[zweben]

I'm one of the few non-programmers here, and I figured this would be a good place to ask: Right now, few applications are coded to utilize so many cores. Is this simply a matter of programmers transitioning to coding for multi-core computers, or are some types of software not good candidates for taking full advantage of so many cores?

I have an 8-core Mac Pro, and I was hoping to see its performance improve over time as software took better advantage of the hardware, but I don't see that happening yet.

[angstrom]

Both, not all tasks are equally parallelized. Some, like the graphics rendering pipeline, data mining, search indexing, and others can reap the benefits and already do. However, there are many tasks that don't parallelize as easily and lead to the programmer waiting for one synchronous task after another due to a dependency. Even if you did run the tasks in parallel the one would need to block and wait until the other completed.

In some cases the programming languages are just ill suited to handle performing concurrent operations free of side effects. Programs written in the languages would have to be rewritten (not likely for many apps). Parallel Studio, by Intel, attempts to make the task easier on programmers working in c++/Windows to identify such cases. Still, others have proposed and attempted to create a layer beneath the programming language that automatically detects code that can be safely executed concurrently without side effects. There's still no silver bullet that's suddenly going to give significant gains without modifications to existing code at this point.

[buro9]

I have an 8-core HP Z800 with 24GB RAM.

For me it's not about improving the speed of end user applications.

My usage is simple, virtualisation.

I have 6 VMs currently fired up, 1 with Oracle installed, a couple of Linux boxes to run an ESB and the remainder run SharePoint on Windows Server.

I wanted each of those to perform pretty well and not be a dev bottleneck (waiting for stuff to happen) and so I've assigned at 1+ cores per VM as well as at 2+GB RAM per VM (Oracle gets more of both). I try and balance my handing out of hardware resource to reduce the amount of scheduling that the system will have to perform between virtual machines... hence, I wanted as many cores as possible and a good chunk of RAM to go with it.

It was cheaper to buy a single high powered workstation than it was to buy 6 cheap servers. Another huge factor was that the running costs (power) and environment (heat generated in the front room + the cabling and physical space + volume in decibels) of a single workstation beat those 6 cheap boxes.

[EvilTrout]

This is a good question and the answer is it's both.

It requires a completely different approach to programming so many programmers haven't made the leap. Additionally, some kinds of software problems seem to be unparallelizable by default.

Having said that, there is an advantage to a many-core machine right now: you can run many tasks (programs) simultaneously. Even if a program can only use one core, your OS should be smart enough to run it on its own core.

Of course then they might contend for other resources, like disk or video throughput.

[_delirium]

One of the bigger traditional market segments for these high-end Apple desktop workstations is graphics and film processing, and the tools there have made some progress on parallelization. A bunch of Photoshop transforms and film-editing tools will happily eat up any cores you throw at them (although Photoshop has also been increasingly moving stuff to the GPU as well).

[gte910h]

Multicore programming is expensive, and in violation of some very common paradigms and libraries. It takes lots of work to accomplish multicore work in a program and actually gain performance