=>
2015年サーベイ論文だから、現在とはかなり違うかもしれないけど…
"A Survey of Interactive Remote Rendering Systems"
https://twitter.com/ogawa_tter/status/701920672734294020…
posted at 22:38:03
Stats | Twitter歴 4,628日(2011/07/29より) |
ツイート数 76,555(16.5件/日) |
表示するツイート :
=>
2015年サーベイ論文だから、現在とはかなり違うかもしれないけど…
"A Survey of Interactive Remote Rendering Systems"
https://twitter.com/ogawa_tter/status/701920672734294020…
posted at 22:38:03
.@houmei ほうめいさんの時代だと、d|i|g|i|t|a|lの i の上が四角形でしたか?
御参考 https://vt100.net/dec/logo https://pic.twitter.com/TwYyJaC1kw
posted at 21:19:51
=>
"Scaling Deep Learning Training with NCCL", Sep 26, 2018 https://devblogs.nvidia.com/scaling-deep-learning-training-nccl/…
NCCL 2.3
NVIDIA Collective Communications Library (NCCL)
Multi-GPU and multi-node collective communication primitives https://developer.nvidia.com/nccl
https://github.com/NVIDIA/nccl
AMD https://twitter.com/ogawa_tter/status/984192231207714817…
posted at 19:11:36
Pixar's Tractor
Job Distribution for Network Rendering https://renderman.pixar.com/product/tractor
Tractor 2, Doc https://rmanwiki.pixar.com/display/TRA/Tractor+2…
Sales & Pricing https://renderman.pixar.com/store
Free Non-Commercial RenderMan https://renderman.pixar.com/intro_logged_in
Cloud Services https://twitter.com/ogawa_tter/status/1046327356955144195…
https://twitter.com/ogawa_tter/status/1039477718172098560…
posted at 18:27:45
Pixar's Renderman Cloud Services
ZYNC (on Google's cloud) https://www.zyncrender.com/
Conductor Cloud Rendering https://www.conductortech.com/
Fox Renderfarm https://www.foxrenderfarm.com/cloud-rendering.html…
GridMarkets https://www.gridmarkets.com/renderman
Ranch https://www.ranchcomputing.com/en
NetRenderMan, 1992 https://twitter.com/ogawa_tter/status/767269366463201280…
posted at 18:14:16
@dylan522p Thanks!
posted at 00:51:10
=>
NVIDIA RTX 2080 Ti Deep Learning Benchmarks, Sep 26, 2018 https://lambdalabs.com/blog/2080-ti-deep-learning-benchmarks/…
https://github.com/lambdal/lambda-tensorflow-benchmark…
NVIDIA Turing Whitepaper, Sep 2018 https://twitter.com/ogawa_tter/status/1043789654581948416… https://pic.twitter.com/7XuOXTTRr2
posted at 22:38:11
=>
"CUDAAdvisor: LLVM-based Runtime Profiling for Modern GPUs", CGO 2018 https://drive.google.com/file/d/1RelfGOkycINBDrQy-UfUZxOL_DhrX8Sw/view…
https://github.com/sderek/CUDAAdvisor…
NVIDIA SASSI
ISCA 2015 https://twitter.com/ogawa_tter/status/618601619400175616…
Tutorial, MICRO 2015 https://twitter.com/ogawa_tter/status/674033170455461892… https://pic.twitter.com/xBebIHkNIQ
posted at 21:46:48
=>
AMD Radeon Open Compute Platform (ROCm), AMD, Lightning Talk, XDC 2018, Sep 29, 2018 PDF https://xdc2018.x.org/slides/AMD_Radeon_Open_Compute_Platform_lightning.pdf…
ROCm Kernel Driver features and HW support
Vega20
Multi-GPU memory mapping
