"At the time of its release in December of 2005, a single chip, eight core, 32-thread, 1.2 GHz UltraSPARC T1 server performed similarly to a two-socket, four-core, eight-thread, 1.9 GHz IBM POWER5 server, performed similarly to a four socket, eight-core, sixteen-thread 3.0 GHz Intel Xeon "Paxville MP" server, and exceeded the performance of a four socket, four-core, four-thead 1.6 GHz Intel Itanium server. Arguably, this made the UltraSPARC T1 the world's most powerful general-purpose commercial server processors, when considering multithreaded commercial workloads. ... One customer has published results showing that a MySQL application running on an UltraSPARC T1 server ran 13.5 times faster than on an AMD Opteron server." -- http://en.wikipedia.org/wiki/UltraSPARC_T1
UltraSPARCs are rather specialised for general purpose processors - one could make the arguments that GPUs, which massively outperform x86 CPUs several times over in terms of FP performance, were faster even than SPARC, but it's very much a different beast.
Also I haven't looked this up but isn't UltraSPARC much more costly?
Given the rising cost of energy, companies include it in the cost of ownership. So, if you had a server which was 13.5 times more powerful and consumed the same energy then you'd definitely save money. Unfortunately, SPARC would cost more up-front.
So, for a fast growing company, using x86 may reduce opportunity cost. For a large, steady company, a rolling replacement programme of SPARC may be the best option.
I am aware of Niagara, and I am skeptical that it will beat Nehalem. Certainly a Nehalem system will beat a Niagara 2 system in most benchmarks for performance and price.
6 comments
[ 4.3 ms ] story [ 32.2 ms ] threadIn October 2007, the T2 made significant improvements ( http://en.wikipedia.org/wiki/UltraSPARC_T2 ) - most notably a 10-fold improvement in floating-point throughput.
Also I haven't looked this up but isn't UltraSPARC much more costly?
Basically: Apples and Oranges etc.
So, for a fast growing company, using x86 may reduce opportunity cost. For a large, steady company, a rolling replacement programme of SPARC may be the best option.