Has Bare Feats helped you? How about donating to Bare Feats?
Does Corsair's Low Latency Memory
really make your Mac go faster?
Originally posted April 25th, 2008, by rob-ART morgan, mad scientist
Updated May 1st, 2008, with "video to layers" test similar to Corsair's
Corsair's press release and "white paper" made a big splash in the news. They claimed as much as a 19% gain in Photoshop performance when a 4GB kit of CAS 5 memory was replaced with their 4GB kit of CAS 4 memory. We were skeptical so we tested the Corsair modules against various brands of "normal" latency SODIMM memory kits in our "Penryn" MacBook Pro 2.6GHz laptop.
First we ran a memory benchmark. We used a specially built version of STREAM to run in 64 bit parallel multi-threaded mode and use 2.7GB of RAM.
Next we ran three real world apps that gobble up over 3GB of RAM and use all available CPU cores.
Finally, we ran a test similar to that offered by Corsair as proof of speed gains. They used Photoshop to convert a video to layers. We took an Apple ProRes 422 movie file and converted the first 500 frames. We rebooted between runs to defeat caching and averaged three runs. They claim this test produced the fabled 19% speed gain when shifting from CAS 5 to CAS 4. We saw vitually no gains.
All memory was PC2-5300 667MHz 200pin SODIMMs
Corsair C4 = Corsair VSA4GSDSKIT667C4 CAS 4, 4GB (2 x 2GB) kit
OWC C4 = OWC OWC53C4DDR2P4GB CAS 4, 4GB (2 x 2GB) kit
Corsair C5 = Corsair VS2GSDS667D2 CAS 5, 4GB (2 x 2GB) kit
OWC C5 = OWC OWC53C5DDR2P4GB CAS 5, 4GB (2 x 2GB) kit
The Corsair CAS 4 low latency memory is "oh so slightly" faster than the Corsair CAS 5 "normal" latency memory. Ditto for OWC's CAS 4 versus CAS 5.
The STREAM benchmark showed a 5% increase comparing CAS 4 memory to CAS 5 memory. But in the REAL WORLD, the average gain for three memory intensive apps using low latency memory was 1%. Though consistent, it's not significant in my estimation. Without a stopwatch, you would not notice the difference.
We saw nothing close to the 19% gain in Corsair's Photoshop test published in their "White Paper" PDF when they compared a CAS 5 four gig kit with a CAS 4 four gig kit. I think somebody in their lab either goofed or moved a decimal point.
If you are getting ready to upgrade from 1GB or 2GB to 4GB -- and you want to squeeze every drop of speed out of your MacBook Pro or MacBook or iMac, then maybe low latency memory is worth getting. If you already have a CAS 5 "normal" latency 4GB kit installed, then it's NOT worth trading up for a CAS 4 kit.
And if you do go for a low latency CAS 4 kit, take a look at OWC's 4GB CAS 4 kit which is priced lower than the Corsair CAS 4 kit.
Thanks goes CanadaRAM and OWC for providing various and sundry memory sample memory kits.
For the record, we tested five brands of CAS 5 memory (Corsair, DMS, Kingston, OWC, and Patriot). They all performed within tenths of each other. We decided to graph only the OWC and Corsair since we had samples of both CAS 4 and CAS 5 from those memory makers.
WHERE TO BUY MEMORY FOR MAC LAPTOPS
We have tested the memory from the following third party companies in our MacBook Pro. We can recommend them wholeheartedly:
Other World Computing sells Samsung, Techworks, and OWC memory kits for the MacBook Pro, MacBook, and iMac. They offer both CAS 5 and CAS 4 latency.
TransIntl.com is selling kits to take your MacBook or MacBook Pro up to 4GB.
CanadaRAM features Kingston memory kits for the Macs. They also provide access to Corsair, DMS, and other brands on request.
WHERE TO BUY A MACBOOK PRO
When ordering products from Apple Store USA, please click THIS TEXT LINK or any Apple display ad as your "portal" to the online store. In so doing, you help to support Bare Feats.
Has Bare Feats helped you? How about helping Bare Feats?
© 2008 Rob Art Morgan
"BARE facts on Macintosh speed FEATS"
, the webmaster