Shared Virtual Memory
7 pp
AMD gem5 APU Simulator: Modeling GPUs https://twitter.com/ogawa_tter/status/1032285636173082626… https://pic.twitter.com/IpBPzETGHY
posted at 18:30:32
=>
"Accelerating Weather Prediction with NVIDIA GPUs", NVIDIA, WS on HPC in Meteorology, ECMWF, Sep 28, 2018
[Video] https://ecmwf.adobeconnect.com/px39opuphddm/
PDF https://www.ecmwf.int/sites/default/files/elibrary/2018/18628-accelerating-weather-prediction-nvidia-gpus.pdf…
Spherical Harmonics: DGX-2 vs. DGX-1V
16 GPUs: 2.4x
Optical Procesors, ESCAPE, Sep 2017 https://twitter.com/ogawa_tter/status/922503720121679872… https://pic.twitter.com/BI1IdV6k5v
posted at 00:59:37
=>
[Video] "How Arm's entry into the HPC market might affect meteorological codes", Simon McIntosh-Smith (Univ of Bristol), Keynote, WS on HPC in Meteorology, ECMWF, Sep 25, 2018 https://ecmwf.adobeconnect.com/p3o8blo2vxfq/
[Slides] https://twitter.com/ogawa_tter/status/1044596259477045248…
Cavium ThunderX2構成図公開は始めてな気がする
posted at 23:06:04
"平成30年度「次世代領域研究開発」(高性能汎用計算機高度利用事業費補助金)申請課題の選定結果について"、平成30年9月28日 http://www.mext.go.jp/b_menu/boshu/detail/1409409.htm…
申請課題数:10件
選定結果
「量子コンピュータ統合型次世代スーパーコンピュータの開発」 東北大学、小林 広明(大学院情報科学研究科 教授)
posted at 22:21:03
@yoshi_and_aki 量子コンピュータ(NISQ)をどの国よりも理解してる米国が"This calls for a “science first” approach that emphasizes supporting basic research because it’s still too early to tell what the best commercial uses for quantum tech will be."と言ってるわけですからね。
https://www.technologyreview.com/the-download/612199/quantum-research-in-the-us-was-just-handed-a-250-million-boost/?utm_campaign=social_button&utm_source=twitter&utm_medium=social&utm_content=2018-09-28…
Retweeted by OGAWA, Tadashi
retweeted at 19:41:50
Sept 27, 1993: @cray_inc unveils the T3D, Cray's first massively parallel processing (MPP) system, which will use Alpha processors; it will be available in sizes from 32 procs up to 2048 procs (http://www.cs.cmu.edu/~scandal/info/Cray.T3D…) #supercomputing #HPC https://pic.twitter.com/qHOV2hod0d
Retweeted by OGAWA, Tadashi
retweeted at 22:51:29
.@kumagi 2018年9月24~27日に開催の SNIA Storage Developer Conference (SDC 2018) のスライドと一部ビデオが公開されています https://twitter.com/ogawa_tter/status/1044971514133544964…
今回も 100以上のスライドと多くのビデオが公開され、ストレージ関連の研究・技術動向の把握にお役に立つと思われます。
posted at 09:28:17
[Slides/Video] SNIA Storage Developer Conference (SDC 2018), Sep 24-27, 2018 https://www.snia.org/events/storage-developer/presentations18…
100+ PDFs
SDC 2018 Agenda https://sniasdc18.pathable.com/meetings
今回もたくさんの公開です。 https://pic.twitter.com/BCfN3RIr66
posted at 00:26:38
.@mhiramat Linaro Connect Vancouverでの富士通 清水さんの基調講演 (2018/9/20):A64FX and the Arm HPC ecosystemのビデオが公開されてますね (スライドは未だ?) https://twitter.com/ogawa_tter/status/1043163829830467584…
連ツイに Bristol大学 McIntosh-Smith教授の ECMWFでの WS (9/25) スライド、TX2の構成図は初公開?
