The Samsung 470 Series Solid State Drive is this company’s first entry into the consumer SSD market. As a producer of many of the components used in other SSDs, as well as providing OEM SSD solutions for various computer manufacturers, we were surprised that they haven’t entered this market earlier, having introduced this drive in November 2010. The Samsung 470 Series SSD comes in three capacities, 64 GB, 128 GB and 256 GB. What follows is our review of the 256 GB model.
Initial Impression
The 470 Series comes in a clear plastic container, the drive contained in a foam enclosure. The drive itself looks stylish, the top being brushed metal with raised SAMSUNG lettering, and orange plastic in the corner of the drive indicating the capacity. The insert lists several features of the drive, such as Samsung 32 nm MLC NAND Flash Memory, Samsung S1MAX SSD Controller, 470 MB/s Read + Write Speed, and a 2.5 inch form factor with SATA II Interface. The 470 MB/s statement was also in larger lettering, with smaller lettering detailing that this was arrived at by combining a Read speed of 250 MB/s and a Write speed of 220 MB/s, which we think is somewhat misleading.
Samsung 470 Series 256 GB SSD
MLC is Multi Level Cell, which is one type of DRAM used in SSDs, the other being SLC, or Single Level Cell. MLC is typically lower performing, but less expensive, than SLC. The 32 nm is the memory technology used in this drive, and it was state of the art when the drive was introduced. The smaller this number, the more memory can fit in a space. The latest reported technology is a 19 nm process from Toshiba and Sandisk.
When we removed the drive from the packaging, at first we thought we had received a mockup, because the drive was so light compared to what we’re used to with a 2.5 inch drive. Whereas the Hitachi drive we’ll be benchmarking against weighs about 4 ounces, the 470 series weighs in at a mere 2.4 ounces. The bottom of the unit, which has screw holes in standard locations on both the side and bottom of the drive, but appears to be plastic, raised concerns about durability and the possibility of stripping it if you’re a bit too enthusiastic with your screwdriver.
Testing Methodology
This drive was tested in a MacBook Pro (Early 2008) with 6 GB of RAM, running Mac OS X 10.7.2, and a Mac mini (Mid 2010) with 8 GB of RAM running Mac OS X 10.6.8. Note that the SATA bus on the MacBook Pro is limited to 1.5 Gb/s, or 192 MB/s, whereas the SATA bus on the Mac mini can achieve 3.0 Gb/s, or 384 MB/s. The rotational drive in the MacBook Pro is a Hitachi HTS725050A9A364 and the rotational drive in the Mac mini is a Toshiba MK3255GSXF.
Our first measure of performance is boot time, which is the time from the boot chime (on the Mac mini) or spinning progress wheel (on the MacBook Pro) to the time the desktop is presented, and a drive read activity indicator indicates no activity. This is a good test for how the drive deals with a large number of small files.
Our second measure of performance is the transfer a single large file between the SSD and another drive on the system. On both the MacBook Pro and Mac mini, we transferred files with an Iomega eGO drive connected via FireWire 800.
Our third measure of performance is synthetic benchmark, using Drive Genius 3.1 from Prosoft Engineering. It performs sustained read, sustained write, random read and random write tests, using a block size ranging from 32K to 16M in size. We also enabled TRIM mode on our Lion machine, to see if we could get increased write performance. TRIM is a method of making sure SSD memory cells are clean before they are written to, resulting in maximum throughput.
Benchmarks
For all tests, items in bold are better.
