The Quest for More Processing Power, Part Two: "Multi-core and multi-threaded gaming"
by Johan De Gelas on March 14, 2005 12:05 AM EST- Posted in
- CPUs
Conclusion
Writing multithreaded code means much higher software development costs while CPU development gets easier and thus cheaper (compared to even more complex superscalar CPUs). No wonder that the CPU developers are very motivated to hype the multi-core route, but the software development community is probably less enthusiastic.Intel and other manufacturers should not simply push the costs of getting higher performance onto the software developers. Because, in the end, it will be the consumer who will pay the final price: either more money or buggier software with more crashes and hangs. One way that Intel and others can help to keep multithreaded development costs under control while offering increasing CPU performance is to keep investing in ILP and thus higher IPC cores; another option is to improve the interCPU communications.
The easiest part of multithreading is using threads that are running completely independent, that don't share any data. But this source of threading is probably already being used almost to the fullest. In order to tap into a new source of multithreading, such as the largely unused potential of multithreaded AI, Phyics and animation, it is important that developers don't have to worry about interthread messaging and synchronization lowering performance.
Very fast interprocessor communications to make sure that thread synchronization comes with little overhead will give a bigger incentive to developers to invest the extra time in multithreading.
Gabriele Svelto:[4]
"Most of the current multi-threaded software is developed with an eye at keeping inter-thread messaging and synchronization as low as possible because both have a significant cost. This cost will be lowered by an ordered of magnitude by multiple cores on a single die giving in turn more flexibility for the programmers.The Pentium-D and Pressler are examples of how not to do it: just slap two CPUs on the same die and call it a day. High clocked single cores like the upcoming Athlon 64 FX-57 will eat these massive chips for lunch in almost all benchmarks while consuming less energy. With the exception of some special far-fetched benchmarks, it will be pretty hard to justify the reason behind these dual cores.
Applications which got low speed-ups by going multi-threaded due to the overhead of fine-grained locking mechanisms will be able to exploit multiple-processors with fast interprocessor communications much better."
Luckily, Intel's Yonah and AMD's Dual Athlon 64 cores show that better multi-core CPUs are on the way. At that point in time, we are entering the multi-core engine for real. And we can only applaud that because it unleashes a massive amount of CPU power upon the developers.
References
[1] Intel multi-core briefingStephen L. Smith, Vice President Digital Enterprise group, IDF Spring 2005
[2] Unreal 3 engine
http://www.unrealtechnology.com/html/technology/ue30.shtml
[3] Galactiv Cilivisations
http://www.galciv.com/index.asp?c=1&u=0
[4] Gabriele Svelto on dualcore CPUs
http://www.aceshardware.com/forums/read_post.jsp?id=20270&forumid=2
Facebook
Twitter
RSS
49 Comments
View All Comments
Pjotr - Monday, March 14, 2005 - link
"dual-cored GPUs are stupid. given the parallel nature of graphics, it makes more sense to just add another pipeline at very little design cost."Unless you hit a power and/or heat output wall.
Tell nVidia that parallell GPUs are bad, they alreay sell their SLI solution for dual-GPU computers.
silverwolf - Monday, March 14, 2005 - link
PPU, is the way to go.defter - Monday, March 14, 2005 - link
"Given the immense complexity involved, I expect dual cores taking a VERY VERY long time to catch on... even then it'll be a half assed job."Well ALL future consoles will use multi core CPUs. Thus if developers want to sell games, their games must take advantage of at least two cores :)
ceefka - Monday, March 14, 2005 - link
#12 That's how it looks, for now. "No one will ever need more than 8 cores." :-DMy dumb question, I was reading:
"Tim clearly emphasizes that only parts of the application can be economically parallelized. Increasing parallelisation, using more threads, is simply not feasible. There is a pretty hard economic limit to TLP."
Isn't a high IPC-count also a form of parallelism? If so, then beyond a certain count won't it be just as hard to take advantage of a high IPC-count.
sandorski - Monday, March 14, 2005 - link
Good article. Most of it was over my head, but the gist was most important. That being that Multicore is a big question mark for Gamers and other common users.I've always preferred Sweeney over others in the industry, he knows what he's doing without getting in everybody elses face about it. I also found it appropriate that he was interviewed on the subject since Unreal Engines have always internally manged 100's of Processes in order to work(I assume other Engines do the same, but my knowledge of hte Unreal Engines is more thorough than them).If anyone can figure out how to use Multicore in gaming my money's on Sweeney.
Calin - Monday, March 14, 2005 - link
Auto-parallelization is of limited use, and it can work only on small pieces of code. You might get a couple of percent extra speed, but no more (or no much more). Managing multi-threaded code interdependent code would be a nightmare.However, some "extra" speed can be recovered in case of multi processors (or multi core) from the reduced state (thread) change. Certainly not extra 24%, but more than a bit.
FDSatyr - Monday, March 14, 2005 - link
Good read - mainly for Tim's comments though! I really enjoy the way Tim isn't arrogant at all in the way he talks. Some fairly silly questions from A though! I still think threads are rubbish, that processes and better schedulers are the way forward. I think the next step - realistically impossible in the industry it may be - would be to create a fresh architecture, and put an x86/87 core on the same die. Ho-hum.mkruer - Monday, March 14, 2005 - link
My magic eightball says that after 4-8 cores, any other core that will be added will be near worthless.AnandThenMan - Monday, March 14, 2005 - link
Anand got owned by Tim Sweeney.AnandTech: Did you make use of auto-parallelisation compiler technology...
Tim Sweeney: Auto-parallelization of C++ code is not a serious notion.
Good article though.
overclockingoodness - Monday, March 14, 2005 - link
Oh my God - the Unreal 3 engine is beautiful. I get amazed everytime I look at it. I can't wait for games to be featured on Unreal 3 and the like engines in the future. So fine...