Linked by Thom Holwerda on Wed 14th Nov 2007 19:49 UTC
Thread beginning with comment 284511
To view parent comment, click here.
To read all comments associated with this story, please click here.
To view parent comment, click here.
To read all comments associated with this story, please click here.
Regarding SPARC and POWER; it depends on what you mean by success or failure? people could look at PPC and claim, because Bob isn't running it on his desktop or that it doesn't dominate small to medium business server sales.
Niagara/Rock have an advantage over x86; SPARC64 IV for example is very competitive with the high end x86 chips put out by Intel and AMD; if you were talking about UltraSPARC, you would have a good case, but the SPARC64 is a fpu/int monster who has no problem holding its own - its just too bad it took this long for Sun to finally move to a superior SPARC implementation given that in the limit benchmarks of SPARC64 such as TPC, it comes out close to the top - especially unclustered.






Member since:
2005-07-08
Sun faces some serious challenges on the SPARC side of its business, especially in terms of performance per thread. I don't think it's feasible in the long-run for Sun to compete with AMD and Intel for serialized workloads. Sun's greatest hardware strength from its up-market pedigree is in system architecture, whereas it never moved the volume required to excel in the modern processor market.
I agree that Sun looks very strong at the moment in thread-dense boxes for transaction computing. However, both Intel and AMD have mini-cores coming in 2009-10 that will scale down to ultra-mobile and tile up to massively-multithreaded rack servers that will compete with Sun's Niagara/Rock descendants. On the other hand, I could definitely see Sun as the leading OEM for these Intel/AMD-based thread monsters.
I don't understand. There are very few operating systems and applications that don't run on x86, and plenty of them run on SPARC and POWER as well. There's no lock-in. The proprietary software vendors that ignore SPARC and POWER do so because the niche markets aren't worth their time and resources, not because of any especially onerous barriers to cross-architecture compatibility.