Author: Robert Hyatt
Date: 19:03:07 01/09/04
Go up one level in this thread
On January 09, 2004 at 14:02:54, Tom Kerrigan wrote: >On January 09, 2004 at 13:00:39, Robert Hyatt wrote: > >>On January 09, 2004 at 12:24:47, Tom Kerrigan wrote: >> >>>On January 09, 2004 at 10:13:39, Robert Hyatt wrote: >>> >>>>On January 09, 2004 at 00:29:51, Tom Kerrigan wrote: >>>> >>>>>On January 09, 2004 at 00:20:48, Robert Hyatt wrote: >>>>> >>>>>>On January 08, 2004 at 23:39:42, Tom Kerrigan wrote: >>>>>> >>>>>>>On January 08, 2004 at 22:50:32, Robert Hyatt wrote: >>>>>>> >>>>>>>>>Where is this going on? Aren't all of the 64 bit Athlons named "Athlon 64"? >>>>>>>> >>>>>>>>They are not being called that everywhere. IE 3400+ what? >>>>>>> >>>>>>>I think you're confused about AMD's product line. 3400+ is the speed rating. >>>>>>>Athlon 64 is the family name. The full name of the chip is "Athlon 64 3400+" so >>>>>>>I don't think you have a valid complaint. >>>>>> >>>>>>athlon. athlon-64. opteron. FX51. And there is "no confusion"? >>>>> >>>>>I'm not confused. Honda makes the Accord, Accord Coupe, Acura TSX, and Acura TL >>>>>and they're all based on the Accord, yet somehow people manage... >>>> >>>>Not all of those are "accords" with different names, however. >>>> >>>>FX51. athlon-64. Opteron are the same with a minor memory controller >>>>difference for one. >>> >>>Not really. The A64 3000+ has a 512k cache, vs. the others' 1M cache. The A64's >>>"minor" memory channel difference is a complete lack of an entire memory >>>channel. And only Opteron can be used for multiprocessor scenarios. >>> >>>Let's compare this to Intel, shall we? Intel sells slightly defective Pentium 4s >>>as "Celerons." The only difference is that half the cache is disabled--how is >>>this different from disabling, e.g., multiprocessing on an Opteron and calling >>>it an Athlon 64? >> >>You are _totally_ missing the point. And I _do_ mean _TOTALLY_. >> >>Intel re-marks PIVs as celerons to (a) sell chips that might normally >>be tossed, or more commonly, to sell 'em at a price point that is cheaper >>than the normal PIV chip. >> >>Who cares? >> >>The point here _is_ that the opteron is really a well-done processor. yet >>AMD is burying flavors of it down in the Athlon product line. If they were >>dominating the market, that might make sense. But now, it would seem that it >>is the right time for AMD to market their chips as AMD, and not as "comparable >>to Intel's xxxx mhz processor." > >AMD has put a lot of money into the Athlon brand--why throw that away? Do you >also take issue with Intel because they keep naming their cores "Pentium" even >though that name only makes sense as a replacement for "586" and each successive >core is dramatically different and performs better? It's all about brand >recognition. > >As for the performance ratings, there must be market justification or AMD >wouldn't do it. I'm pretty sure Jerry Sanders didn't roll out of bed one day and >decide to annoy Bob Hyatt. I don't believe I ever said he did. I simply suggested that this is a _good_ time for AMD to exploit there _own_ name/processor, not exploit the fact that they are Intel-compatible... > >>This is about _marketing_. I suppose if they want to follow Intel until they >>die, that's ok. But _right now_ they actually have something that looks far >>better, yet they are marketing it as "as good as"... > >Per MHz, the Athlon is also far better than the P4, so why are you saying "right >now" specifically? No AMD product is far better than the P4 in terms of absolute >performance right now. The A64 is better at games, compiling, and some other >stuff, and the P4 is better at media encoding, 3D rendering, and some other >stuff. I can only give _my_ results, as I _always_ do. For Crafty, for a couple of other benchmarks I had the chance to run both on a quad-opteron (using 2 cpus for comparison) and on a dual 3.06ghz xeon running linux, the opteron made it no contest on _all_ the tests. Chess and ints (long ints of course). Floating point. Etc... I have not seen that gap on normal athlon vs xeon. IE when I got my xeon, nobody produced a faster Crafty nps except for Aaron and his overclocking... > >>are doing it. They own the high-end marketplace, they are waging a war on the >>low-end. AMD could make a lot of waves with the Opteron. But it is really >>very quiet. > >Are you kidding? There have been a tremendous number of news stories about the >Opteron and it seems like every time I visit a hardware page I see an ad for >Opteron or 64 bit AMD products in general. What would satisfy you? George W. >mentioning Opteron in a State of the Union? News Smews. Who reads news when they are hunting a new computer? I get asked this question _all_ the time by students here, so there is some confusion somewhere, even if it is not apparent where you are... > >>>Of course. I agree completely. That's why the 50MHz int/50MHz ext 486 is called >>>a 486/50 whereas the 50MHz int/25MHz ext 486 is called a 486DX2/50. There's no >>>"faking" of MHz numbers going on at all. >> >>Sure there is. The DX2 ran nowhere near 2x faster than the 486/25. It ran >>faster in some cases, but memory was a bottleneck and the external cache >>had to work to make it run faster at all, except for dinky programs. > >Huh. You're right. I suppose Intel should have gone with something like >performance ratings. Maybe 486 40+? Maybe just "486/25 with a new model number?" > >-Tom I'm not going to change my mind, you aren't going to change your mind. Neither really matters since this is all opinion anyway. If you like the current intel comparison processor IDs, fine. I don't. Particularly when I would advertise as "fastest thing around, and BTW Intel processors are compatible with ours so you can still run existing software while new software tailored to this chip will run even faster than on Intel."
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.