Subject: C468: cJVM: A Cluster JVM Architecture for Single System Image Resent-Date: Thu, 02 Nov 2000 10:47:46 -0500 Resent-From: Geoffrey Fox Resent-To: Geoffrey Fox Date: Mon, 23 Oct 2000 12:49:19 +0300 From: teperman@il.ibm.com To: fox@csit.fsu.edu CC: factor@il.ibm.com, yariv@il.ibm.com Dear Sir, We have reviewed the article and changed the places suggested by the referee. Below you will find our detailed response (lines started with ">>> "). We have also polished the paper a bit and changed the title. Attached please find a ZIP file containing the revised paper. We will appreciate if you could notify us that the article has reached you. (See attached file: ClusterVM.zip) Regards, Avi Referee Report -------------- This is an excellent and very interesting paper. Most of my comments address presentation and grammar. General Comments: 1. Expand JVM one at the beginning of the paper and then use JVM throughout. 2. Expand SSI one at the beginning of the paper and then use SSI throughout. 3. Change all references to Chapters in the paper to Sections. >>> We made the changes suggested by the referee. 4. The reference to super-linear speedup in my opinion less than helpful (Page 20) as your text implies that you run the embarrassingly parallel application with X threads on a single machine and then run the run the same application with X threads on small cluster. A more realistic test would have been to run Y threads on a single machine and NxY threads on N machines. >>> We have eliminated the reference to super-linear test. Particular comments: 1. Change Summary title to Abstract. >>> Done. 2. Abstract, paragraph 2, what is 10Kloc? >>> We have added a footnote explaining the term. 3. Abstract, last sentence, change " --------------------------------------------------------------------- à how we aim at obtaining scalability." to " --------------------------------------------------------------------- à how we aim to obtain scalability." >>> Done. 4. Page 2, last paragraph, change "we enable exploiting opportunities for optimizations" to "we provide optimization opportunities" >>> Done. 5. Page 3, change sub-section title to "The Java Virtual Machine" >>> Done. 6. Page 4, section memory model, paragraph 1, should "integral" actually be "integer"? >>> Changed it as suggested. 7. Figure 2 - this sketch is small and very confusing. I suggest it is broken into two figures, with a caption describing the basics of each. >>> See response on point 9. 8. Page 6, paragraph 1, The wording "application which are concurrent daemons" is confusing and probably incorrect terminology. >>> We rephrased the sentence. 9. Page 6, last two paragraphs, the description here for Figure 2 is confusing to say the least. >>> We have elaborated more on the description and hope it is clearer now and clarifies why it should be one figure instead of two as suggested in point 7 above. 10. Page 7, object model sub-section, list "key challenges" >>> Done. 11. Page 7, object model sub-section, change "in the next chapter." To "in section X." >>> Done. 12. Page 7, Thread sub-section, change wording, "To enable a scalability gain," to "to provide scalability," >>> Done. 13. Page 8, Means to Achieve Scalability sub-section, remove "high" from "to obtain high scalability" >>> Done. 14. Page 8, Means to Achieve Scalability sub-section, paragraph 2, list different approaches. >>> We are not sure what the referee means by this comment at this point in the text. 15. Page 10, first paragraph, change "BACKGROUND Chapter" to "Background Section" >>> Done. 16. Page 12, second last paragraph, add "a" into sentence "if such a method is" >>> Done. 17. Page 12, last paragraph, two lots of "For example" --------------------------------------------------------------------- à Looks odd and should be changed. >>> These statements were re-phrased. 18. Page 14, thread creation sub-section, first paragraph, "subclass" should be hyphenated "sub-class" >>> We have changed the use of "subclass" to be consistent through the whole article. 19. Page 14, thread creation sub-section, first paragraph, the comment "(see below)" should be changed to be explicit, i.e. "(see Fig X)" or "(see sub-section y". >>> Statement was re-phrased to be more clear. 20. Figure 8, the legend of dashes/dots does not appear to correspond to the dot/dashes in the figure. RMI and Proxy Master Relation in the legend appear the same! >>> We have re-checked this and sorry to say we do not agree with the referee; we think the legend accurately describe what it is supposed to. 21. Page 18, footnote 12, typo, change "acquired" to "acquires" >>> Done. 22. Page 19, add space between "translate" and "(GA)" >>> Done. 23. Page 20, citation needed pBOB. >>> Citation added. 24. Page 22, reference 1 need a URL or more details of how to obtain the info that you refer to, i.e. UCSD Tech Doc #, etc. >>> URL added. --------------------------------------------------------------------- Name: ClusterVM.zip Type: Zip Compressed Data ClusterVM.zip (application/x-zip-compressed) Encoding: base64 Description: .ZIP File