System News
 1 2 3 4 5 6 7 8 9 10 next
Articles for the keywords: SPARC T3
27 Apr 2015 Oracle SPARC Processor Documentation [42275]
Includes SPARC Legacy Processor Documentation

Documentation for SPARC Legacy Processors has now been posted on-line.

  • Ultra SPARC Architecture 2007 (privileged edition)
  • SPARC T3 Implementation Supplement
  • UltraSPARC T2 Implementation Supplement
  • UltraSPARC Architecture 2005 (privileged edition)
  • UltraSPARC T1 Implementation Supplement
  • Joint Programming Specification 2
  • UltraSPARC-IV User's Manual
  • UltraSPARC-IV+ User's Manual
  • Joint Programming Specification 1
  • UltraSPARC-IIIi User's Manual
  • UltraSPARC-III User's Manual
  • UltraSPARC IIi User's Manual
  • UltraSPARC-IIe User's Manual
  • UltraSPARC User's Manual, UltraSPARC-I, UltraSPARC-II

(PDF documents)
(Get More Information . .) open to premium members only

04 Sep 2012 The Future of the SPARC T5 in Oracle's Big Iron Lineup [27510]
Timothy Prickett Morgan Reports from Hot Chips

Reporting on this year's Hot Chips event in Cupertino, CA, Timothy Prickett Morgan speculates on the future of the SPARC T5 chip in Oracle's apparent growing interest in big iron featuring multithreaded processors in terms of cores and sockets. In his view, the scalability enhancements coming from the SPARC T5 machines, sporting faster cores that can do superior single-threaded work, will be big enough for all but the absolutely largest Solaris workloads. Prickett Morgan devotes a substantial part of his article to looking at the history of the SPARC T3, T4 and T5 chips and their family resemblances.
(Get More Information . .) open to premium members only

19 Mar 2012 How Memory Allocation Affects Performance in Multithreaded Programs [25732]
Sniffing Out Lock Contention with Several Tools

If your new server has an alarmingly lower transaction rate than you had anticipated, Rickey C. Weisner suggests that you look first at lock contention in the memory allocator. His article demonstrates the use of prstat to determine whether there is lock contention. Demonstrated as well is the use of Professor Emery Berger's MT-hot allocator named Hoard, and of the mtmalloc (both old and new), and the libumem memory allocators. The question of which is best is answered by Weisner's use of a test harness, which appears to give the edge to the new mtmalloc.
(Get More Information . .) open to premium members only

16 Jan 2012 Netra SPARC T4-2 SPECjvm2008 World Record Performance [25258]
The Netra SPARC T4-2 server demonstrates 41% better performance than the SPARC T3-2 server

Oracle's Netra SPARC T4-2 Server equipped with two SPARC T4 processors running at 2.85 GHz set a world record result of 454.52 SPECjvm2008 Peak ops/m on the SPECjvm2008 benchmark. This result surpassed the previous record, which was run on a similar product, Oracle's SPARC T4-2 server. This level of performance is 41% better than the SPARC T3-2 server and similar in performance to Oracle's SPARC T4-2 server. The Netra SPARC T4-2 server with hardware cryptography acceleration greatly increases performance with subtests using AES and RSA encryption ciphers. There are no SPECjvm2008 results published by IBM on POWER7 based systems.
(Get More Information . .) open to premium members only

14 Nov 2011 New World Record for SPARC T4-2 on SPECjvm2008 Result with Oracle Solaris 11 [24934]
Beats SPARC T3-2 Record by 41%

The SPARC T4-2 Server equipped with two SPARC T4 processors running at 2.85 GigaHertz (GHz) and using Oracle Solaris 11 and Oracle JDK 7 Update 2 set a world record of 454.25 SPECjvm2008 Peak ops/m on the SPECjvm2008 benchmark, a benchmark suite for measuring the performance of a Java Runtime Environment (JRE). This level of performance is 41% better than the mark set by the SPARC T3-2 Server. The hardware cryptography acceleration built into the SPARC T4-2 was found to greatly increase performance on subtests using AES and RSA encryption ciphers.
(Get More Information . .) open to premium members only

 
 1 2 3 4 5 6 7 8 9 10 next






!-- end archive_section.tpl -->