A 32GB SSD drive for $60 after MiR sounded like a good idea, especially since that would be great for a Windows Home Server C:\ drive, so I purchased a Patriot PS-100 32GB. (It should be noted that I have not seen the price for $60 again in the following 3 months) 64MB of cache sounded better than a JMicron JMF602 based drive, and surely Patriot would not release in the fall of 2009 a JMicron branded drive. I assumed that the drive would be Samsung, Indilinx (akin to an OCZ Core V2) since 32GB was not an Intel controller sounding size (40GB, 80GB, 160GB MLC to date). My conjecture proved to be inaccurate.
Also, just for comparison here are Raid 0 Benchmarks with the PS-100 32GB and for a comparison (to a much higher-end setup) see: Two OCZ Vertex v1.5 Firmware Raid 0 Benchmarks and see two Intel X25-V 40GB’s in Raid 0 benchmarks for comparison of the Patriot PS-100 32GB to the popular Intel X25-V 40GB.
UPDATE 05 April 2010: See performance with new Patriot firmware #2.008 here
UPDATE 06 June 2010: New Patriot PS-100 32GB Firmware v3.000 benchmarks are up (single drive). This release had significant performance improvements.
Opening the drive one can clearly see the Toshiba flash, Hynix DRAM that is the cache, and a chip that says Phison.
So the question is, what is a Phison? Here’s a closeup of the chip:
Of course the first question I had was, how does it perform? Here’s the CrystalDiskMark of the Patriot PS-100 32GB:
I thought that was an odd result for a SSD on a 3.2GHz Core i7 920/X58 system, especially plugged directly into the ICH10R. I re-ran CrystalDiskMark five times, then tried it three times in the PM45 based laptop before realizing that this was repeatable on both platforms. ATTO was a bit better:
Just for comparison… here are a few other CrystalDiskMark runs on other drives.
It just so happens that my 60GB Agility and 120GB Vertex (non-turbo) perform very similarly so I only posted the Agility CrystalDiskMark. Still, the Patriot PS-100 gets obliterated on the read and 4k write tests. Being fair, the Patriot drive is extremely inexpensive at $89.99 – 30 rebate at Fry’s on 25 November 2009. On the other hand, it does make a strong case for the 40GB Kingston and upcoming Intel SSD’s that went on sale for $115/ 40GB. Also of note, the speed is substantially better than a 7,200rpm drive, and one gets all of the no-noise and vibration that comes along with a SSD.
Here’s a quick Agility ATTO run:
As one can see, the Patriot drive is a fair amount slower, which is to be expected at its low pricepoint.
I’m going to spend the next few days playing around with the drives looking for signs of stutter. Also, I picked up a second drive just to see what will happen in Raid 0.
UPDATE: for comparison here are Raid 0 Benchmarks with the PS-100 32GB and for a comparison (to a much higher-end setup) see: Two OCZ Vertex v1.5 Firmware Raid 0 Benchmarks Also see two Intel X25-V 40GB’s in Raid 0 benchmarks for comparison to the popular Intel X25-V 40GB.
UPDATE 05 April 2010: See performance with new Patriot firmware #2.008 here
UPDATE 06 June 2010: New Patriot PS-100 32GB Firmware v3.000 benchmarks are up (single drive). This release had significant performance improvements.
thanks for the heads up. post raid 0 numbers soon!
Did you noticed stuttering on random 1mb+ writes? I put up some pictures of the drive stuttering over at slickdeals. http://slickdeals.net/forums/showthread.php?t=1695040&page=2
see the last 2 posts. I repeated this test 3 times and all came out the same. Maybe HD Tune is doing something funny. This is tested on a Dell Precision T3500 on Xeon i7 2.66ghz and Intel ICH10R on board controller.
Looking forward to your raid 0 benchmarks.
From what I understand there is no benefit running an SSD as a boot drive.. Please provide information that sat otherwise because Newegg has the OCZ Vertex for $99 after MIR right now.
BTW, thanks for such an informative post. I was looking into this drive myself and was wondering about the controller.
great review…. any signs of stutter?
That odd spiking isn’t exactly stuttering like the JMicron JMF 602 based drives had. I haven’t been seeing the JMicron style stutter. That being said, the access times are really high as quik2 has eluded to. Somewhere in the range of 1ms.
nekon, there is a TON of advantage to using SSD’s as boot drives as boot is one of the areas where there is a lot of random access and small 4k reads/writes. All of my machines currently use SSD’s in them and all spindle based disks are now relegated to servers for storage. If I need to access files it is all over GigE/ Wireless N, and the setup works great.
Try these drivers to see if you see an improvement:
http://download.biostar.com.tw/upload/Driver/Chipset/Intel/SATA/Utility/ICH7_10/iata_cd.exe