posted at 12:44:27
=>
Fujitsu A64FX Post-K Supercomputer: World's Fastest Arm Processor, Charbax, Sep 24, 2018 https://www.youtube.com/watch?v=fHb_L8Ht-hU…
Shirai-san (Software Dev for Post-K), Arai-san (Compiler Dev) and Kato-san (Linaro) from 富士通
A64FX, Fujitsu, Keynote, Linaro, Sep 20 2018 https://twitter.com/ogawa_tter/status/1043163829830467584… https://pic.twitter.com/OB23rezbPc
posted at 12:36:18
=>
"How Arm's entry into the HPC market might affect meteorological codes", Simon McIntosh-Smith, Keynote, WS on HPC in Meteorology, ECMWF, Sep 25, 2018 http://uob-hpc.github.io/assets/Isambard_ECMWF_HPC_WS_Sep_2018.pdf…
この構成図は…
A64FX & the Arm HPC ecosystem, 富士通, Linaro Connect, Sep 2018 https://twitter.com/ogawa_tter/status/1043163829830467584… https://pic.twitter.com/ozVrIRXSiY
posted at 23:35:30
.@tsuchie88 ツッチーさん、MIT Lincoln Labでは Supercomputerで、Wineを 16,000起動しています…
https://twitter.com/ogawa_tter/status/1044585517390557190…
"Specifically, this work demonstrates launching 16,000 Microsoft Windows applications in 5 minutes running on 16,000 processor cores"
posted at 23:26:04
"Interactive Launch of 16,000 Microsoft Windows Instances on a Supercomputer", MIT Lincoln Laboratory, arXiv, Aug 13, 2018 (IEEE HPEC, Sep 27, 2018) https://arxiv.org/abs/1808.04345
LLMapReduce in combination with the Wine Windows compatibility layer
LLSC https://www.ll.mit.edu/r-d/cyber-security-and-information-sciences/lincoln-laboratory-supercomputing-center… https://pic.twitter.com/kmBkomfzxM
posted at 22:52:49
Sept 25, 1984: Amdahl Corp, founded by Gene Amdahl in 1970, introduces 2 new @IBM S/370 compatible supercomputers built by @FujitsuHPC. This is Fujitsu's 1st attempt to target the Western #supercomputing market (US/Canada/W. Europe) (http://www.nytimes.com/1984/09/26/business/amdahl-announces-2-supercomputer-models.html…) #HPC
Retweeted by OGAWA, Tadashi
retweeted at 22:29:05
=>
Integrating Machine Learning with Simulations at Lawrence Livermore National Lab, Invited, DOE CSGF Annual Review, Jul 18, 2018 https://www.youtube.com/watch?v=91BVy-tGBI8…
PDF https://www.krellinst.org/doecsgf/conf/2018/pres/klewis.pdf…
NS16e:16x IBM TrueNorth, LLNL, Mar 2016 https://twitter.com/ogawa_tter/status/714795938951237633…
HEP & ML https://twitter.com/ogawa_tter/status/795045279921676288… https://pic.twitter.com/01I8C3OeNS
posted at 21:25:31
Pacific Northwest National Laboratory@PNNLab
“The most important thing I've accomplished, other than building the compiler, is training young people.” - #GraceHopper #GHC18 #womenintech We're excited to be at @ghc this week. Stop by booth 5908, B3502 👋🏽 https://pic.twitter.com/gAyLV7Cvvw
Retweeted by OGAWA, Tadashi
retweeted at 21:16:53
=>
"次世代HDD技術がついに登場、東芝はMAMRで勝負する
2019年に各社が市場投入"、2018年9月25日 https://limo.media/articles/-/7688
MAMR(マイクロ波アシスト記録):東芝とWD
MAMR, WD, Oct 2017 https://twitter.com/ogawa_tter/status/918449164194742272…
HAMR(熱アシスト記録):シーゲート
昔の HAMR関連の連ツイ https://twitter.com/ogawa_tter/status/711174977400086528…
posted at 18:25:47
=>
IBMによるここ 10年間の (2008-2017) LTO Tape、HDD、NAND動向についての MSST (Int Conf on Massive Storage Systems and Technology) での招待講演スライド (2018/5/15) https://twitter.com/ogawa_tter/status/996722730132099072…
連ツイに AIP Advances掲載論文 (2017/12、Open: 2018/5/1)。
posted at 18:11:54
.@telmin_orca @imakino 紹介したかもですが、AMDで k8s!
