2011 iMac Thunderbolt SSD Benchmark

[supersid]

I have run some benchmarks. I do not know what I am doing so here are some results for you all.

Here is a very un-useful Xbench comparison:

http://db.xbench.com/merge.xhtml?doc1=529855&doc2=1&setCookie=true

 

[rcook]

Do you mind running those same benchmarks on the HDD for comparison?

 

[dwmac]

As a point of comparison, here are XBench SSD benchmarks from an OWC Mercury Pro (SATA2) in my MBP 5,3:

Disk Test 238.36
Sequential 143.04
Uncached Write 170.08 104.42 MB/sec [4K blocks]
Uncached Write 162.11 91.72 MB/sec [256K blocks]
Uncached Read 75.46 22.08 MB/sec [4K blocks]
Uncached Read 375.54 188.74 MB/sec [256K blocks]
Random 714.32
Uncached Write 1073.04 113.59 MB/sec [4K blocks]
Uncached Write 335.46 107.39 MB/sec [256K blocks]
Uncached Read 1603.78 11.36 MB/sec [4K blocks]
Uncached Read 940.46 174.51 MB/sec [256K blocks]

The iMac's SSD has substantially higher overall scores, although it is slower at some specific tests.

 

[philipma1957]

the 4k random write/read are 56 and 17 for the imac vs 113 and 11 for the macbook pro.

now if these are accurate the oem apple ssd does random reads faster the the owc. meaning better app openings etc.

 

[barbarbar]

Can you use Geekbench benchmark to review if it´s more powerfull than his brothers withouth SSD´s?

Thanks, i´m thinking on getting the top imac 27" withouth SSD or the top imac 21" with SSD.

 

[Michael D.]

Can you use Geekbench benchmark to review if it´s more powerfull than his brothers withouth SSD´s?

Thanks, i´m thinking on getting the top imac 27" withouth SSD or the top imac 21" with SSD.

If I'm not mistaken, Geekbench measures mostly processor and memory performance.
SSD's influence is most likely negligeable on Geekbench.

 

[bpeeps]

Can you use Geekbench benchmark to review if it´s more powerfull than his brothers withouth SSD´s?

Thanks, i´m thinking on getting the top imac 27" withouth SSD or the top imac 21" with SSD.

I second this.

 

[barbarbar]

It´s just to see this. If ssd give to this iMac a better benchmark.

P.D. Sorry about my bad english, i´m from Spain.

If I'm not mistaken, Geekbench measures mostly processor and memory performance.
SSD's influence is most likely negligeable on Geekbench.

 

[philipma1957]

xbench shows off ssd or hdd more then other parts of the machine. geekbench shows off ram and cpu more then other parts of the machine. here is an xbench on a 2009 imac just the hdd scores note the 4k write and read. it has a 1tb seagate in it the write score is 1.3 the read is .8 far lower then any ssd. this is the lag that hdds have.

 

[supersid]

Can you use Geekbench benchmark to review if it´s more powerfull than his brothers withouth SSD´s?

Thanks, i´m thinking on getting the top imac 27" withouth SSD or the top imac 21" with SSD.

I have downloaded Geekbench, but that one was paid so I did not run Geekbench.

 

[barbarbar]

If you run it on 32-bit mode, you don´t need to pay, it´s free, and it can provide us an benchmark to compare with another macs withouth ssd. You must use the later option when asked you for registration.

Thanks!

I have downloaded Geekbench, but that one was paid so I did not run Geekbench.

 

[plinden]

I'm no expert, but my understanding is that something like 75% of all disk I/O is in small blocks, so 4K/8K block writes (reads isn't usually the bottleneck) is the number to use in comparisons.

 

[Domino8282]

Can someone who self-installed an SSD in their iMac post the same benchmarks for comparison?

 

[philipma1957]

I'm no expert, but my understanding is that something like 75% of all disk I/O is in small blocks, so 4K/8K block writes (reads isn't usually the bottleneck) is the number to use in comparisons.

both writes and reads under 8k are the most important numbers for users. this is why getting a vertex 3 xbench score from a 2011 iMac showing great random 4k writes and reads would be nice . I will troll around on the xbench result site and look. maybe I can find a score for vertex 3 on a macbook pro. i found a vertex 3 score for a macbook pro

http://db.xbench.com/merge.xhtml?doc1=529787

it show 4k random writes at 280 5x better then the oem ssd but 4k random reads are 20 vs 17. what does it mean opening apps will be close to the same and if you are lite ram the vertex 3 will be faster then the oem ssd. any time you get a page out the vertex 3 will be a lot faster. so power users that go near or over 16gb ram would want a vertex 3 over am oem ssd. most users are not getting much more for the money. time and effort of installing a vertex 3 in an imac. in a macbook pro yes because the install is easy.

 

[iamthedudeman]

Here are my benchmarks I posted in the other thread.

Disk test, xbench overall, geekbench. As expected the SSD has no bearing on geekbench scores, same as the HD version.

I have a 2011 21.5 i7 2.8 with SSD and 1TB HD.

 

[iSayuSay]

For a second there I thought this thread is about benchmarking external SSD plugged into Thunderbolt (looking from the title)

That would be super if true, indeed
But it´s fun to read internal SSD benchmarking too

 

[theSeb]

The problem with random read and write benchmarks is that they become meaningless after a certain point because if the one drive is at that point, it will make no difference to real world performance when you install a drive that is twice as fast. Anandtech has a good explanation of this.

 

[ecstacy2012]

I have the 2011 imac 27" 3.4 ghz with 2gb graphics card. I have done the ssd custom install with the owc 120gb sata III 6g ssd. I get about 530's/516 read write speed, this is benched with aja system test on the highest res setting with 10bit and 128gMB write read file size. It's the fastest ssd you can get. One word of warning to anyone installing it to install osx on it, the osx disk doesn't see the drive nor does disk utlity when using boot from disk. thus you will have to clone another internal drive (such as hd drive) over for OSX to be installed. I contacted owc when they tested the drives they cloned so they were not aware of this, but other then that its unbelievable fast and I'm working on raiding two of them in 0 raid to get near 1gb internal ssd read write speed. Oh and geek bench I got a score of 19,800 of course this was with the stock 4gb's of ram I haven't gotten the 16 gb ram kit I'm putting in yet then I'll rebench!

 

[chfied]

imac i7 3.4 + vertex 3

Replaced internal drive with a Vertex 3. Xbench disk results:

 

[ecstacy2012]

hey xbench doesnt show its full potential and use aja system test please and show results set it to 128mb and 10 bit with full resolution settings to get an accurate result.

 

[snugla]

One word of warning to anyone installing it to install osx on it, the osx disk doesn't see the drive nor does disk utlity when using boot from disk. thus you will have to clone another internal drive (such as hd drive) over for OSX to be installed.

So you're saying there's no way to make a SSD a boot drive...is that correct? Applications can run off of them...just not the OSX?

 

[ecstacy2012]

im saying it only in regards to owc drives will not see the drives no matter what you do the only way around it is to clone over to the internal thus you can't clone onto a raid and no way to build the raid first because it won't see the drives. I'm using vertex 3 drives and they work amazing (now my results are averaging 1100kb read/write. Aslong as you use vertex you can boot and also raid them, Considering the thunderbolts are 600kb with 4 drive raid and the highest most expensive one is 800kb I'm still killing them and wayyy cheaper let alone it is internally