Initial Benchmarks of Patriot Warp V2 SSD on a Macbook

This is a follow up to a previous post in the Apple forum. I'm posting here as some others may be interested. Please note, these tests were all done on a Macbook under Leopard. I am not able to provide any PC based besnchmarks at this time.

---------------------------------

I did some comparisons over the weekend between my new SSD (Patriot 128gb WARP v2) and a couple other notebook drives. I tried to replicate the testing procedures on each drive. I started out with a full TimeMachine backup. Then I waited for all the indexing to finish before starting any tests. All tests were done three times and the average was recorded. Here are some actual "real-world" resuls as I got them.

Boot Leopard
- Patriot WARP SSD = 30 seconds
- Seagate 320gb 7200.3 = 36 seconds
- Fujitsu 120gb MHY2120BH (Stock Macbook Drive) = 41 seconds
- Seagate 80gb 7200.1 = 61 seconds

Open Word 2008
- Patriot WARP SSD = 4 seconds
- Seagate 320gb 7200.3 = 11 seconds
- Fujitsu 120gb MHY2120BH (Stock Macbook Drive) = 13 seconds
- Seagate 80gb 7200.1 = 13 seconds

Start Firefox
- Patriot WARP SSD = 3 seconds
- Seagate 320gb 7200.3 = 9 seconds
- Fujitsu 120gb MHY2120BH (Stock Macbook Drive) = 9 seconds
- Seagate 80gb 7200.1 = 10 seconds

Open "The Matrix" Xvid with Quicktime
- Patriot WARP SSD = 5 seconds
- Seagate 320gb 7200.3 = 8 seconds
- Fujitsu 120gb MHY2120BH (Stock Macbook Drive) = 8 seconds
- Seagate 80gb 7200.1 = 8 seconds

***I tried a whole bunch of different file copy tests, and the SSD and the 7200.3 seemed to work at about the same speeds (with the other two drives lagging behind). I also tried a bunch of different write tests, and found no discernable difference between the SSD and the 7200.3.


Here are some Xbench 1.3 results:

Sequential Uncached Write (4K blocks)
SSD = 56.41 MB/s
7200.3 = 80.40 MB/s
Fujitsu = 43.18 MB/s
7200.1 = 31.31 MB/s

Sequential Uncached Write (256k blocks)
SSD = 62.78MB/s
7200.3 = 65.40 MB/s
Fujitsu = 40.39 MB/s
7200.1 = 29.18 MB/s

Sequential Uncached Read (4k blocks)
SSD = 15.25 MB/s
7200.3 = 15.96 MB/s
Fujitsu = 9.61 MB/s
7200.1 = 10.98 MB/s

Sequential Uncached Read (256k blocks)
SSD = 93.84 MB/s
7200.3 = 75.79 MB/s
Fujitsu = 41.91 MB/s
7200.1 = 29.72 MB/s

Random Uncached Write (4K blocks)
SSD = 1.99 MB/s
7200.3 = 0.89 MB/s
Fujitsu = 0.91 MB/s
7200.1 = 0.68 MB/s

Random Uncached Write (256k blocks)
SSD = 38.77MB/s
7200.3 = 41.65 MB/s
Fujitsu = 22.48 MB/s
7200.1 = 18.16 MB/s

Random Uncached Read (4k blocks)
SSD = 7.64 MB/s
7200.3 = 0.61 MB/s
Fujitsu = 0.47 MB/s
7200.1 = 0.51 MB/s

Random Uncached Read (256k blocks)
SSD = 74.25 MB/s
7200.3 = 25.04 MB/s
Fujitsu = 17.91 MB/s
7200.1 = 16.05 MB/s

 

Oooh very nice. How is the Patriot treating you? Any stutters?

 

I did not experience any slowdowns with small writes or with multi-tasking. It just plowed through everything I threw at it--audio recording, video recording, encoding, files copies, playing videos, opening photos, etc. I even did a whole bunch of these at the same time, and still now slowdowns. From all of my tests, I can say with some confidence that this drive does not suffer random slowdowns (at least not in Leopard).

 

Man... with all these mixed reviews, I really don't know what to get.
Can you please run IOmeter (hope I"m not asking for too much)?

Glad the drive serves you well

 

I would, except I can't find a mac version.

 

do you really start Firefox in 9 seconds with your HDDs ??

my IE in 3-4 seconds shows the google page already, thought the firefox ran even faster

 

My desktop with a seagate 250gb 7200rpm starts firefox in like 1 second . 2 seconds at most. Suppose desktops can't be compared to notebooks, but my HDTune is like nothing compared to many of the notebook HDTunes here

 

That's amazing man. The Patriots are said to have JMicron controllers that suffer from slow random writes. Still some people, including you, do not experience any stuttering.