https://twitter.com/ogawa_tter/status/1040060105142632448…
※二つ前のツイに最近の比較的詳しい EPYC紹介スライド (2018/9/7) があります。
posted at 15:55:04
.@telmin_orca @imakino NVIDIA Turing Architectureの Whitepaperは登録せずに PDFダウンロード可能です (公開:2018/9/14)
https://twitter.com/ogawa_tter/status/1043789654581948416…
86ページと内容豊富です。
posted at 12:07:06
=>
[PATCH v8 00/16] Add support for Hygon Dhyana Family 18h processor, Sep 23, 2018 http://lkml.iu.edu/hypermail/linux/kernel/1809.2/06716.html…
First gen Hygon's processor (Dhyana) originates from AMD and shares most of the architecture with AMD's family 17h
https://www.phoronix.com/scan.php?page=news_item&px=Hygon-Dhyana-Linux-V8…
Jul, 2018 https://twitter.com/ogawa_tter/status/1016337986122731521…
posted at 23:22:53
"Ottawa-based Diablo Technologies files for bankruptcy", Dec 15. 2017 https://obj.ca/article/ottawa-based-diablo-technologies-files-bankruptcy…
"has closed its doors after almost 15 years in business"
"roughly $14.5 million to a long list of creditors"
知らなかった…
https://diablo-technologies.com/
ロシア語サイトがドメイン競売?
posted at 23:00:15
=>
"An analytical framework for cyber security", DARPA, Nov 2011 https://twitter.com/ogawa_tter/status/679565334764883968…
The supply chain is potentially compromised…
F-35C
Approximately 3500 ICs
208 FPGAs
64 FPGA & 9 ASIC types across 12 subsystems
78% of FPGAs & 66% of ASICs manufactured in China & Taiwan https://pic.twitter.com/9OQtNHnbck
posted at 22:44:33
"Boeing Defense, Space & Security Licenses Flex Logix's eFPGA on GlobalFoundries' 14nm Process". Sep 24, 2018 PDF.https://static1.squarespace.com/static/548d5bd7e4b087dc74cd6e6d/t/5ba8c72e0d9297701b1aee7c/1537787694562/2018+09+24+Boeing+Flex+Logix+GF14+Press+Release+FINAL.pdf…
eFPGA for Aerospace Applications
GF14LPP/12LP EFLX4K Gen 2 IN DEVELOPMENT, Sep 2018 PDF https://static1.squarespace.com/static/548d5bd7e4b087dc74cd6e6d/t/5ba6c9d11905f476677847cb/1537657299868/2018+09+EFLX4K+GF14LPP12LP+IN+DEVELOPMENT.pdf…
http://www.flex-logix.com/ https://pic.twitter.com/4aPCy23Z1E
posted at 21:26:21
.@KuniSuzaki DSN 2018、MICRO 2018、SC18の発表全てに以下の NVIDIAの方々が含まれています
S. Hari https://sivahari.github.io/
M. Sullivan http://mbsullivan.info/
T. Tsai https://research.nvidia.com/person/timothy-tsai…
Sullivanさんは Erez教授 https://twitter.com/ogawa_tter/status/1043507349955084288…
のもとで 2015年博士号 https://twitter.com/ogawa_tter/status/783503814334701572…
posted at 15:37:49
=>
"Lifetime memory reliability data from the field", IEEE Int Symposium on Defect and Fault Tolerance in VLSI and Nanotechnology Systems (DFT) 2017, Oct 2017 https://ieeexplore.ieee.org/abstract/document/8244428/…
posted at 14:57:16
@houmei http://www.npal.cs.tsukuba.ac.jp/~nishihara/search1.html…
Quasar Prologの制約評価機構, 情処大会, 第42回, pp.5-121 (Mar. 1991).
硴崎賢一: https://research02.jimu.kyutech.ac.jp/html/302_ja.html
博士論文 (論文博士)
PROLOG処理系の実現方式に関する研究、大阪大学、1993年 https://ir.library.osaka-u.ac.jp/repo/ouka/all/38263/…
要旨だけの公開
※リンク変更で再送しました。
posted at 20:18:25
@houmei http://www.npal.cs.tsukuba.ac.jp/~nishihara/search1.html…
Quasar Prologの制約評価機構, 情処大会, 第42回, pp.5-121 (Mar. 1991).
