The desktop but surprisingly, the difference is not much. Here are the stats when compiling the sample Crestron program: pro2 cen-track demo v1_0_8.smw

Laptop: Sony Z-series, Core 2 Duo Processor running at 2.66 GHz, Hard Disk storage, windows 7 x-64

First run: 9.36
Second run: 5.88
Third run: 5.64
Forth run: 6.41

Desktop: Core i7, 860 Processor running at 2.80 GHz, SSD for boot drive but sample program residing on 1 TB WD Hard Drive

first run: 5.77
2nd run: 5.27
3rd run: 5.32
4th run: 5.22


Excellent question and reason I listed multiple runs above. The way operating systems work is that when you read a file the first time, it is fetched from storage device and given to the application. But, a copy of the data is kept in memory (assuming you have enough memory). Should any application ask for that data again, the operating system checks the in memory copy and should the data reside there, it will fetch it from that location, rather than incurring a read from the storage device. This is called "caching" and is a wonderful way of keeping a CPU from waiting on slow devices such as hard disks.

In the example of the laptop, you see a sharp speed up after the first run. The storage there is entirely hard disk and slower variety in my laptop. The first time the program is complied, there is some amount of disk reads which are cached and as such, subsequent compiles go faster.

The Desktop system has the Crestron tools and libraries all sitting on SSD. This flash memory based drive, has no mechanical parts so it is able to eliminate the costly seek time and rotational latency. Even though SSDs are fast, they are still not as fast as system DRAM. So Windows still caches their content, showing faster performance still in subsequent runs of about 10%.

Seeing how we sit there working on the same program for a while, we can ignore the first invocation and look average of subsequent compiles. There, we see the desktop handling the job at 5.3 seconds relative to notebook at about 6 seconds. The difference then is about 10% faster which ironically, is pretty close to the ratio of CPU clock speeds!

Now, this data is for this one compile. There may be differences if different programs are compiled. But we can conclude that as long as you have enough RAM to cache your program and all the associated Crestron tools/libraries (likely a few hundred megabytes), the clock frequency of the CPU dominates.

That said, what doesn't come across is that my core i7 system "feels" lightning fast. Even though the rest of the cores don't participate in running SIMPL, they do let me open up the browser, do email, etc. with incredible speed because there are spare CPU cores available to service them. More important than the cores though, is the SSD. The difference that makes in the responsiveness of the system is incredible. Imagine rebooting the system in 10 to 12 seconds. It is almost as fast as most TVs these days! :). You basically never wait on anything. I can't recommend the SSD high enough.

Unfortunately high performance SSDs are expensive but you don't need much of them. As implied above, you need enough of it to host your OS and apps and the rest can sit on large hard disks. Mine is 120 Gigabytes but with everything loaded on my machine including everything Adobe makes from Photoshop to Premier, I am at 47 Gigabytes.


Fire away with your questions. In another life, I used to develop operating systems and one of my specialties was computer performance (see a now heavily outdated book I wrote 20 years ago: [Link: amazon.com]).