Winstone Applications Usage
Winstone loads a number of common business applications into memory and then test for system performance. What it does not reflect is the load time of each program, so you need to look at your own use to decide which is better for you – more memory or RAID 0. At work I load the same applications onto the desktop and keep them there all day (actually 24/7 since the system is always on), so Winstone will give an accurate picture of my work system. But my home system is used by a number of different people who load different applications that they close when they finish using them, so Winstone will not be as accurate since the load time of each program is not included in its score (Winstone does measure the load time of a subset of applications, but not all of them). I also deal with disk access at work that consists mostly of small files or reading / writing small amounts of data at a time, so RAID 0 wont have much impact on my performance. On the other hand at home the usage includes working with some very large graphic files along with the normal small read / writes, so RAID 0 would have a positive impact.
What does that all mean to me, the one trying to figure out what to buy? For work, I would opt for a system with more memory while at home I would give some thought to using a RAID 0 array. This just goes to show that not all solutions would be best for all users.
It is also interesting to look at the four 2.2 scores – 256MB, 512MB, 1.0GB of memory, and 1.0GB of memory with the RAID 0 array. Both the 2.0A with 512MB of memory and the 256MB of memory with a RAID 0 array have better scores – 5.21% for the 2.2 and 256MB vs. 6.38% (512MB) and 8.72% (RAID 0). But what is more interesting is the cost factor – the 2.2A with 256MB is 38.64% while the 2.0A with 512MB is 18.62% and the 2.0A with RAID and 125MB is 27.26%. So in the Winstone tests, adding more memory or using a RAID 0 array is not only more cost effective, but also gives better performance for this type of usage.
And lets not forget about the difference between using a 2.0 vs. a 2.0A – that extra L2 cache is worth almost 7% better performance, or about the same as using a 200MHz faster CPU. I’d definitely opt for the new Northwood core – the cost is the same, but the performance is better.
As I’m sure you have guessed, the iP4 2.2 (the fastest currently available) when coupled with 1.0GB of memory and RAID 0 provides the best performance of these configurations, at 26.98% faster than the baseline with a cost premium of 73.67%. This is not really too far out of line if that is what you need (or want).
Be the first to discuss this article!