter does enjoy a technical edge over the former.

Just how fast are the SSDs? Are they really capable of significantly boosting the PC performance, in contrast to the HDD? Can it help notebook computers (NBs) conserve more power and improve the battery lifetime?

Although no definitive answer can be made now, DRAMeXchange recently conducted benchmark tests on SanDisk's new SSD 5000 2.5" SATA device, hoping to shed some light on these questions.

Testing Platform

Taking into account the SATA port interface of the SSD 5000, along with the relevant system performance tests, the Acer PC AcerPower M8 was used as the main testing platform.

Specifications are shown below:

CPU: AMD Athlon 64 X2 dual-core processor 4200+
Memory: 2GB Kingston DDRII 667 Mhz
Motherboard: Acer EM61SM/EM61PM BIOS Version R01-A2
Chipset: NVIDIA nForce 405
GPU: NVIDIA GeForce 6100
HDD controller: NVIDIA MCP61 Serial ATA Controller
OS: Microsoft Windows XP SP2 Version 5.1.2600
HDD: Seagate Barracuda 7200.9 ST3250824AS 250GB 7200 RPM SATA 3.0Gb/s
SSD: SanDisk SSD SATA 5000 2.5-inch, 32GB

Testing Software

H2Bench 3.6
PCMARK05 Test Suite
Iometer
Bootvis
Windows XP SP2
Microsoft Vista Windows Experience Index (WEI)

Test Report

SanDisk SSD 5000 2.5-inch is targeted at enterprise notebooks that are space-constrained, and therefore cannot support a 3.5-inch HDD such as the cutting-edge Seagate model, running at 7200 RPM, used in the Acer PC that we used for our testing platform. The larger 3.5-inch form factor achieves higher performance.
It is important to note that if testing had been performed on notebooks using 1.8-inch or 2.5-inch hard drives, the performance benefits of SanDisk SSD would have been even greater than measured in this report.

Windows XP Startup Time

Currently, a large majority of the world's computers run the Windows XP operating system. However, they are gradually being replaced by Microsoft's latest Vista operating system. One of Vista's most special features is its capability of expanding the system's memory cache by configuring the Flash memory inside the computer. According to a test conducted by SanDisk, it took the SSD 44.07 seconds in opening an Office file on a Vista-running Dell Latitude D420 Notebook, which is in contrast to the 1 min and 12.08 seconds for an average 1.8-inch HDD.

During our own benchmark tests, we installed the Windows XP SP2 on an HDD and an SSD in testing the startup time. The test was performed ten times in order to calculate an average. Results indicate it respectively took the SanDisk SSD 5000 and Seagate ST3250824AS 7.91 and 12.23 seconds in completing the boot-up process, where the SSD was 54.61% faster than the HDD. In addition, the SSD can also access files more rapidly. The performance upgrades should be even more notable on the Vista platform when measuring a 2.5-inch HDD versus the 2.5-inch SanDisk SSD.


Microsoft Vista Windows Experience Index (WEI)

The Vista Windows Experience Index, a benchmark test developed by Microsoft, can be used in evaluating the efficiency of Vista on different computer systems. However, during our analysis, only the data transfer efficiency was studied. Results show the SanDisk SSD 5000 scored 5.8, slightly higher than SanDisk's own 5.7. A probable explanation for the slightly better performance was our testing platform appeared to own a good implementation efficiency. As for the Seagate ST3250824AS 3.5-inch HDD, it earned 5.4 points. When the tests were conducted by SanDisk on a notebook that employed a 2.5-inch or 1.8-inch hard drive, the benefits of SanDisk SSD became more apparent: the HDD was rated at less than 5.0 points for the 2.5-inch form factor, and less than 4.0 for the 1.8-inch  HDD.


H2Bench 3.6

The H2Bench 3.6 was employed to test the SSD's reading and writing performance.


Iometer

The Iometer was used in evaluating the read and write performance of the SSD. The 5000 SanDisk SSD maintained a read rate of 65.89MB/s, similar to the 67MB/s figure tested by SanDisk. On different platforms, the numbers were sometimes even faster. The SanDisk SSD 5000 write speed also reached at least 46.86/s. As for the average read response time, it was roughly 0.9479 milliseconds (ms), while the average writing response time was about 1.3324 milliseconds (ms). The average IO reading speed reached 1054.24 times/s, and the average IO write number hit 750.2 times/s. In general, this is a satisfactory performance.


