Dear Hui and co-authors,

Thank you for submitting a revised manuscript and for largely addressing the reviewers' comments. As one of the reviewers said, the quality of the language could be better and I include below a detailed list of corrections to address this. In future, I would strongly suggest that you seek advice on the language when drafting a paper; it will improve the readability of your papers and help the review process.

I ask that you implement these corrections. Also, it wasn't clear to me that you'd altered the Figure 4 caption in response to one of the reviewers. Can you please do so? Once you've done these changes, I'll be happy to recommend that your paper be accepted and published in GMD.

Regards,
Fiona O'Connor

1. Abstract: line 11: Change “simulation” to “simulations”
2. Abstract: line19: change “reaches maximum 0.025%” to “reaches a maximum of 0.025%”
3. Abstract: line 20: Change “of module test” to “of the module test”
4. Abstract: line 20: Explain MP
5. Abstract: line 22: Change “with performance of old CBM-Z” to “with the performance of CBM-Z”
6. Abstract: line 22: When you write “For the node”, can you explain which node?
7. Abstract: line26: change “13~16%” to “13-16%”
8. Abstract line 26: Change “three-dimension” to “three-dimensional”
9. Abstract line 28: Change “on performance for CTM” with “on the performance of the CTM”
10. Abstract line 29: Change “to 3.32x and 1.96x for gas-phase” to “to a speedup of 3.32 and 1.96 for the gas-phase”
11. Abstract: line 30: Change “speedups for gas-phase” to “speedups for the gas-phase”
12. Abstract: line 31: Change “but whole model” to “but the whole model”
13. Abstract: line 31: Change “performance comparing with” to “performance compared to the”
14. Abstract: line 32: Change “due to bad” to “due to the poor”
15. Abstract: page 2, line 2: Change “the profit of updating CPU improved about 47% by using” to “the benefit of updating CPU improved by about 47% by using”
16. Abstract: page 2, line 3: Change “for new hardware” to “for the new hardware”
17. Abstract: page 2, line 5: Change “scientific simulation” to “scientific simulations”
18. Introduction: page 2, line 8: Change “the Chemistry Transport Models (CTM) is widely” to “Chemistry Transport Models (CTMs) are widely”
19. Introduction: page 2, line 13: Change “the corresponding simulation studies with complicated processes” to “other studies with more complex processes”
20. Introduction: page 2, line 14: Remove “simulation”
21. Introduction: page 2, line 15: Change “In the CTM” to “In a CTM”
22. Introduction: page 2, line 20: Change “Carbon Bond Mechanism” to “the Carbon Bond Mechanism”
23. Introduction: page 3, line 6: Change “test the performances” to “test the performance”
24. Introduction: page 3, line 6: Change “with it, and the codes only contained this single module make it easier” to “with it. The codes test only contained this single module, making it easier”
25. Introduction: page 3, line 8: Change “introduces the detail of CBM-Z” to “introduces the CBM-Z”
26. Introduction: page 3, line 9: Change “operated simultaneous” to “operated on simultaneously”
27. Introduction: page 3, line 10: Change “schemes were called Multiple-Points” to “scheme was called the Multiple-Points”
28. Introduction: page 3, line 11: Change “introduced the test cases and the test results” to “introduce the test cases and present the test results”
29. Method Description, page 3, line 14: Change “complicated” to “complex”
30. Method Description, page 3, line 15: I suggest that you change “too many branches and unbalanced calculations” to “too many options and poor load balancing” as readers unfamiliar with the structure of CBM-Z will not understand what it meant by “branches”. It’s also a word commonly used to add new code to a model that isn’t included in the release and so, probably isn’t appropriate until you’ve introduce CBM-Z in Section 2.1.
