Reprints from my posting to SAN-Tech Mailing List and ...


[san-tech][03259] Video, Slide: QLogic at ISC' 11, Tri-Laboratory Linux Capacity Cluster 2 (TLCC2) RFPとか

Date: Mon, 11 Jul 2011 17:21:45 +0900

"Video: QLogic Hot Seat ? Interconnect on the Path to ExaScale"
 07.9.2011, insideHPC

  "In this video from ISC’11, QLogic's Phil Murphy presents,
   "Interconnect on the Path to ExaScale." The fast-paced Hot Seat
   Sessions have been a trademark of ISC for many years, providing
   HPC vendors an opportunity to present their latest technologies
   in front of a board of inquisitors."

"ISC'11 Hot-Seat - Phil Murphy, QLogic: "Interconnect on the Path to ExaScale""
 2011/07/08, VideoISCEvents, 15:00, 480p

少し前の QLogic大型案件です:
"QLogic Wins Major Deployment in NNSA's Tri-Labs Cluster"
 Jun 17, 2011 
※これは、Department of Energy National Nuclear Security Administration
(NNSA's) Tri-Laboratory Linux Capacity Cluster 2 (TLCC2) への採用です。

"QLogic Grows InfiniBand Reach, Teams with NNS"
 June 20th 2011, SiliconANGEL
"QLogic Sees Supercomputers In The Cloud"
 June 17, 2011, InformationWeek

Tri-Laboratory Linux Capacity Cluster 2 (TLCC2) のサーバは前回と
同様に Appro社です (今回は、プロセッサ・マザーボード共にインテル)。

"Appro Nabs Exclusive Supercomputing Deal with Three US National Laboratories"
  ""Intel is excited to collaborate with Appro to deliver this innovative
   solution, including next generation Intel Xeon processor, code named
   Sandy Bridge-EP, as well as Intel's Server Board which is optimized
   for memory bandwidth performance and maximum density, with a flexible
   IO configuration." said Lisa Graff, vice president and general manager
   of the Enterprise Platforms & Services Division at Intel."

ちょっと驚いたのが Approが 1991年設立ということ
20 Year Anniversary

TLCC2開札に関する NNSAのプレスリリース:
"NNSA Announces Procurement of Capacity Computing Clusters to Support
 Stockpile Stewardship at National Labs"
 Jun 8, 2011
Stockpile stewardship

 2011年6月11日版、最近の話題、Ando's Processor Information

"Appro Comes Up Multi-Million Dollar Winner in HPC Procurement for NNSA"
 June 08, 2011, HPCWire

  "The new Tri-Lab clusters will be outfitted with QLogic QDR InfiniBand
   hardware, ditching the Mellanox parts in the TLCC and Peloton systems.
   In this case, the labs are favoring QLogic gear based on impressive
   scalability and performance results on some of their existing
   QLogic-equipped systems, in particular, the 23-thousand-core Sierra
   cluster at Lawrence Livermore."

"Appro wins 6 petaflops contract from DOE nuke labs"
 Largest deal in company history
 8th June 2011, The Register

なお、Tri-Laboratory Linux Capacity Cluster 2 (TLCC2) RFPは現在でも

Tri-Laboratory Linux Capacity Cluster
Request for Proposal B590550

Draft Statement of Work (SOW) (89 Page) は上記に含まれていますが、
最終仕様書は FBOでも公開されてませんね (調達登録はしてあります)。

Draft SOWによると Hyperionに近い気がします (例として Hyperion構成図を
使ってます)。Tri-Laboratory Operating System Stack (TOSS):OS環境に 

SLURM: A Highly Scalable Resource Manager

"single job scalability to 78,848 processors, and a compatibility library
 to support TORQUE job submission command syntax." (Page 27)
で 1セットの最大規模を推定出来ます。スケジューラは Moabです。

3 TLCC2 Technical Requirements (Page 33)

"These combined SUs shall be capable of supporting a complex workload
 consisting of small (4-256) medium (257?2,048), large (2,049-16,384)
 and occasionally full capability (78,848) MPI task count parallel jobs
 for Tri-Laboratory classified ASC Program and SSP simulations."

とありました (でも↑では processors、ここでは MPI task count parallel
jobs、processors = coresと考えられますが、後から、processor socket、
processor coresという表現が出て来て、定義してくれないと困ります)。

定義ありました (Page 85)、CPU = core = processorです。
単品は VLSI chipと定義してますね。私はソケットとした気がします。

GB(TB)、GiB (TiB) も明確にしておかないと、ベンチマーク等でかなりの差が
  GB:  1,000,000,000 bytes
  GiB: 1,073,741,824 bytes (ベンチマークで 7%の差は大きいです)
  TB:  1,000,000,000,000 bytes
  TiB: 1,099,511,627,776 bytes (当然差が広がります)

0 件のコメント: