In die­ser Sek­ti­on wer­den Pres­se­mit­tei­lun­gen der Her­stel­ler 1:1 ver­öf­fent­licht. Pla­net 3DNow! über­nimmt weder für die Rich­tig­keit der Anga­ben noch für die Art der For­mu­lie­rung jed­we­de Ver­ant­wor­tung. Die Anga­ben kön­nen sub­jek­tiv sein und geben nicht not­wen­di­ger­wei­se die Mei­nung der auf Pla­net 3DNow! schrei­ben­den Autoren wieder.

PGI 2014 Adds OpenACC 2.0 Features for NVIDIA and AMD GPU Accelerators

Verfasst vonDr@

SANTA CLARA, Calif., Feb. 14 – PGI, a lea­ding suite of high-per­for­mance par­al­le­li­zing com­pi­lers and deve­lo­p­ment tools, now fea­tures sup­port for the latest ver­si­on of the Open­ACC pro­gramming stan­dard on acce­le­ra­tor platforms.

Available today, PGI 2014 Com­pi­lers and Tools includes new capa­bi­li­ties for pro­gramming the recent­ly announ­ced NVIDIA Tes­la K40 GPU acce­le­ra­tors using ver­si­on 2.0 fea­tures of the Open­ACC direc­ti­ves-based par­al­lel pro­gramming spe­ci­fi­ca­ti­on. It also pro­vi­des, for the first time, Open­ACC sup­port for AMD Rade­on GPUs and APUs.

“We applaud PGI’s abili­ty to extra­ct per­for­mance from AMD dis­crete GPUs and APUs using Open­ACC,” said Suresh Gopa­l­a­krish­n­an, cor­po­ra­te vice pre­si­dent and gene­ral mana­ger of the Ser­ver busi­ness at AMD. “It will help break down the remai­ning bar­riers to wide-sca­le acce­le­ra­tor adop­ti­on, and decou­ple the choice of acce­le­ra­tor pro­gramming model from the choice of acce­le­ra­tor hardware.”

Key fea­tures of PGI 2014 Com­pi­lers and Tools include:

  • Open­ACC 2.0 Fea­tures – PGI Acce­le­ra­tor nati­ve For­tran 2003, C99 and C++ com­pi­lers expand sup­port for key Open­ACC 2.0 fea­tures, inclu­ding rou­ti­ne direc­ti­ve (pro­ce­du­re calls in acce­le­ra­tor regi­ons), unstruc­tu­red data life­ti­mes and others.
  • New NVIDIA CUDA For­tran Exten­si­ons – Add sup­port for ver­si­on 5.5 of the NVIDIA CUDA par­al­lel pro­gramming plat­form, CUDA ato­mic func­tions and device-side debug­ging using Alli­nea DTT and Total­View from Rogue Wave.
  • Free PGI for OS X – For­tran 2003 and C99 com­pi­lers with all PGI mul­ti-core x64 opti­miza­ti­ons, com­mand-line debug­ging and stream­li­ned online docu­men­ta­ti­on (available in Febru­ary). 

PGI 2014 com­pi­lers deli­ver an avera­ge of 75 per­cent fas­ter per­for­mance on the latest SPEC OMP2012 bench­mark suite, com­pared to GCC using the latest AVX-enab­led mul­ti-core Intel and AMD x64 pro­ces­sors. Addi­tio­nal capa­bi­li­ties of PGI 2014 Com­pi­lers and Tools include full For­tran 2003 sup­port, incre­men­tal For­tran 2008 fea­tures, updated libra­ri­es, sup­port for the latest ope­ra­ting sys­tems and a com­pre­hen­si­ve suite of new and updated code examp­les and tutorials.

For a com­ple­te list of the fea­tures and capa­bi­li­ties of PGI 2014 Com­pi­lers and Tools, visit http://www.pgroup.com/support/new_rel.htm.

“The use of acce­le­ra­tors in high per­for­mance com­pu­ting is now main­stream,” said Dou­glas Miles, direc­tor of PGI Soft­ware at NVIDIA. “With PGI 2014, we are taking ano­ther big step toward our goal of pro­vi­ding plat­form-inde­pen­dent, mul­ti-core and acce­le­ra­tor pro­gramming tools that deli­ver out­stan­ding per­for­mance on mul­ti­ple plat­forms wit­hout the need for exten­si­ve, device-spe­ci­fic tuning.”

PGI 2014 is available today direct­ly from NVIDIA and aut­ho­ri­zed resel­lers. A free 30-day tri­al of PGI 2014 is available for new users at www.pgroup.com. Regis­tra­ti­on is required.

About PGI Software 

An NVIDIA Cor­po­ra­ti­on brand, PGI Soft­ware includes high-per­for­mance par­al­lel For­tran, C and C++ com­pi­lers and tools for work­sta­tions, ser­vers and clus­ters based on x64 pro­ces­sors from Intel and AMD, and HPC acce­le­ra­tors from NVIDIA and AMD. More infor­ma­ti­on is available at www.pgroup.com, sales@pgroup.com or by cal­ling (503) 682‑2806.

About NVIDIA

Sin­ce 1993, NVIDIA has pio­nee­red the art and sci­ence of visu­al com­pu­ting. The company’s tech­no­lo­gies are trans­forming a world of dis­plays into a world of inter­ac­ti­ve dis­co­very — for ever­yo­ne from gamers to sci­en­tists, and con­su­mers to enter­pri­se cus­to­mers. More infor­ma­ti­on at http://nvidianews.nvidia.com and http://blogs.nvidia.com.