Blue Gene/L compute chip: Memory and ethernet subsystem

被引:15
作者
Ohmacht, M
Bergamaschi, RA
Bhattacharya, S
Gara, A
Giampapa, ME
Gopalsamy, B
Haring, RA
Hoenicke, D
Krolak, DJ
Marcella, JA
Nathanson, BJ
Salapura, V
Wazlowski, ME
机构
[1] IBM Corp, Div Res, Thomas J Watson Res Ctr, Yorktown Hts, NY 10598 USA
[2] IBM Corp, Engn & Technol Serv, Rochester, MN 55901 USA
关键词
D O I
10.1147/rd.492.0255
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
The Blue Gene((R))/L computer chip is a dual-processor system-on-a-chip capable of delivering an arithmetic peak peifiornimice of 5.6 gigaflops. To match the memory speed to the high compute performance, the system implements an aggressive three-level on-chip cache hierarchy. The implemented hierarchy offers high bandwidth and integrated prefetching on cache hierarchy levels 2 and 3 (L2 and L3) to reduce memory access time. A Gigabit Ethernet interface driven by direct memory access (DMA) is integrated in the cache hierarchy, requiring only an external physical link layer chip to connect to the media. The integrated L3 cache stores a total of 4 MB of data, using multibank embedded dynamic random access memory (DRAM). The 1,024-bit-wide data port of the embedded DRAM provides 22.4 GB/s bandwidth to serve the speculative prefetching demands of the two processor cores and the Gigabit Ethernet DMA engine. To reduce hardware overhead due to cache coherence intervention requests, memory coherence is maintained by software. This is particularly, efficient for regular highly parallel applications with partitionable working sets. The system further integrates an on-chip double-data-rate (DDR) DRAM controller for direct attachment of main memory modules to optimize overall memory, performance and cost. For booting the system and low-latency interprocessor communication and synchronization, a 16-KB static random access memory (SRAM) and hardware locks have been added to the design.
引用
收藏
页码:255 / 264
页数:10
相关论文
共 7 条
[1]   Embedded DRAM design and architecture for the IBM 0.11-μm ASIC offering [J].
Barth, JE ;
Dreibelbis, JH ;
Nelson, EA ;
Anand, DL ;
Pomichter, G ;
Jakobson, P ;
Nelms, MR ;
Leach, J ;
Belansek, GM .
IBM JOURNAL OF RESEARCH AND DEVELOPMENT, 2002, 46 (06) :675-689
[2]   Early analysis tools for system-on-a-chip design [J].
Darringer, JA ;
Bergamaschi, RA ;
Bhattacharya, S ;
Brand, D ;
Herkersdorf, A ;
Morrell, JK ;
Nair, II ;
Sagmeister, P ;
Shin, Y .
IBM JOURNAL OF RESEARCH AND DEVELOPMENT, 2002, 46 (06) :691-707
[3]  
*IBM MICR DIV, 2004, POWERPC 440 COR
[4]  
*IEEE, 2000, 8023 IEEE
[5]   The eDRAM based L3-Cache of the BlueGene/L supercomputer processor node [J].
Ohmacht, M ;
Hoenicke, D ;
Haring, R ;
Gara, A .
16TH SYMPOSIUM ON COMPUTER ARCHITECTURE AND HIGH PERFORMANCE COMPUTING, PROCEEDINGS, 2004, :18-22
[6]  
Palacharla S., 1994, Proceedings the 21st Annual International Symposium on Computer Architecture (Cat. No.94CH3397-7), P24, DOI 10.1109/ISCA.1994.288164
[7]   Data prefetch mechanisms [J].
Vanderwiel, SP ;
Lilja, DJ .
ACM COMPUTING SURVEYS, 2000, 32 (02) :174-199