Startup Time (seconds)
MacBook Pro Rotational | 175 |
MacBook Pro SSD | 32 |
Mac mini Rotational | 154 |
Mac mini SSD | 22 |
Large File Transfer (seconds)
MacBook Pro Rotational to FW | 55 |
MacBook Pro Rotational from FW | 47 |
MacBook Pro SSD to FW | 39 |
MacBook Pro SSD from FW | 38 |
Mac mini Rotational to FW | 52 |
Mac mini Rotational from FW | 56 |
Mac mini SSD to FW | 29 |
Mac mini SSD from FW | 33 |
Battery Life (minutes)
MacBook Pro SSD | 196 |
MacBook Pro Rotational | 166 |
MacBook Pro - Random Tests (MB/s)
Block Size | Write SSD | Write SSD TRIM | Write Rotational | Read SSD | Read Rotational |
32K | 61 | 83 | 58 | 26 | 1 |
64K | 79 | 97 | 76 | 61 | 3 |
128K | 93 | 106 | 94 | 79 | 6 |
256K | 106 | 119 | 100 | 86 | 12 |
512K | 117 | 122 | 118 | 115 | 22 |
1M | 121 | 125 | 120 | 122 | 32 |
2M | 119 | 122 | 118 | 120 | 47 |
4M | 117 | 123 | 67 | 120 | 55 |
8M | 50 | 51 | 83 | 125 | 73 |
16M | 122 | 121 | 95 | 126 | 79 |
MacBook Pro - Sustained Tests (MB/s)
Block Size | Write SSD | Write SSD Trim | Write Rotational | Read SSD | Read Rotational |
32K | 36 | 57 | 56 | 60 | 58 |
64K | 56 | 76 | 77 | 84 | 78 |
128K | 88 | 103 | 92 | 104 | 73 |
256K | 99 | 117 | 106 | 116 | 67 |
512K | 118 | 125 | 116 | 125 | 55 |
1M | 120 | 127 | 120 | 127 | 85 |
2M | 119 | 126 | 63 | 126 | 88 |
4M | 120 | 123 | 99 | 126 | 71 |
8M | 119 | 125 | 99 | 126 | 60 |
16M | 120 | 127 | 79 | 127 | 95 |
Mac mini - Random Tests (MB/s)
Block Size | Write SSD | Write Rotational | Read SSD | Read Rotational |
32K | 104 | 21 | 13 | 1 |
64K | 129 | 5 | 21 | 3 |
128K | 161 | 73 | 23 | 5 |
256K | 181 | 86 | 79 | 10 |
512K | 195 | 22 | 146 | 16 |
1M | 202 | 99 | 232 | 24 |
2M | 202 | 99 | 230 | 27 |
4M | 201 | 80 | 240 | 34 |
8M | 61 | 42 | 245 | 31 |
16M | 95 | 46 | 249 | 38 |
Mac Mini - Sustained Tests (MB/s)
Block Size | Write SSD | Write Rotational | Read SSD | Read Rotational |
32K | 73 | 47 | 137 | 45 |
64K | 107 | 66 | 195 | 41 |
128K | 142 | 80 | 218 | 39 |
256K | 175 | 91 | 219 | 45 |
512K | 193 | 99 | 240 | 30 |
1M | 202 | 101 | 251 | 45 |
2M | 201 | 101 | 249 | 45 |
4M | 203 | 50 | 252 | 52 |
8M | 203 | 42 | 250 | 53 |
16M | 203 | 49 | 251 | 37 |
Discussion
The startup time for both systems is a clear indication of the advantage of these drives when reading a large number of relatively small files. The time went from hundreds of seconds to tens of seconds. Transferring a single large file showed a less dramatic benefit, but in all cases, the SSD was faster than the rotational drive.
For the synthetic benchmarks on the MacBook Pro, the tests showed little difference in write performance between the SSD and our rotational drive, until the block size got beyond 4 MB, whereas the read speeds showed a clear advantage from the start. The maximum speed achieved of 127 MB/s when writing is quite a bit below the theoretical maximum of 192 MB/s for the SATA I bus in this machine. We also noticed an odd reduction in performance with an 8 MB block size write performance, but we’ll attribute this to shortcomings in the SATA implementation in this machine, since we didn’t see this on the mini. Enabling TRIM showed a measurable increase in write performance for small block sizes on the benchmarks, but we didn’t find it significant in day-to-day use.
For the synthetic benchmarks on the Mac mini, we saw more dramatic results, partly due to the SATA II bus on this machine, allowing us to achieve full bandwidth, and partly due to the Toshiba drive in this machine, which we found would only negotiate a SATA I speed (1.5 Gb/s) with the mini. The drive nearly reached its advertised maximum write speed of 220 MB/s, reaching 203 MB/s in our benchmark, and achieved the advertised maximum read speed of 250 MB/s, both when test block sizes were 1 MB or greater.
Conclusion
The Samsung 470 Series SSD shows a clear performance advantage on our Macs, even one that don’t have the latest SATA II bus supported by this SSD. Due to the relatively low power draw when compared to our rotational hard drive, we noted a significant increase in battery life, always a concern for those on a portable computer. The only downside is the price; for the same capacity, one could get a rotational hard drive for $100 or less. However, the pricing for this SSD is comparable to what we found for other of similar capacity. For those that can afford it, the Samsung 470 256 GB SSD is a worthy upgrade.