parallel roads

It’s been a very long but great day today at PDC. I talked earlier in the day on the Parallel Pattern Library, Asynchronous Agents Library and Concurrency Runtime and demonstrating the CTP of Visual Studio 2010.   Hazim Shafi also spoke on multi-threaded and parallel performance.  I spent the latter portion of the day talking to developers at the booth we’re sharing with the High Performance Computing team.

There will be several more talks from the Parallel Computing Platform team at PDC this week, our architect Niklas will be doing a deep dive on the Concurrency Runtime and discussing extensibility, Daniel Moth will be talking soon about the Task Parallel Library and PLINQ and there’s several Symposium sessions lined up.

While I’m here at the PDC I’ll post on my personal blog summarizing experiences and I’ll also try to get a few more examples posted either here or on our team blog which demonstrate the functionality included in the CTP.  That’s it for now though, it’s been a long day. 

-Rick

Continued from "Is 4 cores really the end? (Part 1)".  I see Jeff has posted another blog in response and it looks like he noticed the application  I use all the time…   paint.net.  I use this a lot because it’s free, multi-threaded and it also happens to be open source.  As I mentioned, most of the filters are multi-threaded and here’s a trace from one, it’s got all 4 cores pegged at 100% for about 7 seconds. I didn’t try hard to find this at all, I just picked a blur filter that I use frequently and ran it…

I’m afraid that since it’s Saturday and I’ve a full schedule, this post is also going to be a bit short, but I will come back and add some more traces over the weekend.

I often read Jeff’s popular blog coding horror and in fact I have every intention of continuing to read it because I often find his blog astute, intelligent and enjoyable to read.  I also often agree with his sentiments, but his post today is something that I have to quite emphatically and respectfully disagree with because it borderlines on disinformation.  

I’ve in fact drafted this post about 3 times now trying to state my disagreement concisely and effectively.  So here goes:

those four cores provide almost no benchmarkable improvement in the type of applications most people use

Let’s take a look at some benchmarks of just a few applications that I think most people use.  I’m going to use my favorite profiler, ‘xperf’ low overhead, accurate and provides a good view of system performance and it’s free.

Quad Core vs Dual Core on some real user scenarios…

Jeff is using a dual core CPU overclocked to 4 GHz.  I have a quad core that’s overclocked to 3GHz, so as near as I can tell as long as I can come up with profiles of several common user scenarios that use more than 75% of the CPU (9GHz > 8GHz) for any measurable length of time then I’ll claim there’s user benefit.

To be continued…