Dual vs single channel results
As I said I was going to post dual vs single channel results, here we are. The computer I have is the Pavilion 17 G101DX with 750GB 5400RPM hard drive 6GB (2x4) CL11 RAM in dual configuration, and an i5-5200U processor. I finally upgraded to 256GB SSD, 16GB CL9 RAM in the last test.
I first ran a test with the stock 6GB CL11 dual channel memory (it was one 2GB samsung ram and another unbranded 4GB stick), then added one stick of 8GB G.Skill CL9 memory, and finally, I upgraded with an SSD and added another stick of G.Skill 8GB ram for a test using 16GB (CL9) of dual channel memory. . Both memory stick were 1600Mhz. The tests were performed with the latest PerformanceTest suite (passmark tests I think)
The stock 6GB dual channel ram (2GB samsung CL11 and unbranded 4GB CL11) and ran two PerformanceTest tests for memory. the mean for the two tests is:
1761
The tests all had scores equal to or higher than the 8GB single channel, except, obviously, the amount of available memory.
In single channel with 8GB of ram the mean of the two memory tests were:
1620
The final test after the addition of an SSD (I don’t know if an SSD affects memory or not, but I assume it doesn’t), and another stick of 8GB of ram to a total of 16GB G.Skill CL9 in dual config. The mean of two Memory tests run is:
2112
So here we can see an increase in total memory performance by switching from single channel to dual channel. I have not performed any statistical tests on the numbers, but maybe I will do so, with more specific detailed analysis, in the near future. So I cannot say conclusively if there is any significant increase in performance (more than due to chance). The small amount of tests run also increases the false positives of the results and my statements about performance increases from single to dual channel configurations are purely subjective.
When running the first test, even with less total memory (weakens the total score) it still had a higher score than the single channel configuration with a lower clock number (faster memory access than a higher clock number). This is due to the ability to theoretically receive “double” the amount of data at once -128 bits per chunk over 64 bits. Although, in Dual channel, it must switch channels as the data comes in. Although I’m not an engineer or computer specialist I propose theoretically that the since my processor is a mid range processor with 2 physical core and 4 logical (i5-5200U), accessing dual channel configurations gives an intermittent wave of data (it must pause some data since so much comes at once. It lines up the data in queue. The speed of the computer is limited by the single channel 64 bit configuration, as the processor is a fairly quick one (the fastest I have had the luxury to use anyway).
The final test shows a nice increase in memory benchmark, but we have to take into account the increase in score from the amount of RAM added -it is double the RAM. If I had done the test with 2x4GB sticks, it would have given a more “realistic” comparison score.
In conclusion, I subjectively conclude that dual channel configurations are better performance-wise than single channel configurations, especially if your processor is not the bottleneck on your PC.
