D. Grunwald. Personal communication, Univ. of Colorado at Boulder, Jan. 1995. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Using Memory-Mapped Network Interfaces to Improve the.. - Kontothanassis, Scott (1996) (24 citations) (Correct)
....what we call CHHN, with the added twist that when only two processors have mappings to a page they write through to each others copies, rather than to a single master copy. Researchers at the University of Colorado also propose the use of NCC NUMA hardware for efficient shared memory emulations [10]. In their approach however, the coherence protocol is simplified to invalidate all shared pages on lock acquisition operations. Thekkath et al. have proposed using a network that provides remote memory access to better structure distributed and RPC based systems [30] The Blizzard system [27] at ....
D. Grunwald. Personal communication, Univ. of Colorado at Boulder, Jan. 1995.
A scalable and explicit event delivery mechanism for UNIX - Banga, Mogul, Druschel (1999) (46 citations) (Correct)
....each activity, allowing one activity to block on I O without affecting the progress of others. Experience with UNIX and similar systems has shown that this scales badly as the number of threads increases, because of the costs of thread scheduling, context switching, and thread state storage space[6, 9]. The use of a single process per connection is even more costly. The most efficient approach is therefore to allocate a moderate number of threads, corresponding to the amount of available parallelism (for example, one per CPU) and to use non blocking I O in conjunction with an efficient ....
D. Grunwald. Personal communication, Mar. 1998.
Tools and Techniques for Building Fast Portable Threads Packages - Keppel (1993) (71 citations) (Correct)
....this, each stack allocation is followed by a large unallocated region that is mapped to cause a fault on access. When the stack overflows, stack references go to the unallocated region. These references invoke a fault handler that maps additional storage in to the region, thus growing the stack [Gru91]. Stacks space can also be scavenged , unmapping storage between the thread s current top of stack and the end of the thread s stack region. 5.11. Shared Stacks The preceding discussion of thread stacks assumed that distinct storage is used for each thread s stack. An alternative is to have a ....
Dirk Grunwald. Personal communication, 1991.
A scalable and explicit event delivery mechanism for UNIX - Gaurav Banga (1999) (46 citations) (Correct)
....allocate a single thread to each activity, allowing one activity to block on I O without affecting the progress of others. Experience has shown that this scales badly as the number of threads increases, because of the costs of thread scheduling, context switching, and thread state storage space[6, 9]. The use of a single process per connection is even more costly. The most efficient approach is therefore to allocate a moderate number of threads, corresponding to the amount of available parallelism (for example, one per CPU) and to use non blocking I O in conjunction with an efficient ....
D. Grunwald. Personal communication, Mar. 1998.
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