硴崎賢一 http://www.cse.kyutech.ac.jp/kakizaki.html
でしょうかね。
博士論文 (論文博士):
Prolog処理系の実現方式に関する研究、1993年、大阪大学 http://hakase.site/article/272077
posted at 20:11:15
=>
"NVIDIA Turing Architecture Deep Dive Whitepaper Available Now For Download", Sep 14, 2018 https://www.nvidia.com/en-us/geforce/news/geforce-rtx-20-series-turing-architecture-whitepaper/…
PDF https://www.nvidia.com/content/dam/en-zz/Solutions/design-visualization/technologies/turing-architecture/NVIDIA-Turing-Architecture-Whitepaper.pdf…
86 pages
Turingの Whitepaperは登録無しで公開!
Thanks @KOMACHI_ENSAKA https://pic.twitter.com/CfszFSUL5K
posted at 18:10:20
@KOMACHI_ENSAKA ありがとうございます!
posted at 16:48:39
=>
"NVIDIA Turing Architecture In-Depth", Sep 14, 2018 https://devblogs.nvidia.com/nvidia-turing-architecture-in-depth/…
Turing Streaming Multiprocessor Architecture
Turing Tensor Cores
GDDR6 Memory Subsystem
Memory Compression
Ray Tracing Technology
Turing RT Cores
Deep Learning Super-Sampling
内容豊富でPDF版が欲しい https://pic.twitter.com/LGui65Mcut
posted at 16:40:43
=>
"SwapCodes: Error Codes for Hardware-Software Cooperative GPU Pipeline Error Detection", NVIDIA, MICRO 2018
"Optimizing Software-Directed Instruction Replication for GPU Error Detection", UIUC, NVIDIA, SC18 https://research.nvidia.com/publication/2018-11_Optimizing-Software-Directed-Instruction…
DSN 2018, UBC & NVIDIA https://twitter.com/ogawa_tter/status/1033004789330661376…
posted at 15:58:51
.@KuniSuzaki 見つけた時は、「お、これは!」でしたが、せめて参考文献があればですね…
インタフェース 1984年8月号にメモリエラー検出・訂正回路例が掲載されてました https://twitter.com/ogawa_tter/status/1043743773837025281…
16 + 6 SBC-DBDがこの規模なので、CRAY-1時代の 64 bitエラー訂正はそれなりになり当初躊躇するのも…
posted at 15:15:33
=>
"エラー検出・エラー訂正"
特集;メモリ ICの基礎と応用、インターフェース、1984年8月号
[図 4.9] シングル;ビット・エラー訂正 (SBC) の回路例
8 bit + 4 bit
[図 4.10] シングル;ビット・エラー訂正ダブル・ビット・エラー検出 (SBC-DBD) の回路例
16 bit + 6 bit https://pic.twitter.com/G9pQwwSJ5k
posted at 15:08:02
"空飛ぶクルマ 欧米に挑む日本、官民に潜む壁
霞が関2018"、日本経済新聞、2018年9月19日 6:30 https://www.nikkei.com/article/DGXMZO3539550014092018000000/?n_cid=SNSTW001…
第1回 空の移動革命に向けた官民協議会、経産省、国交省、2018年8月29日 https://twitter.com/ogawa_tter/status/1036125182836916225…
posted at 14:57:55
.@KuniSuzaki 歴史になりますが、Reed–Solomon codesの Irving S. Reed教授の Symposium in honor of Solomon Golomb's 60th birthday (1992/5/29-31) での講演をもとにしたペーパー:A brief history of the development of error correcting codes, 2000 が公開されています https://www.sciencedirect.com/science/article/pii/S0898122100001127… https://pic.twitter.com/sxjzltdUzd
posted at 01:16:10
.@KuniSuzaki ECCの歴史のまとめはちょっと思いつきませんが、信頼性、Soft Error等との長い々連ツイ https://twitter.com/ogawa_tter/status/783483468894175233…
のメモリ障害解析、信頼性、新しい符号化の研究で、過去との比較をしていた気もします。
Erez教授の講演スライドも参考になりそうです https://lph.ece.utexas.edu/merez/MattanErez/Talks…
