OWC Shares Mid-2014 Retina MacBook Pro Unboxing, SSD Tests

The analysis is extremely poor. You should not be looking at the average data rate when testing multiple transfer sizes because the difference between 4KB and 1024KB transfer speeds is close to tenfold. OWC's testing is also fairly bad because they only ran the standard suite, which tops out at the IO size of 1024KB and thus cannot take full advantage of the drive.

The SSDs are the same as in the original Haswell rMBP, so to be honest there is nothing to see here. Bad tests and even worse analysis are just confusing everyone now.

I didn't realize how much MR dropped the ball on this one until I clicked over to the OWC article. Besides MR listing the average speed instead of the top speed (OWC just posted the screenshot and didn't single out the pointless "average" number), OWC has two more tests for each drive with larger file sizes and much higher numbers.

http://blog.macsales.com/25770-owc-unboxes-tests-ssd-speeds-of-new-13-and-15-retina-macbook-pros

Top numbers were actually 791 read 744 write for the 256G and 762 read 454 write for the 128G. Those are really good numbers, faster than any SATA drive can handle and I suspect comparable the previous generation of these machines if not a bit faster.

Really misleading spin by MR, you guys really should correct your article as you give the opposite impression of what the actual speeds of these drives are.

The analysis is extremely poor. You should not be looking at the average data rate when testing multiple transfer sizes because the difference between 4KB and 1024KB transfer speeds is close to tenfold. OWC's testing is also fairly bad because they only ran the standard suite, which tops out at the IO size of 1024KB and thus cannot take full advantage of the drive.

The SSDs are the same as in the original Haswell rMBP, so to be honest there is nothing to see here. Bad tests and even worse analysis are just confusing everyone now.

FYI, OWC updated its original post several hours after it posted the first two tests, which were the only available at the time. I've updated the post based on your feedback and the new tests.

Granted I'm not that familiar with how SSDs perform their reads/writes vis-a-vis against magnetic media, but the numbers cited in the article seem a little suspect:

R-Read R-Write S-Read S-Write
256GB 194 322 262 245
128GB 165 131 263 244

How is a Random Write faster than anything else?

Is it just me are these supposedly "superior" flash drives actually fairly poor? My neutron gtx ssd in my 2010 macbook still posts scores after 2 years of use that absolutely smoke these...

As another data point, here is my BlackMagic test of the new version 128GB 13" Retina I picked up yesterday. This is with Filevault on.

The analysis is extremely poor. You should not be looking at the average data rate when testing multiple transfer sizes because the difference between 4KB and 1024KB transfer speeds is close to tenfold. OWC's testing is also fairly bad because they only ran the standard suite, which tops out at the IO size of 1024KB and thus cannot take full advantage of the drive.

The SSDs are the same as in the original Haswell rMBP, so to be honest there is nothing to see here. Bad tests and even worse analysis are just confusing everyone now.

Thank you. That makes much more sense now. Shame on MR for the very misleading spin.