i860 TM xp microprocessor. Programmer's Reference Manual, 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....smaller average page miss penalty than VM disk paging. 4 Implementation To evaluate our method of combining fine grained threading with memory server, we need a platform that can support both sequential and message passing parallel applications. The Intel paragon exactly matches our requirement [11]. The memory server model implementation requires a small modification to the kernel, while the thread system runs at user level. 4.1 Experiment Environment The Intel Paragon multicomputer used in our experiments consists of 10 nodes for computation. Each node has one compute processor and a ....

....Environment The Intel Paragon multicomputer used in our experiments consists of 10 nodes for computation. Each node has one compute processor and a communication co processor. Both processors are 50 MHz i860 microprocessors with 16 Kbytes of data cache and 16 Kbytes of instruction cache [11]. The two processors share 64 Mbytes of local memory, with around 56 Mbytes memory available to user applications. The memory bus provides a peak bandwidth of 400 MBytes sec. The nodes are interconnected with a wormhole routed 2 D mesh network whose peak bandwidth is 200 Mbytes sec per link [23] ....

i860 TM XP Microprocessor. Programmer's Reference Manual, 1991.

....multicomputer used in our implementation consists of 64 nodes for computation. Each node has one compute processor and a communication co processor, sharing 64 Mbytes of local memory. Both processors are 50 MHz i860 microprocessors with 16 Kbytes of data cache and 16 Kbytes of instruction cache [18]. The data caches 4 diff diff apply diff Acquire(l) Write(x) Release(l) diff fetch make twin Acquire(l) Write(x) Read(x) make twin Acquire(l) Acquire(l) fetch page Release(l) proc Comm co proc proc Compute Compute proc diff apply Comm Compute ....

i860 TM XP Microprocessor. Programmer's Reference Manual, 1991.

....of 64 nodes for computation. Each node has one compute processor and a communication co processor, sharing 64 Mbytes of local memory. Both processors are 50 MHz i860 microprocessors with 69 CHAPTER 4. IMPLEMENTATION AND EXPERIMENTS 70 16 Kbytes of data cache and 16 Kbytes of instruction cache [Int91a] The data caches are coherent between the two processors. The memory bus provides a peak bandwidth of 400 MBytes sec. The nodes are interconnected with a wormhole routed 2 D mesh network whose peak bandwidth is 200 Mbytes sec per link [TD92] The operating system is OSF 1 with multicomputer ....

i860 TM XP Microprocessor. Programmer's Reference Manual, 1991.

....multicomputer used in our implementation consists of 64 nodes for computation. Each node has one compute processor and a communication co processor, sharing 64 Mbytes of local memory. Both processors are 50 MHz i860 microprocessors with 16 Kbytes of data cache and 16 Kbytes of instruction cache [18]. The data caches are coherent between the two processors. The memory bus provides a peak bandwidth of 400 MBytes sec. The nodes are interconnected with a wormhole routed 2 D mesh network whose peak bandwidth is 200 Mbytes sec per link [31] The operating system is a micro kernel based version of ....

i860 TM XP Microprocessor. Programmer's Reference Manual, 1991.

No context found.

i860 TM xp microprocessor. Programmer's Reference Manual, 1991.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC