What the hell, I'll bite.
Originally posted by: BenSkywalker
Is that a joke? Dealing with cache hierachy differences is why Cell is so rough? There are a lot of different ways that Cell is a bit different to deal with then LRB, but dealing with DMA level code is far more one of control versus laziness then extreme levels of complexity. It is more time consuming to manually handle it all, but it certainly helps avoid stalls from cache misses.
Have you ever even read the Cell documentation on how to do multi-threading? _It is insane_. And you can always tell the armchair-developers from the real deal when they tell you that "time consuming" is an A-OK trait for a development platform. TIME IS MONEY. "Cache misses"? These are video games. How many cache misses are you expecting? Think about how video games use memory for like ten seconds.
Larrabee is entirely in order execution units relying extremely heavily on vectorization to extract reasonable performance. The days of vomiting on your keyboard and having Intel's compilers make it run well are over.
Larrabee is a _GPU_, and it is not even pretending to be a CPU. This is _not_ the same as the Cell! I don't know how much more clear I can be.
And, again, typical armchair-developer talk. If the compiler can optimize, I damn well want it to optimize. TIME IS MONEY. Sony's compilers certainly optimize, I promise you that.
The programmers who do realize that need to rehearse this phrase over and over- "Would you like fries with that?". It will help them out tremendously within the next decade. Single core performance has ceased to progress in any meaningful way and there is nothing on the horizon that hints in any way whatsoever that is going to change. If you are a programmer and can't handle it, I reccomend you go back to school and find yourself a different profession, you aren't smart enough for the job moving forward. The current trend is going to change anytime soon. If you want to be a remotely decent programmer you learn to work in the new model. Crying that you suck too much to do the job isn't going to get you anywhere. Crying that someone needs to come up with an easier solution is a good way to land yourself with a top notch cell phone app job until they go mutli core not all that long from now.
You seem to not even understand what's being talked about here. Writing multi-threaded code is hard on an SMP architecture. Writing it on something like the Cell, which is asymmetric, is even more difficult. Unless there is some sort of compelling performance reason for this, and best I can tell, there is not, it was a dumb design decision.
And, just FYI: I used to write soft real-time space flight simulators for NASA in my last job. Each component was multi-threaded, and had components running simultaneously across multiple systems. I have a bit of expertise with writing high-performance multi-threaded code. What about you? Seriously, do you even have a degree in CS? (I do, and from a highly-ranked school, too.) I really don't appreciate being told I'm incompetent by someone who can't even understand what's being talked about.
And how do those programming challenges compare to the laws of physics? Since more talented coders have already been managing to handle extracting parallelism from game code without too much of an issue, even on Cell, I would say that it is more reasonable to assume that top tier coders are going to figure out how to program effectively under the new model of architectures then any of the processor manufacturers are going to figure out how the laws of physics don't apply to them.
You seem to not comprehend that "programming for the Cell" is not the same problem as "writing multi-threaded code". If the latter were so hard, the 360 would be experiencing myriad performance issues. Interestingly, it's not.
Why would they cripple their system with a wannabe Cell instead of using Cell with a GPU? Actually, in your world they should probably use a Pentium4, it was easier to develop for then anything that came after it.
Because, um, Larabee is a GPU, and the Cell has already proven itself not to be? Honestly, did you even read what you were responding to before responding in a fanboy rage?
Since the top tier developers all seem to be having no problem with the current setup, I don't think that is likely to happen. If they wanted to target the platform for the shittiest coders in the world they could have just thrown a single core x86 in there and been done with it, clearly that wasn't their goal.
Ah, so the definition of "top-tier" is reduced to "people who write reasonably-performing PS3 games". That's a definition that only a fanboy could love.
I feel like you didn't even comprehend what I wrote, which is starting to not surprise me. SPEs are hard to fill, because they're so specialized in what they do well. PPEs are generalized, and thus easier to fill. If programmers are finding that the ratio of SPEs to PPEs is too heavily weighted towards SPEs, it's time to adjust that ratio. Why do you think Sony has a magic crystal ball that told them that 7:1 is the right ratio?
Unchartered2, KZ2, GT5. We know the 360 has a stronger GPU then the PS3, there isn't really an argument on that one. Why is the PS3 offering flat out superior visuals in games? How can it be possible? How is it that Sony is doing things in games that the 360 can't match?
Funny that your examples of "unmatched visuals" are games that haven't even come out yet. Let me know how reality matches up with bullshots.
Facebook
Twitter