31. Method Description, page 3, line 16: Change “to the low performance” to “to poor performance”
32. Section 2.1, page 3, line 23: Change “simulation” to “simulations”
33. Section 2.1, page 3, line 25: Change “In the NAQPMS and GNAQPS model” to “In the NAQPMS and GNAQPS models”
34. Section 2.1, page 4, line 7: Change “saving the computation resources” to “saving computational resources”
35. Section 2.1, page 4, line 9: Change “uses the other” to “uses other”
36. Section 2.1, page 4, line 11: Change “indexes” to “indexing”
37. Section 2.1, page 4, line 12: Change “efforts” to “effort”
38. Section 2.1, page 4, line 12: Change “CBM-Z” to “The CBM-Z”
39. Section 2.2, page 4, line 23: Change “the automatic” to “automatic”
40. Section 2.2, page 4, line 25: Change “directive need to” to “directives are needed to”
41. Section 2.2, page 4, line 23: Change “the vectorization” to “vectorization”
42. Section 2.2, page 4, line 29: Change “the model grids” to “the model grid boxes”. This requested change occurs on this line but elsewhere throughout section 2.2
43. Section 2.2, page 5, line 4: Change “the vectorization” to “vectorization”
44. Section 2.2, page 5, line 8: Change “loop were” to “loops were”
45. Section 2.2, page 5, line 9: Change “the balance of” to “load balancing of”
46. Section 2.2, page 5, line 10: Change “to balancing the calculation” to “to this”.
47. Section 2.2, page 5, line 15: the line “The VLEN number grids contain …” is very unclear and I would ask that you re-write this with more clarity
48. Section 2.2, page 5, line 22: Change “results from copying back” to “results from being copied back”
49. Section 3, page 5, line 32: Change “regional named CTM” to “regional CTM named”
50. Section 3, page 5, line 32: Change “with optimized” to “with the optimized”
51. Section 3, page 6, line 1/2: Change “the real scenes” to “three dimensions with varying chemical regimes”
52. Section 3.1, page 6, line 13: Change “compile flags about vectorization” to “compile flags for vectorization”
53. Section 3.1, page 6, line 13: Change “accuracy control of CBM-Z” to “accuracy of the CBM-Z”
54. Section 3.1, page 6, line 14: Change “NAQPMS model” to “the NAQPMS model”
55. Section 3.1, page 6, line 15: Change “were adopted” to “were adopted for MP-CBM”
56. Section 3.1, page 6, line 16: Change “options of compile” to “options for the compile”
57. Section 3.1, page 6, line 16: Change “control” to “controls”
58. Section 3.1, page 6, line 16: Change “calculation” to “calculations”
59. Section 3.1, page 6, line 17: Change “discussed” to “consider”
60. Section 3.1, page 6, line 18: Change “designated” to “selected”
61. Section 3.1, page 6, line 18: Change “would improve” to “improves”
62. Section 3.1, page 6, line 19: Change “performance but decline” to “computational performance but reduces”
63. Section 3.1, page 6, line 19: Change “accuracy of floating-point calculation” to “accuracy of the floating-point calculations”
64. Section 3.1, page 6, line 19: Change “Using precise would keep value safe and force” to “Using precise is a safer option and forces”
65. Section 3.1, page 6, line 20: Change “avoid” to “to avoid”
66. Section 3.1, page 6, line 20: Change “some calculation to improve the accuracy” to “some calculations to improve accuracy”
67. Section 3.1, page 6, line 20: Change “We compared the results of two options” to “We compare the results of the two options”
68. Section 3.2, page 6, line 23: Change “of single” to “of the single”
69. Section 3.2, page 6, line 24: Change “the results under the different parallel frames” to “the results using different parallelization techniques”
70. Section 3.2, page 6, line 24: Change “repeated for 10” to “repeated 10”
71. Section 3.2, page 6, line 25: Change “to isolate the impact of the random state of platforms” to “to reduce the impact from any platform variability”
72. Section 3.2, page 6, line 27: Change “two cases were used for the box model of CBM-Z” to “two cases that were used for the CBM-Z box model”
73. Section 3.2, page 6, line 27: Change “single point” to “single gridbox”
74. Section 3.2, page 6, line 28: Again, change “grids” to “gridboxes”
75. Section 3.2, page 6, line 28: Change “this module” to “the module”
76. Section 3.2, page 6, line 29: Change “under the ideal scenario” to “under a more realistic scenario”
77. Section 3.2, page 6, line 29: Change “values of the single point case were showed” to “values for the single gridbox are shown”
78. Section 3.2, page 6, line 30: change “time step-size” to “time step”
79. Section 3.2, page 6, line 30: Change “for two cases” to “for the two cases”
80. Section 3.2, page 7, line 1: Change “The case used for validating output the simulation results every 5 minutes” to “For validation purposes, output every 5 minutes was used”
81. Section 3.2, page 7, line 1: Change “the performance” to “the computational performance”
82. Section 3.2, page 7, line 1: Change “did not open the” to “did not include the”
83. Section 3.2, page 7, line 2: Change “to isolate the disturbance of I/O” to “to eliminate any impact form I/O”
84. Section 3.2, page 7, line 2: Change “compiling flags for precisions of” to “compile flags for the precision of”
85. Section 3.2, page 7, line 2: change “calculation were” to “calculations are”
86. Section 3.2, page 7, line 3: Change “in the Table” to “in Table”
87. Section 3.2, page 7, line 3: Change “We tested” to “We test”
88. Section 3.2, page 7, line 3: Change “and optimized” to “and the optimized”
89. Section 3.2, page 7, line 4: Change “time of the computation portion” to “computational time”
90. Section 3.2.2: I would suggest that you change the name of this section to “Box model validation”
91. Section 3.2, page 7, line 6: I felt that this line is repetitive and could be removed. However, the point about the meteorology seems to contradict what was stated in 3.2.1. Was the meteorology constant or updated?
92. Section 3.2, page 7, line 7: Change “validated” to “evaluate”
93. Section 3.2, page 7, line 9: change “are relatively representative and suitable” to “important for tropospheric gas-phase chemistry and sulphate aerosol formation and hence, suitable”
94. Section 3.2, page 7, line 10: Change “Figure 4 show” to “Figure 4 shows”
95. Section 3.2, page 7, line 11; Change “and optimized (opt) version model” to “and the optimized (opt) model”
96. Section 3.2, page 7, line 12: Change “results by the baseline codes” to “results with the baseline code”
97. Section 3.2, page 7, line 12: Change “is as the benchmark” to “is the benchmark”
98. Section 3.2, page 7, line 13: Change “results from baseline and optimized codes with same” to “results from the baseline and optimized codes with the same”
99. Section 3.2, page 7, line 14: Change “relative” to “relatively”
100. Section 3.2, page 7, line 14: Change “and would prohibit the optimizations that would affect” to “and prohibits optimizations that can affect”
101. Section 3.2, page 7, line 15: Change “would lead to error” to “can lead to errors”
102. Section 3.2, page 7, line 15: Change “codes” in two places to “code”
103. Section 3.2, page 7, line 15: Change “compile flag is” to “compile flag relative to the benchmark is”
104. Section 3.2, page 7, line 16: Change “As stated” to “As shown”
105. Section 3.2, page 7, line 16: change “induced about maximum 0.025% RE” to “results in a maximum RE of 0.025%”
106. Section 3.2, page 7, line 17: Change “We can find that the error brought by fast=1” to “We find that the error caused by the fast=1”
107. Section 3.2, page 7, line 18: Change “for the species with small magnitude like” to “for species with low concentrations”
108. Section 3.2, page 7, line 18: Change “We would further discuss the impact of fast=1” to “We will further discuss the impact of the fast=1”
109. Section 3.2, page 7, line 19: Change “in the Section 3.3.2 under a real simulation scenario” to “in Section 3.3.2 in the context of CTM simulations”
110. There seems to be two Section 3.2.2 – Can you correct section numbering?
111. May I suggest that you change the name of the second Section 3.2.2 to “Box model computational performance”
112. Section 3.2, page 7, line 21: Again, can you change the use of the word “grids” to “gridboxes” here and elsewhere
113. Section 3.2, page 7, line 21: Change “ideal performance” to “computational performance”
114. Section 3.2, page 7, line 23: Change “precise compile flag was considered as the benchmark time to evaluate the speedups” to “the precise compile flag was considered as the benchmark time”
115. Section 3.2, page 7, line 24: Change “CPUs and KNL” to “of CPUs and KNL”
116. Section 3.2, page 7, line 26: Change “codes” to “code”
117. Section 3.2, page 7, line 27: Change “with precise” to “with the precise”
118. Section 3.2, page 7, line 27: Change “speedups for baseline codes on to generation” to “speedups for he baseline code on both CPUs”
119. Section 3.2, page 7, line 28: Change “updating CPU enable” to “updating the CPU enables”
120. Section 3.2, page 7, line 2”module gains about 1.52 and 1.59 times acceleration with” to “module to gain a speedup of about 1.52 ad 1.59 with”
121. Section 3.2, page 7, line 30: Change “MP CBM-Z” with “The MP CBM-Z”
122. Section 3.2, page 7, line 30: Remove “two generation”
123. Section 3.2, page 7, line 30: Change “old architecture of Broadwell” with “Broadwell architecture”
124. Section 3.2, page 7, line 31: Change “codes” to “code”
125. Section 3.2, page 7, line 32: change “Regarding to” to “In regard to”
126. Section 3.2, page 8, line 2: Remove “could”
127. Section 3.2, page 8, line 2: Remove “with impact of diverse compile flags” – at present, this doesn’t make sense but is demonstrated by the next sentence so can be omitted.
128. Section 3.2, page 8, line 3: Change “with original CBM-Z codes” to “with the original CBM-Z code”
129. Section 3.2, page 8, line 5/6: Replace “codes” with “code” in 3 places
130. Section 3.2, page 8, line 5: Change “with old” to “with the old”
131. Section 3.2, page 8, line 5: Change “of optimized” to “of the optimized”
132. Section 3.2, page 8, line 5: Remove “on two generations codes”
133. Section 3.2, page 8, line 6: Change “we can find” to “we find”
134. Section 3.2, page 8, line 7; Change “CPU generation” to “the CPU”
135. Section 3.2, page 8, line 8: Change “1.5 times of baselines codes” to “1.5x speedup gained with the baseline code”
136. Section 3.2, page 8, line 9: Change “affect the performance of codes” to “affect the code performance”
137. Section 3.2, page 8, line 9; change “optimized codes consumed” to “optimized code took”
138. Section 3.2, page 8, line 10: Remove “to finish the single core”
139. Section 3.2, page 8, line 10; Change “with precise” to “with the precise”
140. Section 3.2, page 8, line 10; change “with majority” to “with the majority”
141. Section 3.2, page 8, line 11: Change “while it only consumed 214.09 seconds and gains 4.74x speedup with fast=1” with “it only took 214.09 seconds and obtained a speedup of 4.74x with the fast=1”
142. Section 3.2, page 8, line 12: Change “with old” to “with the old”
143. Section 3.2, page 8, line 13: Change “baseline” with “the baseline”
144. Section 3.2, page 8, line 14: Change “On Intel” with “On the Intel”
145. Section 3.2, page 8, line 15; Change “with original” to “with the original”
146. Section 3.2, page 8, line 16: Change “that majority” to “that the majority of”
147. Section 3.2, page 8, line 18: Change “MP CBM-Z by compiling by fast=1 option with the MPI” to “MP CBM-Z by compiling with the fast=1 option and with MPI”
148. Section 3.2, page 8, line 19: Change “We valued” to “We evaluated”
149. Section 3.2, page 8, line 19: Change “of old” to “of the old”
150. Section 3.2, page 8, line 20: Change “were showed” to “are shown”
151. Section 3.2, page 8, line 20: Change “Table 5 and the MPI” to “Table 5. The MPI”
152. Section 3.2, page 8, line 22; Change “for MPI” to “with MPI”
153. Section 3.2, page 8, line 24: Change “with same” to “with the same”
154. Section 3.2, page 8, line 25: Change “resulted” to “results”
155. Section 3.2, page 8, line 27; Change “which can be significant to fully use the new MIC architecture processors such as KNL” to “and can be important in making full use of the new MIC architecture processors such as KNL”
156. Section 3.3, page 8, line 30: Change “CTM, Nested” to “CTM, the Nested”
157. Section 3.3, page 8, line 31: Change “in real scenarios” to “under more realistic conditions”. Also, replace the next sentence with “The following sub-sections will describe the CTM test case, and will present results from the scientific validation as well its computational performance.”
158. Section 3.3, page 9, line 1; Change section name to “CTM test case description”
159. Section 3.3, page 9, line 2 & 3: Change “NAQPMS” to “The NAQPMS”
160. Section 3.3, page 9, line 4: Change “multi-phase” to “and multi-phase”
161. Section 3.3, page 9, line 5: Change “Same as the box model, NAQPMS” to “In a similar way to the box model test case, the NAQPMS”
162. Section 3.3, page 9, line 5: Change “with baseline and optimized CBM-Z” to “with the baseline and optimized CBM-Z modules”
163. Section 3.3, page 9, line 6: Change “with diverse compiling flags shown” to “with various compile flags as shown”
164. Section 3.3, page 9, line 7: Change “simulation covered East” to “simulation covering the”
165. Section 3.3, page 9, line 7: Remove “of grids”
166. Section 3.3, page 9, line 8: Change “driven” to “driving”
167. Section 3.3, page 9, line 12: Change “ended in 00:00 UTC” to “ended at 00:00 UTC”
168. Section 3.3, page 9, line 14: Change “calculation of performance was based on the average value of experiments” to “performance was assessed on the basis of the average value”
169. Section 3.3.2: I suggest that you change the name of the sub-section to “CTM Validation”
170. Section 3.3, page 9, line 16: Change “includes” to “including”
171. Section 3.3, page 9, line 16: Remove “or simulation at 00:00 UTC, August 20, 2015”
172. Section 3.3, page 9, line 17: Change “valid” to “evaluate”
173. Section 3.3, page 9, line 17: There are many places where you refer to the “old” CBM-Z. I suggest that you remove all references to “old” and replace with either “unoptimized” or “benchmark” or “baseline”. Whichever you choose, use consistently throughout the manuscript
174. Section 3.3, page 9, line 18: Change “”of old NAQPMS codes” to “of the XXX NAQPMS code”, where XXX is a suggested alternative to “old” as indicated in comment no 173
175. Section 3.3, page 9, line 18: Change “with fast=1” to “with the fast=1”
176. Section 3.3, page 9, line 18: Change “and optimized” to “and the optimized”
177. Section 3.3, page 9, line 21: Change “concentration of these species” to “their concentrations”
178. Section 3.3, page 9, line 21: Change “other experiments” to “other experiments relative to the baseline”
179. Section 3.3, page 9, line 21: Change “We can find” to “We find”
180. Section 3.3, page 9, line 21: Change “show same reasonable spatial” to “show the same spatial”
181. Section 3.3, page 9, line 22: Change “AEs of majority grids” to “AEs in the majority of gridboxes”
182. Section 3.3, page 9, line 22: Change “of other” to “for the”
183. Section 3.3, page 9, line 23: The use of “old” – see comment above
184. Section 3.3, page 9, line 23: Change “with same fast=1 out that range shows the more obvious AEs” to “with the same fast=1 flag are larger”
185. Section 3.3, page 9, line 25: Change “options enable the results of two version codes” to “option enables the results of the two versions”
186. Section 3.3, page 9, line 26: Change “in near” to “in the near”
187. Section 3.3, page 9, line 26: Change “For majority points” to “For the majority of points”
188. Section 3.3, page 9, line 27: Change “existed the exceptional obvious” to “show exceptional and obvious”
189. Section 3.3, page 9, line 28: Change “after 72 integration” to “after 72 hrs of integration”
190. Section 3.3, page 9, line 28: Change “and fast=1” to “and the fast=1”
191. Section 3.3, page 9, line 28: Change “error” to “errors”
192. Section 3.3, page 9, line 29: Change “version codes” to “versions”
193. Section 3.3, page 9, line 29: Again replace reference to “old”, as above
194. Section 3.3, page 9, line 29: Change “For old MAQPMS codes” to “For the XXXX NAQPMS code”, where XXXX is your chosen replacement for the word “old”
195. Section 3.3, page 9, line 29 & 30: change “maximum 0.13, 4.5, 0.17 and 2.6 ppbv AEs” to “maximum AEs of 0.13, 4.5, 0.17 and 2.6 ppbv
196. Section 3.3, page 9, line 30: Change “CO.” to “CO, respectively.”
197. Section 3.3, page 9, line 31: Change “map, which is relative better” to “domain, which is relatively better”
198. Section 3.3, page 10, line 1: Change “Besides of” to “In addition to”
199. Section 3.3, page 10, line 1: Change “determination of –fp-model option should also takes its impact on the performance in to account” to “impact of the –fp-model option on performance should be taken into account”
200. Section 3.3, page 10, line 2: Change “prediction, in an acceptable range” to “prediction,”
201. Section 3.3, page 10, line 3: Change “to gain is more reasonable” to “to gain computational performance”
202. Section 3.3, page 10, line 3: Change “the applications” to “applications”
203. Section 3.3, page 10, line 3: Change “climate simulation” to “climate simulations”
204. Section 3.3, page 10, line 4: Change “more value safe” to “safer”
205. Section 3.3, page 10, line 4: Change “calculation should be required” to “calculations are”
206. Section 3.3.3: I suggest that you change the sub-section name to “CTM Computational Performance”
207. Section 3.3, page 10, line 6: Remove first line of Section 3.3.3
208. Section 3.3, page 10, line 7: Change “of old NAQPMS” to “of the XXX NAQPMS”
209. Section 3.3, page 10, line 7: Change “was as the” to “was the”
210. Section 3.3, page 10, line 8: Change “showed” to “shown”
211. Section 3.3, page 10, line 8: Change “for 72.26% wall-time of whole simulation” to “for 72.26% of the wallclock time for the whole simulation”
212. Section 3.3, page 10, line 9: Change “trade” to “improve”
213. Section 3.3, page 10, line 9: Change “would lead” to “leads”
214. Section 3.3, page 10, line 12: Change “incline” to “improve”
215. Section 3.3, page 10, line 12: Change “using fast=1” to “using the fast=1”
216. Section 3.3, page 10, line 13: Change use of “old” as above
217. Section 3.3, page 10, line 13: Change “codes, which implying the need of optimizing” to “code and supports the need for optimizing”
218. Section 3.3, page 10, line 13: Change “The performance of gas-phase” to “The computational performance of the gas-phase”. Because you are considering both the scientific and computational performance of the models, I think it’s important to be distinguish between them. Hence my suggestion here.
219. Section 3.3, page 10, line 15: Change “MP CBM-Z described” to “MP CBM-Z, as described”
220. Section 3.3, page 10, line 16: Change “showed” to “shown”
221. Section 3.3, page 10, line 16: Change “gets” to “shows speedups of “
222. Section 3.3, page 10, line 17: Remove “speedups”
223. Section 3.3, page 10, line 17: Change “with same” to “with the same”
224. Section 3.3, page 10, line 18: Change “With same” to “With the same”
225. Section 3.3, page 10, line 19: Same comment as above for “old”. Also please add “the” in appropriate places e.g. the gas-phase chemistry, the whole NAQPMS, the XXX CBM-Z
226. Section 3.3, page 10, line 20: Change “releases the potential of existing hardware, the” to “improves the potential on existing hardware and the”
227. Section 3.3, page 10, line 20: Change “relative” to “relatively”
228. Section 3.3, page 10, line 22: Change “enforced” to “improves”
229. Section 3.3, page 10, line 23: Change “speedups could reach” to “speedups reach”
230. Section 3.3, page 10, line 23: Change “with precise” to “with the precise”
231. Section 3.3, page 10, line 24 & 25: Change “adopting fast=1” to “adopting the fast=1”
232. Section 3.3, page 10, line 24: Change “with benchmark” to “with the benchmark”
233. Section 3.3, page 10, line 26: Change “with old CBM-Z” – as above
234. Section 3.3, page 10, line 26: Change “for gas-phase” to “for the gas-phase”
235. Section 3.3, page 10, line 26: Change “and whole” to “and the whole”
236. Section 3.3, page 10, line 26/27: Change “the time consuming proportion of gas-phase” to “the proportion of time taken by the gas-phase”
237. Section 3.3, page 10, line 27: Change “30.74%, which is largely lower than that of 72.26% in baseline version model” to “30.74% compared to 72.26% in the baseline model”
238. Section 3.3, page 10, line 29: Change “MP CBM-Z extended” to “the MP CBM-Z extends”
239. Section 3.3, page 10, line 30: Change “old” here and elsewhere throughout the manuscript
240. Section 3.3, page 10, line 31: Change “times of that on AVX-2” to “times that on the AVX-2”
241. Section 3.3, page 10, line 31: Change “using new CPUs improved about 47%” to “using the new CPUs improved by about 47%”
242. Section 3.3, page 10, line 32: Remove “ability”
243. Section 3.3, page 10, line 32: Change “that the applications, like CTM” to “that applications, like the CTM”
244. Section 3.3, page 11, line 1: Change “relied” to “reliant”
245. Section 3.3, page 11, line 1: Change “to improve its” to “for”
246. Section 3.3, page 11, line 2: Change “with benchmark” to “with the benchmark”
247. Section 3.3, page 11, line 3: Change “old”
248. Section 3.3, page 11, line 3 & 4: Change “MP CBM-Z” to “the MP CBM-Z”
249. Section 3.3, page 11, line 4: Change “of whole” to “of the whole”
250. Section 3.3, page 11, line 4: Change “declined 24%” to “declined by 24%”
251. Section 3.3, page 11, line 4: Change “large” to “largely”
252. Section 3.3, page 11, line 5: Remove “ability”
253. Section 3.3, page 11, line 5: Change “optimizations were required to let the whole CTM adapt to the architecture of the KNL” to “optimizations are required for greater efficiency of the whole CTM on the KNL architecture”
254. Name of Section 4: Change “Conclusion” to “Conclusions”
255. Section 4, page 11, line 9: Change “packing the multiple” to “packing multiple”
256. Section 4, page 11, line 9: Remove “grids”
257. Section 4, page 11, line 10: Change “reconstructed” to “re-structured”
258. Section 4, page 11, line 11 & 12: Change “grids” to “gridboxes”
259. Section 4, page 11, line 13: Change “had similar chemistry processes, the impact of these process on performance was largely limited, and the codes were” to “have similar chemistry processes, the impact of this on the scientific performance was largely limited, and the code was”
260. Section 4, page 11, line 15: Change “with the two generations” to “with two generation”
261. Section 4, page 11, line 16: Remove “respectively”
262. Section 4, page 11, line 16: Change “codes” to “code”
263. Section 4, page 11, line 17: Change “on accuracy” to “on the accuracy”
264. Section 4, page 11, line 19: Change “discrepancies” to “errors in”
265. Section 4, page 11, line 19: Change “by single float” to “by the single float calculations”
266. Section 4, page 11, line 20: Change “with fast=1” to “with the fast=1”
267. Section 4, page 11, line 21: Change “in baseline” to “in the baseline”
268. Section 4, page 11, line 23: Change “to 5.16x and 8.97x speedups” to “to speedups of 5.16x and 8.97x”
269. Section 4, page 11, line 24/25: Change “comparing with performance of old CBM-Z on Intel” to “compared with the unoptimized CBM-Z on the Intel”
270. Section 4, page 11, line 25: Change “to the KNL” to “on the KNL”
271. Section 4, page 11, line 26: Change “version CBM-Z” to “version of CBM-Z”
272. Section 4, page 11, line 26: Remove “For the node”
273. Section 4, page 11, line 28: Two sentences repeated – please remove one
274. Section 4, page 11, line 31: Change “were used” to “was also used”
275. Section 4, page 11, line 31: Change “the real situation” to “more realistic scenarios”
276. Section 4, page 11, line 31: Change “old” as above
277. Section 4, page 11, line 32: Change “with precise” to “with the precise”
278. Section 4, page 11, line 32: Change “tested the on” to “tested on”
279. Section 4, page 12, line 1: Change “outputs of models” to “model outputs”
280. Section 4, page 12, line 1: Change “validate and present” to “evaluate”
281. Section 4, page 12, line 1: Change “as compile” to “as the compile”
282. Section 4, page 12, line 2: Change “between old” to “between the XXX”, as above
283. Section 4, page 12, line 2: Change “using precise” to “using the precise”
284. Section 4, page 12, line 2/3: Change “maximum” to “maximum discrepancy”
285. Section 4, page 12, line 3: Change “leads to more obvious error” to “leads to larger errors”
286. Section 4, page 12, line 4: Change “performance” to “computational performance”
287. Section 4, page 12, line 5: Change “between the accuracy” to “between accuracy”
288. Section 4, page 12, line 6: Change “of codes would largely be improved but” to “of the code is largely improved by”
289. Section 4, page 12, line 6: Remove “even using same codes”
290. Section 4, page 12, line 8: Remove “of” in phrase “of floating-point”
291. Section 4, page 12, line 8: Change “calculation” to “calculations”
292. Section 4, page 12, line 8: Change “to using fast=1option to using the codes” to “to use the fast=1 option”
293. Section 4, page 12, line 14: Change “using a mediate or indirect index” to “using indirect indexing”
294. Section 4, page 12, line 14: Change “the following” to “subsequent”
295. Section 4, page 12, last sentence: Change to “Therefore, it is essential to adopt good practice, e.g. commenting code well and controlling the compile process, for ease of maintenance and development”