posted at 23:28:34
=>
CDC 7600にはパリティがあるのに?、と思ったら
「次のCDC 7600にはパリティを組み込んだため、専門家から「多くの農夫がコンピュータを買うだろう」と皮肉られた」
との続きが…
https://ja.wikipedia.org/wiki/ECC%E3%83%A1%E3%83%A2%E3%83%AA…
https://twitter.com/KuniSuzaki/status/1043500703207968768…
posted at 23:12:25
.@KuniSuzaki FPS AP-120Bは PDP-11アクセラレータなので少し特殊でしょうが、CDC7600と CRAY-1 (1976, LANL) の MTTFの大きな差はメモリ保護にありそうです https://twitter.com/ogawa_tter/status/1043475053969469443…
CRAY-1, 1976, LNALは 64 bit + 1 Parityで検知しかありませんが、連ツイの製品版は 64 + 8に改善されていますね。
posted at 21:55:46
Sept 22, 2008: In latest attempt to break into #HPC market, @Microsoft releases Windows HPC Server 2008, which includes a version of MPI based on MPICH2 that is compliant with the MPI spec (previous MS-MPI implementations were not fully compatible w/ the spec) #supercomputing
Retweeted by OGAWA, Tadashi
retweeted at 21:50:03
=>
CRAY-1 Evaluation, Final Report, LANL, Dec, 1976 https://twitter.com/ogawa_tter/status/1043407093057740801…
64 data bits plus 1 parity bit
The CRAY-1 computer system, Cray Research, Inc., 1977 http://www.computerhistory.org/brochures/doc-43729572b49ef/…
64 data bits plus 8 check bit
CDC7600, 1968, Parity https://twitter.com/ogawa_tter/status/1043475053969469443… https://pic.twitter.com/V5PN4jrhdo
posted at 21:46:33
=>
FPS AP-120B/FPS-164, IEEE Computer, Sep, 1981 https://www.computer.org/csdl/mags/co/1981/09/01667502.pdf…
CDC 7600, Nov 1968 http://www.computerhistory.org/collections/catalog/102646087…
Memory Parity
CPU small core mem
60 bit + 5 parity
CPU large core mem
60 + 4
PPU core mem
12 + 1
CRAY-1 (LANL, Dec 1976)
64 + 1
https://twitter.com/ogawa_tter/status/1043456763071098881…
=> https://pic.twitter.com/7L4hC6GZAn
posted at 21:20:14
.@KuniSuzaki NASAで 1977年10月に開催されたワークショップ:Future Computer Requirements for ... にhttps://twitter.com/ogawa_tter/status/1043456763071098881…
MFLOPS/MTTF/相対 HWコスト
FPS AP-120B Array Processor
12/300 hours/0.05
CDC7600
12/days/1.00
CRAY-1
25 (60:計画)/7 hours/0.60
とありますね…
posted at 20:22:35
=>
Future Computer Requirements for Computational Aerodynamics, NASA, WS, Oct, 1977 https://ntrs.nasa.gov/search.jsp?R=19780011835&hterms=19780011835…
"Computational Advances in Flued Dynamics"
TABLE I
Comparative Computer Statistics (Average)
Page 479
FPS: https://en.wikipedia.org/wiki/FPS_AP-120B
Cray, 1979/77 https://twitter.com/ogawa_tter/status/1043407093057740801… https://pic.twitter.com/TxwIbrtmxI
posted at 20:07:33
.@KuniSuzaki もしかしたら、仕様書策定時の CRAY-1の実情にあわせて Mandatory Requirementでは 4時間にしたのでは、と思ったりします…
CRAY-1について歴史的背景やその後のビジネスも含めた面白そうなスライドが見つかりました https://twitter.com/ogawa_tter/status/1043408629297115136…
残念ながらドイツ語です…
posted at 18:47:23
.@kotatsu_mi どちらの digital がお好みですか?
