The latest Intel SSD, the 730 series brings enterprise class hardware to the consumer market. Intel made some tweaks to its Data Center SSD, the DC S3500 and re branded it as the 730 series. Unlike the DC S3500 line, which offers 9 SSDs, sizing from 80 up to 800 GB, the 730 series is only available in two capacities, 240 and 480 GB. It is rather a smart move considering those two sizes represents the bulk of the SSD sales. It is going to be interesting to see how an business class product could met the consumer needs. Enterprise computer hardware requirements are based on 24/7 availability, performance consistency while the consumer based craves on burst performance at an affordable price.
I want to extend my thanks to Intel for loaning us two 730 480 GB. I will take that opportunity and benchmark both SSD as a RAID 0 and RAID 1 volume.
|Model Name||Intel® Solid-State Drive730 Series|
|Capacity||240GB & 480GB|
|Cache||1GB / 480GB|
|Bandwitdh||Sustained Sequential Reads / Writes|
|Single Drive||2 Drive RAID-0|
|240GB: up to 550 / 270 MB/s||240GB: up to 1020 / 530 MB/s|
|480GB: up to 550 / 470 MB/s||480GB: up to 1020 / 910 MB/s|
|Read / Write Latency||50 µs / 65 µs|
|4KB Reads / Writes|
|Random I/O per sec (IOPS)||Single Drive||2 Drive RAID-0|
|240GB: up to 86K/ 56K||240GB: up to 136K / 136K|
|480GB: up to 89K / 74K||480GB: up to 168K / 145K|
|Interface||SATA 6Gb/s, SATA 3Gb/s, 1.5Gb/s|
|2.5”||Up to 7mm / up to 78 grams|
|Life Expectancy||2 million hours (MTBF)|
|Lifetime Endurance||240GB – 50GB WpD||480GB – 70GB WpD|
|Power Consumption||12V 240GB 480GB||5V 240GB 480GB|
|Active: 3.8W 5.5W||Active: 3.8W 5.0W|
|Idle: 1.5W 1.5W||Idle: 1.2W 1.3W|
|Operating Temperature||0° C to 70° C|
|Pricing||240GB / $249||480GB / $489|
Controller has been factory overclocked from 400 to 600MHz and the NAND bus speeds was raised from 83 to 100MHz. Higher clock speeds improve internal maintenance operations such as garbage collection, CRC, TRIM, wear leveling for performance consistency purposes.
The cons for higher speed is power consumption and heat output. Although the DC S3500 was already pushing around 5W at 400Mhz. Regardless, both of those, don’t make the 730 an ideal drive storage for laptops. For reference, Samsung EVO power consumption average 0.1W vs.5.5W for the 480GB 730 and idle is about 0.045W vs 1.3W for the Intel 730.
The absence of encryption hardware is understandable, since notebooks are more prone to thefts and losses than a desktop computer. However the DC S3500 supports AES-256 hardware encryption, I do wonder if the increased clock had some “issues” with the 730 series.
Higher capacitors give the drive enough time to the cached data to commit to the NAND in case of power loss. Per Intel, the power loss protection are equivalent to the data center drives. Based on the DC S3700/3500, Intel switched from a compressed binary tree system to an 1:1 ratio indirection mapping table. Advantage, faster access to the data, trade off, the uncompressed table requires more cache storage. What is left over is used for the controller firmware code. No user data is cached.
When an SSD drive advertised an user storage of 480 GB, the actual NAND capacity on board is 512 GB. That way you can (almost) always tells if the drive is provisioned or not. For instance, the PRO does not but the EVO series does. The 730 is packed with fourteen 32 GB, one 64 GB and one 16 GB modules, totaling 528 GB of NAND storage. The 46 GB reserved storage, unavailable to the user, is dedicated to cover wear leveling, garbage collection, spare cells, and other internal maintenance. The extra 16 GB explains, in part, a higher price tag compared to other mainstream SSD products.
Although not included in the packaging, Intel provides two applications to manage the SSDs on their website, Intel Solid-State Drive Toolbox and Intel Data Migration Software.
Intel SSD Toolbox with Intel SSD Optimizer (v3.2.1).
Even though the Intel SSD 730 is not listed as a supported drive by Intel SSD toolbox, the utility recognizes the drive as a standalone but also as a RAID volume. Beside “Secure erase” and “Firmware update”, most of the features are supported in the RAID configuration. Intel SSD optimizer supports manual TRIM in a RAID configuration.
Intel Data migration Software.
In case the user does not feel like re-installing the Operating System from scratch, Data Migration Software makes the process painless. It is pretty straight forward, start the application, select the source, then the destination on the next screen and confirm. The computer will reboot, migrate the drive content from one to the other, and reboot again. At that point, make a stop in the BIOS to select the new drive to boot from. I would advise you to keep the previous drive for a little while just in case. The custom made migration tool by Acronis for Intel is only available when it detects Intel drive signature. In a RAID configuration, the cloning utility failed to detect the Intel drives as components of the RAID. It generated a BSoD, I resorted to used an Acronis CD rescue to clone the RAID volume.
I went through most of the popular benchmark tools, AS SSD, CrystalDiskMark, ATTO, IoMeter, Anvil’s Storage Utility v1.1.0 and PCMark Vantage. But I also used performance monitoring tools such as DiskMon and hIOmon, primarily to validate the tests. Instead of posting chart after chart, I believe, as a consumer, what is important is how the product fits the needs and not chasing after uber high numbers which are only attainable during benchmarking. I narrowed it down to Anvil’s Storage Utility and PC Mark Vantage Licensed Pro version.
Drive conditioning: The SSDs were prepped with Windows 7 (from an image), filled with about 120GB of data total and benchmarks were run from the tested unit acting as the OS drive.
Steady state: This state occurred overtime when the drive went through enough write cycles, or to be more specific program/erase (P/E) cycles, that write performances were consistent or stable. It may take a few weeks before the SSD reaches it, depending on the computing usage but it can be accelerated using IoMeter.
In summary, Steady State is: Written Data = User capacity x 2, at least.
Both RAID 1 and RAID 0 volumes were setup at the BIOS level, RAID (vs. AHCI), with the “Intel(R) Desktop/Workstation/Server Express Chipset SATA RAID Controller” driver. In RAID 0, the recommendation was to use 128K stripes, which favor large files size. But since I wanted to test the RAID 0 in a desktop environment, I lowered the stripe size down to 32K.
The default caching settings were left to default. “Write-cache buffer flushing” is enabled and Cache Mode: OFF. This setup prioritizes data integrity over performance. To improve performance, “Write-cache buffer flushing” is set to “disabled” and Cache mode to “Write back”. UPS is a must in this configuration.
What numbers are relevant in a real world usage?
Keep in mind that unlike synthetic benchmarks which perform only one specific operation at the time for a predetermined duration, seq read, then seq write then random read, and so on and so forth, real world usage paints a different picture. All four access types can occur at any time, and different transfer rates and different (I/O access) percentages. For instance, a storage subsystem on a streaming server would mostly see high seq read I/O, large block reads, with very little to none write. Looking at a database server without blob data type, we would probably see 75% random read, 20% random write and 5% random and seq write. I could either guesstimate the different ratios or figure a method to define a more accurate I/O usage baseline.
While it is entertaining to run a bunch of benchmarking tools, expecting huge numbers, the purpose of testing the units is to get a good look at how they perform under realistic desktop usage pattern. That is why I picked PCMark Vantage suite as my usage pattern. By capturing and analyzing I/O during the PCVM run, disk operations are breakdown to percentage read vs. write, random vs. sequential, queue depth and average file transfer size.
With that information, benchmarking makes more sense since all the numbers do not carry the same importance thus some results are more valuable than others.
In summary, I/O pattern defines what I need from the device vs. what can the device do overall.
The I/O baseline process was explained in the Intel 525 mSATA review.
From the numbers, I rated the I/O usage by activity as follow: Random Read > Random Write > Seq Read > Seq Write and average file size is 128K.
To cover Queue Depth, I used hIOmon during the PC Vantage full run. There is a trial version for a week which is enough time to build the baseline. Based on the chart below, it is obvious that a benchmark score from a QD 16 (or more) does not carry the same weight as a score from a QD 1.
Setting aside Samsung RAPID technology numbers, the Intel Random Read performance single drive is on par with the Samsung products we have tested so far. Even in both RAID setups, I only start seeing a 25% (+/-) increase once the QD is 16 and higher. Keep in mind, a RAID 0 configuration will not take full advantage of the data stripping when working with 4K data. In RAID 1, since the data is mirrored, the data can be read from either drive, in other words, two different files can be read at the same time.
Working with larger blocks, 32K and 128K, the numbers lag behing the 840 PRO and EVO, even in a RAID setup. Which is not surprising since RAID performance only kicks in when dealing with large files, the larger, the faster.
Writes I/O is comparable to the competitors. I was actually expected a lower number based on the fact that the Intel 730 does not cache user data. It appears the “overclocking” and the uncompressed indirection mapping table makes up for it, while adding write performance consistency. The performance gains only show with a RAID 0 volume, and again, with large files. Large data can be efficiently stripped across both drives while RAID 1 duplicates data.
PC Mark Vantage, trace based benchmark “productivity” and “gaming”, showed only a small performance increased from a single drive compared to a RAID setup. Looking at “HDD” benchmark, the numbers revealed a 35% increased between single drive and RAID 1, up to 50% when a single is matching with RAID 0. The numbers, from PC Mark Vantage, demonstrates that RAID 0 or RAID 1 does not just straight up “double” the system disk performance because the numbers of drives is doubled. The performance gain is only noticeable when the computing system deals with massive amount of data.
Going through the specifications alone already gave me the assumption that the product focuses around reliability and consistency.
- 70 GB host write a day.
- 20 mn NAND, Data Center grade
- Power Loss Protection
- Life expectancy is rated at 2 million compared to 1.5 million hours MTBF for the EVO, as a reference.
- 5 Years Warranty
- Uncompressed indirection mapping table.
- Higher controller and NAND clock.
- No user data is cached internally. Caching available through the Operating System.
Performance wise, the 730 480 GB, did not break any records, based on synthetic and traced benchmark, the numbers are acceptable. I doubt that the difference in the results will be noticeable in every day computing usage. How would the product fares in the consumer market? My take is, only consumers who understand and/or has a real need for the tradeoff between raw performances vs. consistency would make the investment worth their money. Applications such as photo or video editing, 3D animation rendering or running a dedicated VM host with several active VMs would benefit from the 730 Series in terms of reliability.