How long have you used the drive now?

I thought the same. 9 seconds seems to be very slow. Maybe he has many plugins, history, cookies and bookmarks?

 

I have had it a week.

About the slowdowns, my understanding is that the v2 drives have an optomized firmware that helps minimize the slowdowns. I also am using OS X, and understand the slowdown issues can be more pronounced on an NTFS file system.

But, yes, I have not had any slowdowns--even when I tried.

 

Well done benchmarks by the way. Very interesting to see the real world differences measured like that.

By the way, I find it hard to believe that the OS X would not have a problem with slow writes. Maybe your MB uses a different SATA controller though.

 

Like I said, I have tried to replicate the slow-write issue. Any suggestions on how I might try to really tax the drive into a slow down.


PS - I just wrote a little script to open 15 apps at once (word, excel, powerpoint, etc.). The SSD had them all open in a matter of a few seconds. No traditional HD can do that.

 

One of the things I heard that worked on a OCZ Core was to open (let's say) 5 tabs in Firefox with Youtube movies playing at the same time. Or see how many tabs you can open that way.

At Anandtech they describe at situation that worked on a Mac Pro with Supertalent SSD, but it does not sound like you had any of that:
http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3403&p=7

You could also try to emulate one of the tests that Laptopmag did. then you can see how you're drive compares to Intel en OCZ Core.
http://www.laptopmag.com/review/storage/intel-x25-m.aspx?page=4

 

OK. Tried this. Opened 15 tabs in Firefox with Youtube movies playings. It only started to studder when the CPU reached about 100%.

While these 15 tabs were opened and all the movies were playing, I opened Google Earth and it opened normally. The only studders came when the CPU hit 100%.

Actually, I am posting this while another instance of Firefox is running 5 Youtube videos and Safari is running 5 youtube videos. And, I'm also IMing with my sister-in-law. CPU is running near 80%.

Sorry, this is pretty hard to replicate on a mac platform. There is so much variety here, like CPU speed, RAM amount, number of files and their content, and the program being used to compress and decompress.

 

Sounds like your CPU is the bottleneck there, not your hard drive . Glad to hear its working good for you

 

Ok sounds pretty convincing to me.

Now it would be nice to see if someone could repeat this with XP and Vista.

 

This is not directed to you OP...but I just personally feel that it's unfair to compare SSDs to HDDs at this point in time. From the reviews and info. i've seen and read; SSDs are just plain out better than HDDs. They should really be compared to their own kind now...they're in a class by themselves...seriously!

 

Well, if the stutters are fixed, they definitely are going to own Hard Drives. I would have bought the Patriot SSD had it not been the risk I would have been taking if it stutters.

 

If I had a few extra days and more patience, I would try it. But for now, somebody else is going to have to try that one.

There are some rumours floating around that these v2 Patriots are not using the jmicron controller. THAT would be something worth conforming. I was under the impression that it was just a better tuned firmware.

Hmm....maybe I should open it and look? Um. Not gonna void the warranty after shelling out the bucks.

 

I think I was the one who started those
I got that information from an official post @ patriot forums. Not sure if they're saying the truth or not though

 

Email sent to patriot. Lets see what they say.

 

I installed Vista on a BootCamp partition on my Macbook today. I'll play around with it and update with some benchmarks.

The one thng I have done is load IE with 6 tabs of YouTube videos. No slow downs. I'll post more later.

 

Awesome numbers and benchmarks Cant wait to get mine .

 

Must say that I do find this quite interesting indeed. Though really, I don't find it all too surprising that OSX does a better job with file IO than microsoft...

Same setup as with my other testing, but now in OSX (no, it's not a mac >.>)
Xbench numbers, in same order as those in the original post.
4k sequential write - 86.49 MB/sec
256k sequential write - 83.23 MB/sec
4k sequential read - 25.79 MB/sec
256k sequential read - 142.52 MB/sec
4k random write - 1.87 MB/sec
256k random write - 36.95 MB/sec
4k random read - 14.25 MB/sec
256k random read - 118.01 MB/sec

Now then, what to make of this? Well, the lack of difference in sequential writes between 4k and 256k really implies to me that OSX is doing write caching, despite Xbench's claim that it's uncached. This would go along purrfectly with the random write performance being decent - these drives only crash and burn because the jmicron controller (yes, it is a JMF602 controller on there) doesn't cache writes.