Author: Robert Hyatt
Date: 15:04:46 02/13/03
Go up one level in this thread
On February 13, 2003 at 14:50:09, Tom Kerrigan wrote: >On February 12, 2003 at 20:59:21, Robert Hyatt wrote: > >>No, but if the compiler does it perfectly, the OOO processor executes everything >>_in order_. Which is the point of the idea. OOOE lets the processor overcome >>bad instruction scheduling as produced by the compiler. If the compiler gets >>it right, OOOE means very little, unless you count X86 with the register issue. > >This argument isn't going anywhere. We don't have processors where we can turn >OOOE on and off (or at least we don't know how) and even if we did and OOOE >provided a significant speed increase, you would fall back on your >blame-the-compiler position. The UltraSPARC 3 is fast on paper except for its >lack of OOOE, and it performs like crap, but I suppose Sun's just been slacking >off on their compiler. Sun has been slacking off on their compiler _and_ their processor. The sparc hasn't been competitive in several years now... > >>>The point of predication is to eliminate dependency on a branch. How can a >>>compiler do this? In other words, how can a compiler say "we'll sort out later >>>which was crap and which was important" without a branch on a non-predicated >>>ISA? >>Easy. Do _both_ pathways at the same time, then (say) a cmov (yes, alpha >>has 'em as they came up with the idea in the first place) to pick the right >>one rather than branching at all... > >How is cmov not predication? I guess because it's not called "predication." Because it is not the same thing. It is just a simple scalar instruction that tests the flag register and moves the data if the right condition is set. The effect is certainly akin to predication, I would not argue that point... > >>I wouldn't mind a G4 at all. In fact, I tried to find a good multiple-cpu >>PPC machine when I bought my first quad p6/200, but could not find anything >>that was much beyond vaporware at the time. The processor looks pretty good >>to me however. > >I suggest you try using one. They're crap. > >-Tom It depends on your definition of "Crap". Crap because it is slow? Crap because it has a poor instruction set? Crap because the architecture itself makes programming cumbersome? IE back in the good old days, when the war was 8086 vs the MC6809, the motorola lost on speed, but had a better architecture, particularly when the 68000 came along. The PPC architecture looks to me to be reasonable. Whether the performance is or not is another issue of course.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.