Upgrading RAM – How does it affect Kontakt 5?

Upgrading RAM – How does it affect Kontakt 5?

In theory, faster ram equals less waiting time. At least that is the principle of RAM speeds. But how does RAM speeds affect loading times in Native Instruments Kontakt?

I recently upgraded my DAW computer with new RAM (memory) from 1600 MHz DDR3 to 3200 MHz DDR4 and I decided to do a preliminary test to see if the loading time in Kontakt was any better with the new RAM.

About the test

The way I tested the RAM in Kontakt 5 was to video capture the screen and measure how long it took for Kontakt to load the samples into the RAM. This is probably the best and easiest way to measure how long a user would have to wait in a real world situation. I recorded the loading times, making sure i restarted the computer for each test. Once I had the results I compared the old RAM timings with the new RAM results.

The two RAM sets I have compared are:

Both RAM sets ran at max speed, which means that both RAM were maxed out with their respective X.M.P profiles. With the X.M.P profiles enabled, the DDR3 had a latency timing of CL11 and the DDR4 had a latency of CL16. This means that that the DDR3 has theoretically lower speed but might benefit from the low latency.

The actual loading time in this test is from a library is opened (double clicked) to it is completely loaded into RAM. All settings in Kontakt are set to the default values on a clean installation of Kontakt 5. All samples are loaded from a traditional 3.5″ spinning hard drive.

Test results

Ram speeds compaction. Less is better.
Ram speeds compaction. Less is better.

 

Instrument RAM (1600MHz) RAM (3200MHz)
The Giant 15,5 15,2
Action Strikes 23,5 23,0
India 13,8 14
Action Strings 56,0 57,2
Una Corda 19,8 20,0
Average 25,72 25,896

The test shows that the results with the different RAM modules are almost identical. Since we are talking about milliseconds in difference it is only fair to say that DDR4 at 3200MHz does not reduce loading times in Kontakt 5. At the most there is a difference of only a second. On average there is only milliseconds difference. As I have discussed in previous videos the test results indicated that even if the DDR4 ram speed is double the DDR3 speed Kontakt is not able to benefit from the RAM with higher speed. The reason for this is completely open for discussion. It is most likely that the speed of the DDR4 gets cancelled out by the latency, but in theory there is no consensus for this. Crucial is criticising that the latency-speed theory is far more complex. You can read more about this here. It might also be that the DDR4 chosen for this test is not the best on the marked. Other than that Kontakt might also not me programmed to take full advantage of DDR4 modules. It is also important to note that these results are limited to a traditional hard drive. The results might be different on an SSD drive.

Conclusion

Either way this test concludes that just upgrading RAM to a higher speed does not necessarily give any performance boost in Kontakt 5. If it does, we are talking about milliseconds less waiting time for libraries to load. If you have any tips on how to reduce loading times in Kontakt 5 let us know in the comments.

  1. Your test is, unfortunately, the most stupid I’ve ever seen. Firstly, loading time is obviously the same for different RAM modules, you don’t need any tests to know this (the loading speed is limited by performance of storage subsystem), and, the second, people install faster ram not because they just want libraries to load faster. The faster ram you use, the lower buffer and latency you, potentially, can achieve.

    1. Hi Alex and thank you for your input. You say that “loading time is obviously the same for different RAM modules”. I have never said otherwise and I completely agree. “the loading speed is limited by performance of storage subsystem” -while this statement is true in general, Kontakt will always try to load the whole library on to the RAM and not the disk. Unless you have insufficient RAM. If you have insufficient RAM, well yes, Kontakt will try to store samples on disk, in which case it really doesn’t matter which RAM you have because the disk speed will be a major bottle neck for Kontakt. In this test I had enough RAM to load the whole library on the RAM and so the disk speed was completely irrelevant. “The faster ram you use, the lower buffer and latency you, potentially, can achieve” This is actually wrong. The buffer and latency (either in Kontakt or your DAW) is completely dependent on the capacity and speed of the CPU, not the RAM so this has nothing to do with this test. Oh by the way, Merry Christmas 😀 !

  2. Sorry, ya this is a really pointless test. 1600Mhz of RAM means it can load 1,600 million bytes of data per second or 1.6Gb/s. If you don’t have a disk drive with at or above 1.6Gb/s read speed you’re immediately bottle-necked from the speed of the non volatile drive. RAM like any other memory can only store data as fast as it is received.

    There was no real need to test this…speeds are measured and detailed in all hardware specs and promo materials.

    The differences in your results, of maximum a second difference, are likely due to other processes taking reads from the same disk, background processes, or quite possibly even writing to disk variations in your screen capture if this was on the same drive.

    1. What you are saying is almost correct, but Kontakt does not write the samples to the hard drive. Instead it loads the samples to the RAM, when possible. So the hard drive is not a bottleneck at all here. The hard drive can only be a bottleneck in a situation where there is not sufficient RAM capacity.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.