Slashdot Mirror
Sun's Zippy New Chips
Mark the Revelator writes: "Reuters has a story about Sun unveiling it's latest and greatest UltraSparcIII chips. The new chips are being made by TI and are the first UltraSparcs to use copper instead of aluminum for transistor connections. Although they're supposed to compete with Intel's Itanium chips, they only run at 900MHz ... for now."
Ars Technica has a fantastic article comparing and contrasting Sun's future MAJC (Microprocessor Architecture for Java Computing) CPU architecture with Intel's IA-64. It's going to be very interesting to see if Sun can carve out a large enough market to ensure MAJC's viability. My uninformed opinion after reading the article--Sun has been making decisions since its founding that have given it the only chance to survive. By almost totally eschewing both Intel and Microsoft, Sun has been forced to innovate on both hardware and software to compete with these giants. Sun simply had to invent Java--what was the alternative, reselling NT "workstations"?! Now Sun has leveraged Java into strategic partnerships with IBM, Oracle, etc. to create from scratch a major software niche, not to mention Java's future in the embedded markets. MAJC it seems to me is the logical step in hardware once Sun made the commitment to Java and once Sun decided not to become a reseller of Intel chips like say HP. Without having to worry about what Intel wants, Sun can use its traditional RISC approach to registers to once again offer a fantastic alternative--read the Ars Technica article cited above: "MAJC, however, spends so much of its die space on registers that it can have the register states for four different threads loaded at once. Since it doesn't have to save and load register states to switch between threads, its context switches are very fast". In the 1980s HP saved the company investing in PA-RISC. Maybe that was because the engineer founders David Packard and Bill Hewlett were still alive and strong. I believe that it is Sun that has applied that lesson of not surrendering control over the CPU architecture, and that HP will continue to pay a heavy price for deciding to go with Intel. Financing new chip architectures is difficult, but in my opinion there is no future for being a reseller unless one is IBM or Dell. (And note that IBM resells only because it wants to since it already manufactures alternatives, it is beholden to no one. Just who will be able to compete with IBM's Global Services?)
Doesn't apply well at all. They're all RISC now, but with CISC wrappers on some, you decide which. BTW: The Sparc III, code name Cheetah, is designed with multi-processor scalability in mind, including local and global cache coherency. Its databus is 128bits wide and I hear from reliable sources so close, that the 1.2 Ghz is only limited by product yield.
> keep in mind that these are pure RISC processors
... have always toasted any CISC or CISC-to-RISC" is a solid lie. There are plenty of occasions where a CISC processor outperforms a RISC processor.
> and have always toasted any CISC or CISC-to-RISC
> processor of a much higher processor rating.
That is misleading and, in fact, bordering on the level of a total lie. The benefits of RISC architectures are not performance. They're simplicity. This simplicity, in the past, sometimes had the benefit of increasing performance, but higher performance is not a rule in and of itself.
Saying that "pure RISC processors
In specfp_2000, the lowest frequency Pentium 4 scores a 516, while the highest frequency UltraSPARC III scores a 482. The slowest Pentium 4 is 7.1% faster than the fastest USPARCIII.
In specint_2000, the slowest P4 gets a 490, while the fastest Sun processor gets a 467. Here, the wimpiest current generation Intel processor is 4.9% faster than the best thing Sun can offer.
These above factors keep in mind that the Sun chips are *specifically* architected to achieve the highest performance possible, pretty much regardless of cost. They are full-on server chips. The Intel Pentium 4 series are designed with cost factors in mind. The Pentium 4 cannot be a three thousand dollar behemoth due to its target market (actually, the 750MHz USPARCIII processor module costs about $7k on Sun's website). So the USPARCIII can have the benefit of loads of added performance enhancing features while the Netburst (P4) architecture has to cut corners at every step.
the UltraSPARC III outperforms the Pentium 4 on a clock for clock basis. Of course, the original Pentium outperforms the Pentium 4 on a clock for clock basis on many benchmarks, too. This means nothing. It is merely reflective of different design strategies. I can easily point out the fact that the Pentium 4 offers higher performance per watt or higher performance per number of integer ALUs. But, like "performance per megahertz", those are also stupid measurements.
There is nothing out there which would cause me to believe that an x86 processor made with the design strategy of the UltraSPARC III ("we're gonna sell this for thousands of dollars, so throw in the kitchen sink, too!") would not outperform the UltraSPARC III at like frequency. Well, except if the the fp instructions on the USIII are three operand, but that's a special case. ^_^
-JC
http://www.jc-news.com/
Is this a good idea though? I mean, using one of today's compilers, ported to a IA64/Itanium architecture, a compiled program might run very slowly, since today's compilers probably let a bit of the optimization (within reason) up to the CPU. This would also mean that it may be a little while until some quality IA64 compilers are released. Or am I misinformed?
Where are we going and why am I in this handbasket?
I don't know what APPLICATIONS you people are running but it seems all you are comparing are SPEC marks. We run Cadence Verilog and Synopsys synthesis on our Dell PIII 1Ghz machines and they blow the frigging doors of all of our SPARC's. For IC design applications at least 64Bit vs 32 Bit is a non issues as most of the apps are not compiled for 64bits. Performance is mostly just CPU clock speed so a 1Ghz PIII runs at just over 2x a Sun Ultra 450. Given that there is about a 3-4x difference in cost and about a 2x performance bonus you get about a 6x-8x cost/performance benifit. I'll take Intel machines all day long. Sun better wake up or they are going to lose the CAD market. Course, I don't think they really care about that one anyway.
"TV, a medium as it is neither rare nor well done." Ernie Kovacs
It blows any of the pentium-based machines we have here out of the water.
-fialar
From my point of view, the Iantium shares a lot of similarities with the iAPX432 of 20 years ago. Both are new architectures that purport to emulate the previous technology, but from all reports at least, don't do that emulation very well. Both rely on software technology that exists only in the laboratory, if there.
Only time will tell if both share the same marketing fate.
Attempting to measure how fast a computer can go by its CPU's clock speed is tantamount to measuring how fast a car can go by its engine's horsepower. There are many more factors at play here.
Let's start with the whole RISC vs. CISC thing. Everyone knows that RISC is more efficient; the only thing that has kept CISC alive this long is backwards compatibility with the Wintel juggernaut. You develop a lean, efficient instruction set, then you write compiler back ends that take advantage of it.
Also keep in mind that Sun's motherboard designs are true performers. The path between the CPU, memory, and bus are designed to move data around in ways that just aren't possible with Intel.
Did you know that SPARC is more or less an "open" CPU design? It was designed to be a multi-vendor instruction set, one that would be 'common' without having one vendor calling all the shots. Read www.sparc.org for more details.
Tired of FB/Google censorship? Visit UNCENSORED!
There was something linked from solaris central about how the Ultra III was for massively multi-processor machines, not workstations. The Ultra IIIi will be the workstation processor clocking in(I think) at well over 1GHz. I'd provide the link, but I can't remember it.
-blah
The following sentence is TRUE. The previous sentence is FALSE.
Actually, this will happen. Current processors are designed so that with each cycle they load and decode the instructions they're going to execute (and, of course, the data the instruction is going to work on). When the instruction is completed, it's thrown away. This is highly inefficient for loops, because the same code is loaded again and again. Think about audio or video decoders -- the same decoding instrcutions are reloaded all the time.
Future processors will, at least partly, be reconfigurable, that is they will load a set of instructions and save it, and then have to load the data only. This is supposed to be the optimum between a hardware-only implementation (fastest, but can't change when, for example, encodings change) and current "software-only" implementations (most flexible, but processors must (re)load instructions in each cycle).
Take a look at, for example, PACT if you're interested in this technology, they're one of the companies developing such processors.
Sig (appended to the end of comments I post, 54 chars)
PA-RISC chips may have been fast, but the cost is outrageous. Last year, my company put 4 more 552 MHz CPU's into our N-Class, at a cost of $20,000 each. Figure out the price/performance ratio on those and it is nowhere close to Intel or AMD cpu's. But, non-techie managers don't understand such things. They'll buy 1 HP cpu instead of 10 Intel cpu's because HP hardware must be better... I'm not too disappointed that HP is trying to switch to Intel processors, especially if it will make them more price competitive.