Mismatching RAM: Benchmarks

blibbax · Jan 2, 2009, 01:00 PM

I'm being tempted at the moment by a move to 3GB of RAM, in my single channel dual slot DDR2 motherboard. I'd then complete the move to 4GB when I get round to getting 64 bit. So, I decided to run a few tests using some old sticks I had lying around. The results aren't surprising, but I thought people might be interested.

Test setup:
-Motherboard: Foxconn P4M9007MB 8RS2H
-Processor: Intel E2220 @2.4ghz
-RAM:
-2x1GB Kingston DDR2-667
-1x512mb Crucial DDR2-533

I did each test three times and took a best result.

I included the 667mhz setup just for some perspective, but the relevent comparison is betwen the two 533mhz sets. They manage very similar results, despite one of them being two identical sticks and one being mismtached.

nicnik · Jan 2, 2009, 04:10 PM

the 2x1 in dual channel, same brand, match etc?

Matth · Jan 2, 2009, 09:45 PM

Single channel chipset, and shows that there is no advantage in matching on single channel - well, one advantage, if the speeds/timings are different, it will have to use the slowest.

Puzzled by the 667 gain in read speed, versus negligible gain in write speed.

blibbax · Jan 2, 2009, 10:50 PM

Quote:

Originally Posted by nicnik

the 2x1 in dual channel, same brand, match etc?

It's not a dual channel motherboard, but yes they're even same week.

Quote:

Originally Posted by Matth

Single channel chipset, and shows that there is no advantage in matching on single channel - well, one advantage, if the speeds/timings are different, it will have to use the slowest.

Puzzled by the 667 gain in read speed, versus negligible gain in write speed.

Same here!

I added the Super Pi chart, forgot to the first time.

Jan 2, 2009, 01:00 PM	#1
blibbax What does this do? Join Date: Sep 2008 Location: Oxford, UK Posts: 4,230 Rep Power: 254 System Specs	Mismatching RAM: Benchmarks I'm being tempted at the moment by a move to 3GB of RAM, in my single channel dual slot DDR2 motherboard. I'd then complete the move to 4GB when I get round to getting 64 bit. So, I decided to run a few tests using some old sticks I had lying around. The results aren't surprising, but I thought people might be interested. Test setup: -Motherboard: Foxconn P4M9007MB 8RS2H -Processor: Intel E2220 @2.4ghz -RAM: -2x1GB Kingston DDR2-667 -1x512mb Crucial DDR2-533 I did each test three times and took a best result. I included the 667mhz setup just for some perspective, but the relevent comparison is betwen the two 533mhz sets. They manage very similar results, despite one of them being two identical sticks and one being mismtached. __________________ Donate the spare computing power of your PC to help to cure Alzheimer's, ALS, Huntington's, Parkinson's disease and cancer: Fold for HH! Last edited by blibbax; Jan 2, 2009 at 10:51 PM.

Jan 2, 2009, 04:10 PM	#2
nicnik In the Land of Snow Join Date: Oct 2005 Location: London, UK Posts: 569 Rep Power: 46 System Specs	Re: Mismatching RAM: Benchmarks the 2x1 in dual channel, same brand, match etc? __________________ Follow Me on Twitter Battlefield 3 \| youtube \| TwitchTV

Jan 2, 2009, 09:45 PM	#3
Matth Flash Banner Hater Join Date: Jun 2002 Location: UK Posts: 3,426 Rep Power: 93 System Specs	Re: Mismatching RAM: Benchmarks Single channel chipset, and shows that there is no advantage in matching on single channel - well, one advantage, if the speeds/timings are different, it will have to use the slowest. Puzzled by the 667 gain in read speed, versus negligible gain in write speed. __________________ Mary had a little lamb, Her father shot it dead Now Mary takes her lamb to school, Between two crusts of bread