Feel free to dig for some info on the details of the "branch predictor" on the PPU, and then re-evaluate your comment, in terms of icache misses, especially since the PPU threads are sharing a cache.

If you argued that SPU branches tend to be faster because they never end up in a cache miss, because they SPU code is much smaller, better written, etc. than PPU code tends to be, I would have to agree with you.  That still says nothing about their ability to run general purpose code blazingly fast, especially when doing it in parallel.

I posted two links from authoritative sources which support my claims. You posted a sarcastic comment. What exactly are you trying to accomplish here?

@ Groucho

You're statements qualify as "nit-picking" in my book, and I don't feel like wasting my time digging for backup, when I know I'm right, except for a few irrelevant circumstances?

How is posting a source which contradicts your claim nitpicking?

If you say SPUs are better at branching than the PPU, you need extraordinary proof for that, considering that the PPU has special hardware to improve branch prediction that the SPU doesn't have. Are you suggesting that the PPU's branch predictor is a contrarian indicator due to bad design?

Your statements qualify as "wrong" in my book.

@ NJ5

True, but as you say that's something you can also do on the PPU, so Grouch still has to explain why his statement goes against what the Cell experts say.

NJ5,

Let me throw you a bone.

You are correct, in your nitpicking, in that a single SPU is, in some operations, slower than a single PPU thread (which is only 1.6 GHz, not 3.2 GHz).  I was thinking (sadly not the same as typing) in terms of.. multiple SPUs, but I can see that, from my wording, it is easily interpreted as though I was suggesting that one SPU was absolutely the supreme champion of the processing universe, when compared to the PPU.

I was oh so wrong.  

Good thing I have 6 SPUs to run general purpose code on, independantly of the PPU threads, eh?  I hope I have soothed your ravenous nitpicking for the day, because who knows what other insignifigant-to-the-discussion points you may be able to drudge up otherwise.

Untamoi said: afree_account said: According to TR, even a 2 years old core 2 duo tops at 200 gflops. http://techreport.com/articles.x/10351/4 Did you look at the side of the picture? Those numbers are mflops, not gflops. These again, these don't represent real theoretical maximum, only numbers from test program which is very unoptimized.

lol i did!

But i'm having a hard time with numbers.

Let me have a second try:

According to this chart the Core 2 X from 2006 tops at 2 gflops.

Pretty low, someone got another source ?

-ignore, double post.