御参考
https://vt100.net/dec/logo
http://homepage.divms.uiowa.edu/~jones/pdp8/logos/… https://pic.twitter.com/iahkWB5oW0
posted at 18:15:33
.@KuniSuzaki 見返したところ、
CRAY-1
Mandatory Requirement and Qualification Criteria
Mean-Time-to-Failure (MTTF)
"must equal or exceed four hour"
とありますね…
https://twitter.com/ogawa_tter/status/1043407093057740801…
その後劇的に改善され
MTTF (H)、1977
Jan: 5.16
Feb: 10.31
Mar: 18.22
Apr: 53.13
posted at 17:12:41
=>
"Cray-1, Ikone des Supercomputing – Wie die Maschine zur Welt kam, und was danach passierte", Wolfgang Stief, Vintage Computing Festival Berlin (VCFB), Oct 7, 2017
Video https://media.ccc.de/v/vcfb2017_-_63_-_de_-_201710071030_-_cray-1_ikone_des_supercomputing_-_wolfgang_stief…
Slides https://speakerdeck.com/stiefkind/vcfb2017-cray-1-ikone-des-supercomputing-wie-die-maschine-zur-welt-kam-und-was-danach-passierte…
CRAY-1, LANL, 1976/77 https://twitter.com/ogawa_tter/status/1043407093057740801… https://pic.twitter.com/WT41bNGyZs
posted at 16:56:17
=>
CRAY-1 Evaluation, Final Report, LANL, Dec, 1976 PDF https://permalink.lanl.gov/object/tr?what=info:lanl-repo/lareport/LA-06456-MS…
Mandatory Requirement and Qualification Criteria (P. 67)
Mean-Time-to-Failure (MTTF)
"must equal or exceed four hour"
The CRAY-1 at LASL (LANL): An Update, LANL, May, 1977 PDF https://permalink.lanl.gov/object/tr?what=info:lanl-repo/lareport/LA-UR-77-1069… https://pic.twitter.com/rjbu0WLUHu
posted at 16:50:11
=>
クレイリサーチ (Cray Research: Original Cray) 手持ちコレクション等の連ツイ
クレイリサーチのカタログを発掘 (1/n)!
CRAY Y-MP、1990年4月
「それは一台のスーパーコンピュータから始まった。」
https://twitter.com/ogawa_tter/status/794551815865962497…
Cray Research関連のツイも今後はここに繋げる。
posted at 16:20:55
=>
Cray Research (Original Cray) 関連のツイは以下この連ツイに追加
"Vector architectures: past, present and future", ICS 1998 https://twitter.com/ogawa_tter/status/768705511818616832…
特集:速さを競うスーパコンピュータ
日経エレクトロニクス、1983年4月11日号
FACOM VP-100/200
HITACHI S-810/10, 20
---
NEC SX-1/2
等々
posted at 16:18:17
=>
"GraphIA: An In-situ Accelerator for Large-scale Graph Processing", MEMSYS 2018, Oct 2, 2018 PDF https://nicsefc.ee.tsinghua.edu.cn/media/publications/2018/MEMSYS18_None.pdf…
"GraphIA couples large-capacity memory and computing resource in DRAM by connecting multiple chips with computation circuits inside."
Ring interconnection https://pic.twitter.com/76acwpMHbD
posted at 01:17:42
=>
A64FX and the Arm HPC ecosystem, T. Shimizu, Fujitsu, Keynote, Linaro Connect Vancouver, Sep 20, 2018 https://www.youtube.com/watch?v=kq4XfVtFme4…
A64FX, Hot Chips 2018 https://twitter.com/ogawa_tter/status/1032241831679938560…
Tofu D, Cluster 2018 https://twitter.com/ogawa_tter/status/1042347940495351809…
初版 CPU評価概要、理研、2018年9月4日 https://twitter.com/ogawa_tter/status/1042554666146455552… https://pic.twitter.com/GQm7jL2yGg
posted at 00:43:32