ServeTheHome.com

The Intel X25-V 40GB drive is well known to be one of, if not the best low-cost SSD on the market at the moment. The SSD is basically an Intel X25 G2 drive with TRIM support except with half the NAND flash memory and using only half of the SSD controller channels. The major impact is well known at this point, the Intel X25-V 40GB drives have poor sequential write performance that is approximately half that of the Intel X25-M 80GB. Of course, with two X25-V’s in RAID 0, write speed is essentially doubled and read speed goes well beyond a single drive.

Continue Reading

OCZ’s Agility 2 120GB is a new generation of drive based on the SandForce SF-1200 controller. Unlike the OCZ Vertex LE and OCZ Vertex 2 drives, the Agility 2′s SandForce controller has an IOP limit artificially limited in firmware. With that being said, it is also cheaper than the Vertex 2 120GB and is still a very fast SSD, especially with compressible loads.

The short story of the SandForce controller is that it uses an internal encryption/ compression algorithm to write data to NAND faster than other drives. The SandForce based Agility 2 120GB is therefore fairly fast with compressed workloads (i.e. pictures, music files, zip files, ISOs, and etc) and much faster with compressible workloads such as office documents. Another major storyline is that SandForce is a fairly new player and the controllers are of unknown reliability. Finally, originally SandForce reserved much more space on drives in this capacity range for use in wear leveling, cache and etc. OCZ Agility 2 and Vertex 2 SandForce drives were originally 100GB but after being criticized for mediocre cost per GB and with little performance penalty moving from 28GB reserved (old drives were 100GB usable 28GB reserved) to 8GB reserved (new drives are 120GB usable 8GB reserved), the choice to enable the extra capacity was clear.

Continue Reading

This weekend I woke up one morning and decided that I wanted to know if retail packaged drives had a lower DOA rate than OEM drives from Newegg. In all fairness, I think I was just trying to put off a 5am Saturday morning gym trip for a few hours. I ended up filtering Newegg’s hard drive category by internal drives of 1TB, 1.5TB, and 2TB in capacity. I then went through each result and recorded the quantity of reviews for each of Newegg’s awesome egg-scale along with a few other parameters. It turns out that Newegg did not have as much information on retail packaged hard drives as I had wanted, but I found some interesting results nonetheless. Anyone can do this survey, but hopefully this saves some time. For the most up-to-date information see Newegg.

Continue Reading

The Patriot PS-100 32GB continues to be a very low cost SSD after rebates. I purchased two of the drives early on and did both single drive and raid 0 benchmarks. After that I did PS-100 firmware 2.008 benchmarks. Recently, Patriot released a new firmware, v3.000 so I decided to try the flash and see if the claims of better small (4K) reads/ writes were better with the new firmware. I found that they are better, but there is a significant drawback.

Continue Reading

I got a note asking to update my original benchmark for the Patriot PS-100 32GB with results from the new Patriot PS-100 Firmware v2.008. Performance is said to be better and it is!

Update 6 June 2010: New Patriot PS-100 32GB Firmware v3.000 benchmarks are up.

The specs of the update are:

Firmware update #2.008

• Firmware v2.006 doesn’t support Active Protection feature in Lenovo laptops. Firmware 2.008 fixed it by enabling Active Protection identify bit.
• Some platforms had issue installing Windows 7 into SSD with firmware v2.006. The root cause is partial command will be NAK in background process and bank (0) is not loaded in PRAM. Firmware v2.008 has fixed this issue.
• Firmware v2.006 has power on and loading Windows XP issue with some NVIDIA chipsets. Firmware v2.008 fixed it by changing default setting to MDMA mode 2.
• In Apple Mac Book and Mac Book Pro, sometimes SSD with Firmware v2.006 will fail in reboot. Firmware v2.008 fixed this issue.

Quick Teaser: The results got better!

Continue Reading

Shortly after posting my last benchmarks on the dual OCZ Vertex 120GB in Raid 0 setup, I received an e-mail to try the new Intel Rapid Storage Technology (RST) 9.6.0.1014 drivers, which were supposed to enable TRIM in raid. These new drivers were released, taken down, then re-released in March 2010. The big improvement was supposed to be TRIM is enabled for Raid 0, 1, 10, but not enabled for Raid 5. Unfortuneately, this seems not to be working so the best that can be done is the background garbage collection in firmware v1.5 for OCZ’s Indilinx based SSDs. Since I am only working with two 120GB drives at the moment, I decided to give them a shot. Just for reference, the original benchmarks were done with the standard, off the shelf, Windows 7 64-bit RTM drivers.

Continue Reading

A few weeks ago, I had my main system’s OCZ Vertex 120GB drive die after a Firmware version 1.5 update. After applying the allegedly successful v1.5 firmware update, I started seeing data corruption issues. A few days later, the drive failed to be recognized by any system I own regardless of which onboard SATA controller I tried or even with raid card was used. Just to give an idea of the timeline:

1. Day 0- the SSD drive died, RMA request sent to OCZ (past working hours)
2. Day 1- SSD information such as serial numbers was given to OCZ and a RMA number was issued
3. Day 2 (Friday)- SSD was shipped to OCZ’s San Jose, CA facility (which is <20 miles away)
4. Day 5 (Monday)- SSD was delivered to OCZ
5. Day 9 (Friday)- Another SSD arrived via FedEx from OCZ in shrinkwraped retail packaging

Somewhere in there I ordered another 120GB Vertex from Amazon.com which arrived on Day 8 of the timeline above since I was running out of space anyway.

Continue Reading

I was checking Amazon the other day and saw both the 100GB and 200GB versions of the OCZ Vertex LE are available for pre-order. Both drives are next-gen (as of Feb 2010) performance wise and are said to be shipping sometime late February or early March.

I preordered the 100GB Vertex LE from Amazon.com. I’m guessing it will be available less expensively, but I wanted to get one of the 5,000 before they are sold out. It looks like these are going to be the only low-cost SF-1500 controller based drives for awhile, depending on what Other World Computing (OWC) does with their drives.

Introduction

For Small Business Servers and home servers, one thing becomes apparent, the need to use raid with large data sets. Raid 1 and software mirroring technologies like Windows Drive Extender are too cost ineffective for large storage arrays. For example, if one has 1 TB of storage need and uses one of these mirroring redundancy option, they need 1TB extra storage for a complete backup. Likewise, if an administrator uses four 1TB drives for data storage, then another four 1TB drives would be needed for redundancy. Raid 5 offers a user with both more performance than Raid 1 and lower “lost” capacity due to redundancy storage. Instead of effective storage capacity of n/2 for Raid 1, Raid 5 allows for n-1 of storage capacity.

Continue Reading

Here’s a quick update to the Patriot PS-100 32GB review including Raid 0 benchmarks with two drives.

Also, for comparison (to a much higher-end setup) see: Two OCZ Vertex v1.5 Firmware Raid 0 Benchmarks

Update 6 June 2010: New Patriot PS-100 32GB Firmware v3.000 benchmarks are up (single drive).

Continue Reading