Bootvis

BootVis was invented by Microsoft in boosting the boot time on the Windows XP operating system. It is used to track and analyze the computer boot up process and perform further optimizations when possible. The core program loading time can be seen in the table below. Our results showed only a minor advantage for SanDisk SSD. But measuring the boot performance of a 2.5-inch hard drive against a 2.5-inch SSD substantially boosts this advantage, according to SanDisk: such a hard drive boots in the range of 21 to 26 seconds versus SanDisk SSD boot time of 8 seconds.


PCMARK05 Test

The PCMARK 05 was used at the end in testing the overall disk drive performance.


By comparing the SanDisk SSD 5000 with the HDD in the PCMARK 05 test, the two were able to respectively achieve a virus scan speed of 64.304MB/s and 59.93MB/s. However, in terms of the write speed, the SSD was slower than the HDD, attaining only 23.756MB/s, which can be attributed to the slower write speeds of the Flash chip. However, when compared with a 5,400 rpm 2.5-inch HDD, the difference is not as evident. It should be noted that the PCMARK 05 has a different criterion from the H2Bench 3.6 in gauging the read speed. As the H2Bench directly controls the HDD, the figures from the H2Bench are higher than the PCMARK 05.

However, when the H2 Bench is used to assess the SSD and HDD, the difference is less notable. As for the XP Start Up test, the SSD is three times more efficient than the HDD. Furthermore, with the transfer speed of a 5400 rpm 2.5-inch hard disk reaching only 5.9MB/s-.4MB/s, the SSD is 5 times faster than the HDD. In terms of the application software's reading speed, the efficiency of the SSD is 4.46 times better than the HDD, a demonstration that the SSD's read speed is quite efficient. This is due to the fact that an SSD has almost "zero"access time since it requires neither seek nor latency time like the hard drive. This improves the input/output per second (IOPS) rate to enable much faster random read performance.


SSD Power Consumption

As the SSD conservers more power than the spindle motor of a small HDD, its read speed is faster, which thus reduces the data search time. In our notebook tests, multiple application software were run. Results revealed that the battery lifetime of the HDD was roughly 2 hours 35 minutes, while SanDisk SSD reached 2 hours and 52 minutes, more than 10%.

Conclusion: All in all, the SSD performs outstandingly, where it is capable of significantly improving the program execution efficiency. Will first be employed in high-end NB PC.

Despite the various NAND Flash advantages, people may still remain doubtful on the SSD lifetime. This stems from the fact that data losses may begin to occur after the chip is accessed roughly hundreds of  thousand ~1 million times. SSD manufacturers have developed algorithms in balancing the number of read/write times for each block of a chip. According to SanDisk's product info, the SanDisk SSD 5000 has undergone an Error correcting algorithm in ensuring the data integrity and security. In addition, they are also equipped with SanDisk patented technology for flash management. Other features of SanDisk SSD include continued operation even in the absence of power, high durability, along with a two million hour mean time to failure (MTTF) rate.

Flash memory-based SSDs can conserve more power than the DRAM-based memory products. By utilizing merely 0.9 watts, it releases less heat. When compared with the HDD, the heat release is even lower. In our SanDisk SSD tests, there is no heat problem. With no heat, there is no requirement for a fan in cooling the device. This allows the DT or NB to be both quieter and cooler. Without fans, it reduces the amount of noise to the SSD, which is conducive to both the product and environment.

SanDisk SSD 5000 is the company's 5th generation SSD product. Its high quality and small weight allows it to be a suitable choice in the mid to low capacity HDD market. Despite being more expensive, notebook manufacturers such as Dell  are already promoting SSDs by unveiling products that are more power-saving and sport longer battery lifetimes. This should help increase the overall popularity of SSDs among consumers. (http://www.dell.com/content/topics/global.aspx/corp/pressoffice/en/2007/2007_04_24_rr_000?c=us&l=en&s=corp)

 



About DRAMeXchange

DRAMeXchange is a global primary provider of future intelligences, in-depth analysis reports and advisory services on DRAM and Flash memory industry with coverage including current business, spot trading prices, and market trends, capital spending and wafer capacity trends, the impact of DRAM/flash memory products on the market, and other relevant PC industry information.

©2000-2007 DRAMeXchange ® Tech.Inc. All rights reserved.